1
00:00:00,000 --> 00:00:03,000
Okay, or not?

2
00:00:03,000 --> 00:00:04,000
Sounds great.

3
00:00:04,000 --> 00:00:06,000
I like it now.

4
00:00:06,000 --> 00:00:09,000
Awesome.

5
00:00:09,000 --> 00:00:20,000
Welcome to Game of Nodes, a weekly podcast on the Cosmos from independent Validator teams.

6
00:00:20,000 --> 00:00:23,000
Okay, so now the phrase is punking the guests.

7
00:00:23,000 --> 00:00:24,000
That's apparently what we're doing now.

8
00:00:24,000 --> 00:00:29,000
Hello and welcome to Game of Nodes, a weekly podcast on the Cosmos from independent Validator teams.

9
00:00:29,000 --> 00:00:34,000
And we have the lovely, lovely voiced Greg O'Seary to with us today,

10
00:00:34,000 --> 00:00:37,000
despite his complaints to the contrary.

11
00:00:37,000 --> 00:00:39,000
Thank you for humoring us, Greg.

12
00:00:39,000 --> 00:00:44,000
We're going to be talking about, well, funnily enough, we're going to be talking about Akash today.

13
00:00:44,000 --> 00:00:48,000
Is it, well, actually, first question, Greg, to kick us off, warm us up.

14
00:00:48,000 --> 00:00:50,000
Is it Akash or Akash?

15
00:00:50,000 --> 00:00:51,000
Akash.

16
00:00:51,000 --> 00:00:52,000
Akash.

17
00:00:52,000 --> 00:00:53,000
Okay.

18
00:00:53,000 --> 00:00:56,000
And what is there a deeper meaning behind the name?

19
00:00:56,000 --> 00:01:03,000
Of course, Akash means the sky in Sanskrit or the ether, the first element in Sanskrit.

20
00:01:03,000 --> 00:01:04,000
That is interesting.

21
00:01:04,000 --> 00:01:13,000
Since clouds are formed in the sky, since Akash is a cloud, a cloud, or marketplace, or meeting point for clouds,

22
00:01:13,000 --> 00:01:17,000
the name was derived from there, essentially.

23
00:01:17,000 --> 00:01:21,000
As soon as you pronounced it, ash, I was like, he's going to say Sanskrit.

24
00:01:21,000 --> 00:01:25,000
And many of my ancestors are going to roll over in their graves.

25
00:01:25,000 --> 00:01:26,000
Thank you for that.

26
00:01:26,000 --> 00:01:29,000
I now feel very embarrassed.

27
00:01:29,000 --> 00:01:31,000
Okay.

28
00:01:31,000 --> 00:01:38,000
So this week, obviously we've got Greg, we're going to talk quite a bit about Akash.

29
00:01:38,000 --> 00:01:43,000
And but to kick us off at the top of the show, we're going to do some regular follow ups,

30
00:01:43,000 --> 00:01:47,000
basically rolling off of stuff we had last week.

31
00:01:47,000 --> 00:01:54,000
And the three topics I've got on here are Docsgate, Brontide, and Prop 2021.

32
00:01:54,000 --> 00:01:58,000
I want to go with Brontide first, because that is the most personally exciting.

33
00:01:58,000 --> 00:02:02,000
So Yussepe, you have some thoughts on this excellent band?

34
00:02:02,000 --> 00:02:07,000
Well, I think I've been kind of following, based on some of your recommendations and kind of where Spotify's taking me and everything else.

35
00:02:07,000 --> 00:02:14,000
Brontide has been showing up more and more on my pretty excellent prog slash metal playlist here.

36
00:02:14,000 --> 00:02:24,000
And that's the number of great bands coming out of the UK that have a lifespan of less than three to five years,

37
00:02:24,000 --> 00:02:25,000
but produce amazing albums.

38
00:02:25,000 --> 00:02:28,000
And then those members go do something else and produce other amazing albums.

39
00:02:28,000 --> 00:02:35,000
It's very 2000-y type of thing where you don't have a, you know, somebody's just playing for 40 years just to try to knock it out.

40
00:02:35,000 --> 00:02:39,000
It seems like there's a lot less, a lot more jumping around from bed to bed, if you will.

41
00:02:39,000 --> 00:02:45,000
Maybe that's just creatively to try to find the right thing, or maybe that is, you know, they just understand it's not going to happen or something else.

42
00:02:45,000 --> 00:02:49,000
But it sure produces a great amount of music and very diverse music too.

43
00:02:49,000 --> 00:02:59,000
Fun fact, I think the basis from that band went on to, I think is now the basis in Roto Tamsi, if you kind of more of a metalcore type of vibe.

44
00:02:59,000 --> 00:03:04,000
But, you know, there is a lot of moving around on all these sorts of different projects.

45
00:03:04,000 --> 00:03:10,000
When you're programming hot new features for a cash, Greg, are you, what kind of music do you listen to?

46
00:03:10,000 --> 00:03:17,000
Boy, my favorite programming music is probably

47
00:03:17,000 --> 00:03:23,000
Eagles of Death Metal.

48
00:03:23,000 --> 00:03:25,000
I mean, they're not Death Metal band.

49
00:03:25,000 --> 00:03:27,000
It's called Eagles of Death Metal.

50
00:03:27,000 --> 00:03:29,000
Yeah.

51
00:03:29,000 --> 00:03:32,000
Or BMRC or Black Rebel Motorcycle Club.

52
00:03:32,000 --> 00:03:36,000
Something that's just like jing-jing-jing on the background.

53
00:03:36,000 --> 00:03:44,000
It's quite raunchy, kind of, sort of almost down back to basics, kind of bluesy rock and roll, right?

54
00:03:44,000 --> 00:03:54,000
Yeah, blues rock and roll is usually my go-to, but something that gets away in the background and doesn't take my focus away to the words, right?

55
00:03:54,000 --> 00:03:56,000
Like, yep.

56
00:03:56,000 --> 00:04:02,000
But I'm listening to music at least 10 hours a day.

57
00:04:02,000 --> 00:04:04,000
I'm plugged in all the time when I'm recording.

58
00:04:04,000 --> 00:04:06,000
Yeah, especially when I'm recording.

59
00:04:06,000 --> 00:04:10,000
Like, yeah, what about yours?

60
00:04:10,000 --> 00:04:17,000
I think we've got over our tastes quite a bit on the show previously, but this is obviously why we're talking about music as a regular thing,

61
00:04:17,000 --> 00:04:25,000
but I think Söpp and I are kind of more of a progressive rock kind of vibe a lot of the time.

62
00:04:25,000 --> 00:04:30,000
And I think the actualsy I think is into mainly achingly cool stuff, right?

63
00:04:30,000 --> 00:04:34,000
Yeah, something like a more melodic type music.

64
00:04:34,000 --> 00:04:39,000
Yeah, yeah. And then I feel like Noel's a stealth metalhead. He keeps really quiet.

65
00:04:39,000 --> 00:04:44,000
Yeah, like metal and trance.

66
00:04:44,000 --> 00:04:47,000
I like a bit of trance while I'm like bashing on the keyboard.

67
00:04:47,000 --> 00:04:53,000
Stuff that I can't understand any words to, because I'm just not distracting.

68
00:04:53,000 --> 00:04:59,000
I actually went and saw Dvorak's New World Symphony over the weekend, which is like, it's got some nutcracker in there.

69
00:04:59,000 --> 00:05:03,000
It's got all sorts of nonsense with the organ symphony. So that was great timing.

70
00:05:03,000 --> 00:05:05,000
That's cool.

71
00:05:05,000 --> 00:05:11,000
It might be a bit like classical music.

72
00:05:11,000 --> 00:05:21,000
I mean, in the classical music, I've got like a 101 symphony album that I just like put on in the car and when I'm working sometimes.

73
00:05:21,000 --> 00:05:23,000
I reckon it's great. She'll see.

74
00:05:23,000 --> 00:05:27,000
It is. Not the age myself, but I'm a big rock man enough and most are fan too.

75
00:05:27,000 --> 00:05:31,000
I don't know. I think those are not that far from these other type of genres we're talking about.

76
00:05:31,000 --> 00:05:36,000
I know it seems like there's a big divide between some of that stuff and trance and those types of things.

77
00:05:36,000 --> 00:05:38,000
Not that much of a divide actually if you get into the right.

78
00:05:38,000 --> 00:05:46,000
It's a percussion, isn't it? Like that. I think that's the big, like having kind of like a really specific backbeat is like very uncommon in classical music.

79
00:05:46,000 --> 00:05:50,000
And if you cannot live without that.

80
00:05:50,000 --> 00:05:52,000
For example, like, kind of quite common.

81
00:05:52,000 --> 00:06:08,000
That thing of what people say like ambient music is like not that far away from like string, string sort of classical, maybe that is very sort of based on a large string section, moving at a slowish pace, you know, obviously not like Sturm and Drang or something where it's very dramatic and

82
00:06:08,000 --> 00:06:10,000
with brass and whatnot.

83
00:06:10,000 --> 00:06:19,000
But, you know, the snare drum only comes along as an instrument in the orchestra in the romantic period, right?

84
00:06:19,000 --> 00:06:22,000
So please, musicologists, tell me.

85
00:06:22,000 --> 00:06:29,000
I mean, I can only go as far as back to like violin of like the 1600s, everything after that. I'm like, I don't know.

86
00:06:29,000 --> 00:06:33,000
Yeah, dude, I don't know the history of it. I just enjoy to see it a bit, all right.

87
00:06:33,000 --> 00:06:36,000
Yeah, that's exactly right.

88
00:06:36,000 --> 00:06:55,000
I'm just saying that like the best classical music is the succession intro where it is a dissonant string section with a tritone in it and the hip hop beat that that is like, what class if classical music had a hip hop beat, it would be est here in my

89
00:06:55,000 --> 00:07:06,000
If classical wasn't classical, the best part of any show is probably the tune up for me. When you get to hear everything started coming and harmonize. That's the best.

90
00:07:06,000 --> 00:07:10,000
No, no, harmonization, the best.

91
00:07:10,000 --> 00:07:16,000
Okay, so I have a successfully completely derailed what should have been like a one minute how are we doing.

92
00:07:16,000 --> 00:07:31,000
I sorry about that. So do we want to talk very, very briefly about what's going on with prop 20 prop 21. As we've just we've just done a prop 20 upgrade on on Juno.

93
00:07:31,000 --> 00:07:37,000
And there is a slight little problem with it isn't there.

94
00:07:37,000 --> 00:07:53,000
Not with the prop. Well, the upgrade went swimming well, right? It sounds like that in the the contract says we're testing everything else associated with prop and the software upgrade was was actually a cosmo visor running I actually forgot about it until it was over.

95
00:07:53,000 --> 00:08:02,000
I just got notified just with the pause but then I assume that prop 21 you said there there is something that was found or at least something that needs to be updated.

96
00:08:02,000 --> 00:08:10,000
Yeah, so there's a there's a basically a small problem in the IBC handler which we've actually been running until now without noticing.

97
00:08:10,000 --> 00:08:21,000
But essentially, if you want to update an IBC channel via gov at the moment, you can't do you know we've never had to do it before which why we didn't notice that there was a bug like a line of code that is changing.

98
00:08:21,000 --> 00:08:33,000
So I think that's in prop 21 the upgrade and also crucially the the funds need moving from a from a.

99
00:08:33,000 --> 00:08:42,000
It's basically like I think it's more of a it's either a fat finger or miscommunication essentially that the funds were moved to a null address.

100
00:08:42,000 --> 00:08:52,000
And then we're supposed to be moved to smart contract but instead they would just move to another dress, which was not what we did in testing.

101
00:08:52,000 --> 00:08:59,000
So there was a little bit of a disconnect between what we ran on the test net and then what actually ran on main net in terms of like each one of the steps.

102
00:08:59,000 --> 00:09:11,000
So now the second half needs to happen where it just gets moved over where after being on stake to moved it gets moved over to smart contract but I didn't really have any involvement in that part.

103
00:09:11,000 --> 00:09:18,000
So it was a surprise to me as well because I didn't write or work on any of the code for the actual move side of things.

104
00:09:18,000 --> 00:09:28,000
So I obviously saw it tested because you know we're helping out with the test nets and whatnot and helped out with smart contract but because it obviously worked in the test nets.

105
00:09:28,000 --> 00:09:29,000
It was like okay cool.

106
00:09:29,000 --> 00:09:33,000
Well, you guys who are working on that bit you've seemed to have under control whatever.

107
00:09:33,000 --> 00:09:46,000
And then today like I was also away from keyboard when the upgrade happened and it worked whatever got back and it's like, oh, hey guys, what was what and I'm like, what's happening.

108
00:09:46,000 --> 00:09:49,000
So yeah, it's gonna be an upgrade in a couple of days.

109
00:09:49,000 --> 00:09:50,000
Not really a big deal.

110
00:09:50,000 --> 00:09:52,000
I think that's kind of the end of that.

111
00:09:52,000 --> 00:09:58,000
Actually, you know, as another, since we're doing follow ups and tie ups from last week that have come back rounds.

112
00:09:58,000 --> 00:10:06,000
Well, something we were talking about last week on last week's episode Greg was that the the upgrade model of cosmos.

113
00:10:06,000 --> 00:10:16,000
I think has gotten a lot of attention recently because of the variety of chains of upgrades and halts, etc. that have been happening.

114
00:10:16,000 --> 00:10:27,000
And we were we've sort of had like quite a long conversation on whether or not it was some fundamental property of the way cosmos works or whether it's just that these chains are now.

115
00:10:27,000 --> 00:10:38,000
Scaling and sophistication and user base and it's sort of correlation rather than causation just that there are a lot of chains moving very fast now.

116
00:10:38,000 --> 00:10:47,000
Naturally, sometimes they're going to be issues they get stuck the teams do some work on stick them and it's not really anything to worry about.

117
00:10:47,000 --> 00:10:49,000
So what's the question.

118
00:10:49,000 --> 00:11:03,000
So the question is, do you think it's like correlation or causation or do you think it's like, do you think there's some fundamental thing about cosmos that means these problems sort of bubbling up now or do you think it's just noise because we're moving quickly.

119
00:11:03,000 --> 00:11:06,000
I think it's lack of coordinate.

120
00:11:06,000 --> 00:11:16,000
I mean, it's hard to say right like we had an issue with upgrade module recently we upgraded the chain and that was a disaster.

121
00:11:16,000 --> 00:11:24,000
I feel like I mean, and it was also like, well, we didn't include the upgrade model in a testnet.

122
00:11:24,000 --> 00:11:31,000
So we weren't able to test the upgrade model module, assuming that upgrades are going to just be smooth.

123
00:11:31,000 --> 00:11:39,000
And assuming that's just going to work out of the box because that's the least sort of like critical piece of the the whole main net upgrade.

124
00:11:39,000 --> 00:12:08,000
But that didn't end up to be true. And it was just a holy hell of like pushing out a patch waiting for people to, you know, come online the validators are biggest validators dropped and then people were out of sync so that we sink and it's a whole nasty sort of like, you know, the mediation process right so I think, I think it's just

125
00:12:08,000 --> 00:12:17,000
I think the people are moving quite fast and there are different use cases I think are not being tested properly.

126
00:12:17,000 --> 00:12:22,000
The test is the idea of testness have sort of like gone out of fashion.

127
00:12:22,000 --> 00:12:30,000
Some of them do it, but we even the testness there no there's no guarantee that if the code passes through the test is going to actually like work.

128
00:12:30,000 --> 00:12:31,000
Right.

129
00:12:31,000 --> 00:12:44,000
So, with the kind of with the kind of speed people are moving at so I think they need to be a cosmos collective for testing these like, like modules.

130
00:12:44,000 --> 00:13:00,000
I don't know if there is one or if ignite team is fully like plugged in to what's going on in the ecosystem and also like it's wrong for us to actually expect ignite team to to be the, you know, the

131
00:13:00,000 --> 00:13:11,000
responsible parties right so yeah things are moving way faster than the core team of cosmos members can keep up with.

132
00:13:11,000 --> 00:13:14,000
I think that's what's causing all these issues right like.

133
00:13:14,000 --> 00:13:36,000
Great that was a SDK issue or a wasmish I forget, but the SDK issue module supposed to introduce cosmos wiser and supposedly the ideal scenario is when you when you vote on on a proposal agree on what block you're going to grade it'll automatically do that great.

134
00:13:36,000 --> 00:13:48,000
Supposing that was a promise, but the reality is like well, you know, there was some no pointer exception some memory some some nasty thing.

135
00:13:48,000 --> 00:13:49,000
Right.

136
00:13:49,000 --> 00:13:54,000
And I don't know why we didn't.

137
00:13:54,000 --> 00:14:13,000
Why we assume just going to work and that just gives you the importance of actually having to test the actual binary you're going to know the other commit hash you're going to like upgrade on and not assuming that this coming touch is going to or this this version is going to work right like low hand and most of

138
00:14:13,000 --> 00:14:20,000
I think Juno also had the problem or even sorry and most have the same problem right like.

139
00:14:20,000 --> 00:14:33,000
Juno and have more so hard holds recently Juno was a cyber attack F moss was with Sir, she'll see no you guys.

140
00:14:33,000 --> 00:14:36,000
That was a upgrade snafu.

141
00:14:36,000 --> 00:14:46,000
Basically during the upgrade there was a we didn't get consensus straight away and then there was a lot of node resetting and losing state files.

142
00:14:46,000 --> 00:14:50,000
So was it was it validator error then.

143
00:14:50,000 --> 00:14:54,000
It was just it was a communal like error.

144
00:14:54,000 --> 00:15:00,000
I guess you would say like there was a lot of people telling people to do things that they shouldn't be doing and.

145
00:15:00,000 --> 00:15:15,000
The impetus on that was I think there was there was a wasm there was a zero day was some sort of zero day wasm type Dylan Pratt keep you going but there was some there was some issue that was kind of forcing an emergency upgrade and then that emergency upgrade.

146
00:15:15,000 --> 00:15:21,000
Turned into a mix of bad communication and then there were spreadsheets and trying to get the make sure we had the consensus on it.

147
00:15:21,000 --> 00:15:27,000
And then there was a version switch at the last minute which is always a bad thing right like somebody set it up.

148
00:15:27,000 --> 00:15:31,000
That was was it last minute or during trying to get consensus.

149
00:15:31,000 --> 00:15:34,000
I think it was during what I'm trying to get it through.

150
00:15:34,000 --> 00:15:36,000
That's right. Yeah, it was so yeah.

151
00:15:36,000 --> 00:15:49,000
And so they did a bin switch and then you know something and then consent it just the banners were just confused as shit and then trying different things to make it work and people to do stuff they shouldn't do.

152
00:15:49,000 --> 00:15:53,000
And then a bunch of nodes got double signed big ones as well.

153
00:15:53,000 --> 00:15:56,000
So they just had to stop the network.

154
00:15:56,000 --> 00:16:00,000
What happened to the double signers to the lose money.

155
00:16:00,000 --> 00:16:12,000
No, because they they wound back to the block before the doubles. So they when they restarted the network, they did like a state migration and then wound back a block.

156
00:16:12,000 --> 00:16:14,000
That's right. Yeah.

157
00:16:14,000 --> 00:16:17,000
Oh, they are the rollback. Yeah.

158
00:16:17,000 --> 00:16:20,000
Most useless rollback ever.

159
00:16:20,000 --> 00:16:23,000
But yeah, it was like one block.

160
00:16:23,000 --> 00:16:26,000
One block. That's right. Yeah.

161
00:16:26,000 --> 00:16:31,000
Just that one block that they were trying to, you know, get going.

162
00:16:31,000 --> 00:16:38,000
So it was the I think it was pretty much just it picked up where the original upgrade was supposed to pick up.

163
00:16:38,000 --> 00:16:39,000
That's right.

164
00:16:39,000 --> 00:16:44,000
Right. So yeah, that so the upgrade module leads to the consensus failures.

165
00:16:44,000 --> 00:17:00,000
I've seen that quite a happen quite a bit and worst scenario is if you have a big validator that was, you know, that has a state that's out of sync, and they just keep like printing blocks.

166
00:17:00,000 --> 00:17:13,000
And everyone else is like out of consensus right so you're going to have these like crazy scenarios that you have to re coordinate and then the big validator has to like re sync the state to

167
00:17:13,000 --> 00:17:19,000
a to a to a to a trust group that's much smaller.

168
00:17:19,000 --> 00:17:40,000
But also like we're learning quite a bit like we're learning the importance of like better distribution right like is one of the one of the reasons why my gosh upgrade took about 1618 hours, and we're up all the time and the big reason was we had one of our biggest validators didn't show up on time or at least they showed up on time

169
00:17:40,000 --> 00:17:58,000
they actually did what they're supposed to do, but they weren't present for us to be able to debug and like coordinate and overclock lapsed the company, the bills, cash network, the core contributors we don't have enough voting power to influence a vote in the in the, you know, in the thing so

170
00:17:58,000 --> 00:18:06,000
it's decentralized launches that extremely hard and especially highly risky when you have validators with with a lot of voting power.

171
00:18:06,000 --> 00:18:07,000
Yeah, right.

172
00:18:07,000 --> 00:18:20,000
And we've spent a lot of time on Juno, especially given the cadence of the many upgrades that have happened in the last two months, what was the cyber attack what the three security patches that we've had for cosmogism.

173
00:18:20,000 --> 00:18:28,000
Plus the plus, you know, the fact that we've had routine stuff going on as well that we've been trying to get through.

174
00:18:28,000 --> 00:18:50,000
What we've had to do is basically have there is a coordination channel where we say we have a whole, there's now a releases procedure in the GitHub repo say even if people don't really want to engage with, I suppose the Juno team, they know where they can get those updates and then we've also had to be like well for a security

175
00:18:50,000 --> 00:19:02,000
perspective we can't reveal everything publicly all the time anymore. Right. So some of the validators who sort of were relying on us just tweeting all the information and then just like knocking it out.

176
00:19:02,000 --> 00:19:15,000
We've had to say well, look, the problem is we can't do that because there's security risks so we kind of need to know if not exactly who you are you don't have to dox yourself we do need to have a place to contact you otherwise it's not going to work.

177
00:19:15,000 --> 00:19:25,000
Or at least it's not, I mean, well, it will work but the first notification you're going to get of an upgrade is when everybody else stops upgrades and you have hash.

178
00:19:25,000 --> 00:19:39,000
That was one of my main concerns with upgrade or cosmos prop 69 the cosmogism upgrade, because I was like well, for Juno we've had a couple of like shadow upgrades right and that's going well because many of the validators are pretty plugged in.

179
00:19:39,000 --> 00:19:51,000
However, with cosmos those upgrades tend to take quite a bit longer because it seems like those validators aren't quite as engaged so if there is a security breach or a topic that needs to come out now, or we can be able to react as quickly.

180
00:19:51,000 --> 00:19:53,000
That was one of my main concerns.

181
00:19:53,000 --> 00:20:04,000
But we're ultimately voting yes anyway. Yeah, yeah, so you told me the addition of cosmogism to the hub. Yeah, yeah, yeah, yeah.

182
00:20:04,000 --> 00:20:12,000
I think, I think in Juno we're a little bit more lucky in that we don't have as many corporate validators in Juno.

183
00:20:12,000 --> 00:20:18,000
So it's like a little bit more accessible to the actual people who are running the nodes.

184
00:20:18,000 --> 00:20:47,000
I think the other things as well I mean with Manita we did two tests we did so we did two tests that's pre launch and then pre monitor we had the mainline test net which got killed by we are cyber attacking it me and Ben X for and then we had to restart it didn't we because we killed it and so any validator that was a part of those tests

185
00:20:47,000 --> 00:21:00,000
nets had like the experience of like chainhawk restart with Cosm wasm and all the like, oh now you need to shuffle the wasm folder and all that kind of stuff and then obviously then the final upgrade out.

186
00:21:00,000 --> 00:21:04,000
We did like what was it for test nets in three weeks.

187
00:21:04,000 --> 00:21:23,000
I think that you mean there's like a core of validators there who really know their stuff and like can help other validators when stuff goes wrong. I mean, and to be on site sometimes there's things that when I'm tired I screw up and I'll literally ask and one of those validators that was a part of that

188
00:21:23,000 --> 00:21:35,000
a starty set will literally be like DM me and like right dude here's what you need to do don't worry about it and you're like ah this is why these guys are great this is all these guys are so good.

189
00:21:35,000 --> 00:21:46,000
We have a pretty good participation in uni like in terms of number of validators from the mainnet set there's quite a few in our test net which is good.

190
00:21:46,000 --> 00:21:48,000
Yeah running.

191
00:21:48,000 --> 00:21:53,000
The whole upgrade is a bonding process for sure. It's a good way to put it.

192
00:21:53,000 --> 00:22:03,000
It was amazing like most people forget how important validators are for supporting these networks and and really.

193
00:22:03,000 --> 00:22:22,000
I mean a lot of cash validators is just amazing to see how supportive you know especially the small guys are right like the big guys don't really care as much as a small guys do so makes you really reevaluate your your delegations and we're going to do a whole revamp of our own foundation delegates

194
00:22:22,000 --> 00:22:42,000
to people that actually added support and were critical when we needed them but boy it's just been such a such a bonding experience and also a learning experience and and a reminder to thank you validators should have a validator day.

195
00:22:42,000 --> 00:22:44,000
Well that's a validator day.

196
00:22:44,000 --> 00:22:50,000
Every day is validator day for us.

197
00:22:50,000 --> 00:23:04,000
Well that was part of the reason to actually talk about game of notes was was that the independent validators I think sometimes get overlooked or you have you know everybody and I think there's a there's a pretty blinds pretty large blind spot in in terms of.

198
00:23:04,000 --> 00:23:19,000
Delegators knowing how to pick validators what's important and then they're all ranked right so when you see one to 10 out of 125 people I'm safe right that's what that's what you see and there's there's really that idea around decentralization and some of these errors errors that we were talking about issues we're talking about.

199
00:23:19,000 --> 00:23:35,000
Involvement and buy in and those types of things like like somebody in that top 10 I know like maybe a an exchange wallet or something similar that doesn't have they're not involved the way you know the phrase or cake notes is or lavender five is right.

200
00:23:35,000 --> 00:23:46,000
But again that's difficult to you really have to take some time to really build the marketing and the branding around that for people to understand what's the difference there so I think that that whole idea on proof of stake.

201
00:23:46,000 --> 00:23:51,000
That is a real misconception I think in terms of delegators in terms of what's important.

202
00:23:51,000 --> 00:24:01,000
Do you think you surfer the the the difficulty there is I mean I say from like my my capacity and our capacity as a small validator company is like.

203
00:24:01,000 --> 00:24:22,000
Maybe as well because we do that development and we have other projects but like we can basically support to change to like a meaningful like we're more engaged than the average right and more than that we have to be just another validator in the pack even though we're an independent validator and we rely on.

204
00:24:22,000 --> 00:24:34,000
Like to the extent this is why we've only we've only run three chains and the third chain we run as a meme chain because it's a way for us to test stuff out in production to some extent right.

205
00:24:34,000 --> 00:24:38,000
But it's quite I mean so you know Greg we've looked at.

206
00:24:38,000 --> 00:24:52,000
Joining the cash set because it's a really interesting project right yes but I kind of feel like it's the amount of work I do other than just validating personally and that we try and do as a company.

207
00:24:52,000 --> 00:25:05,000
For the chains that we work on like I already feel like I don't do nearly as much for Stargat is say then we did before they launched like when they were doing the test and stuff we tried to be.

208
00:25:05,000 --> 00:25:12,000
Because those folks were in the US we tried to be helping out in the European time zone and doing some kind of stuff and you're chipping in what we could and what not.

209
00:25:12,000 --> 00:25:17,000
And I feel like because the amount of work in June is quite big.

210
00:25:17,000 --> 00:25:26,000
And we were quite involved over there that's meant that we can't be as involved over here and so then I look at other chains like like yours really interesting project.

211
00:25:26,000 --> 00:25:41,000
Seems very involved validator set you know we're for some depth development stuff we're doing we're looking to move stuff long term it's like you know dowels running their own infrastructure on a cash and all that's really really interesting to us but we're also just like well if we can't make.

212
00:25:41,000 --> 00:25:55,000
A meaningful contribution outside of just validating and we try and buy our way into the set and maybe knock somebody out who we don't know maybe is really helping out area right what are we adding should we even be in that validator set probably.

213
00:25:55,000 --> 00:25:57,000
Not right.

214
00:25:57,000 --> 00:25:59,000
I mean.

215
00:25:59,000 --> 00:26:04,000
Folks that are adding value on a cash will will get delegations.

216
00:26:04,000 --> 00:26:13,000
I'm very active in the discourse in the Twitter's and especially in the upgrades I was there for the whole 16 to 18 hours right.

217
00:26:13,000 --> 00:26:17,000
So I know what validators added value.

218
00:26:17,000 --> 00:26:25,000
And.

219
00:26:25,000 --> 00:26:28,000
And about it is that don't add value right so if I see a validator getting knocked out of the act to set a validator that I feel like should not be knocked out.

220
00:26:28,000 --> 00:26:34,000
I'm going to give my personal or and I'm going to propose our foundation to give delegation right.

221
00:26:34,000 --> 00:26:44,000
So it's a very much approved statement very much not very much like can like can contextual conditional is no.

222
00:26:44,000 --> 00:26:56,000
You know rulebook right like but if there is someone who just you know has a lot of money I mean we have some validators who you know just bought their way in they don't do anything.

223
00:26:56,000 --> 00:26:57,000
Yeah.

224
00:26:57,000 --> 00:27:05,000
And they're not going to get the color they might personal love and foundation delegation right so.

225
00:27:05,000 --> 00:27:15,000
I mean there's no right way to do it and of course we want our validators to be hyper involved.

226
00:27:15,000 --> 00:27:22,000
But you know buying tokens and holding on to those tokens is also contributing to security network right so it's not they're not like useless they are contributing to security network.

227
00:27:22,000 --> 00:27:24,000
So.

228
00:27:24,000 --> 00:27:26,000
But.

229
00:27:26,000 --> 00:27:35,000
Delegations are one way to keep the validators in the active set without having to buy the tokens right and I think that should be a more natural.

230
00:27:35,000 --> 00:27:41,000
We want to see that because the value of contribution is much higher than than than value of capital.

231
00:27:41,000 --> 00:27:44,000
So.

232
00:27:44,000 --> 00:27:53,000
And we're also increasing our validators set right we do are I think the proposals every three or four months to increase the validators set.

233
00:27:53,000 --> 00:28:00,000
I mean you got to keep in mind because the validators at the slower and the larger the chain gets right.

234
00:28:00,000 --> 00:28:11,000
But they're always straight out but you know we feel like if the if the minimum required to be in the active set is not reasonable.

235
00:28:11,000 --> 00:28:15,000
There's somebody in our community and most of the proposals don't come from us.

236
00:28:15,000 --> 00:28:24,000
Somebody in our community will put a proposal and we'll have you support it in whatever way we can.

237
00:28:24,000 --> 00:28:26,000
And and then you have like.

238
00:28:26,000 --> 00:28:32,000
Ethereum and like these other chains here thousands of thousands of validators.

239
00:28:32,000 --> 00:28:33,000
I wonder how.

240
00:28:33,000 --> 00:28:39,000
Upgrades if there is anything of such in Ethereum is going to happen in the proof of stake network right.

241
00:28:39,000 --> 00:28:42,000
Very curious about networks that have.

242
00:28:42,000 --> 00:28:45,000
Enormous amounts of nodes right.

243
00:28:45,000 --> 00:28:53,000
I guess Ethereum you don't really have to worry too much because I think it's the leader finance folks that control like 50% of the nodes is some crazy number like that.

244
00:28:53,000 --> 00:29:01,000
I just need to go to one and the whole notion of like Ethereum is more decentralized because has no more nodes is just.

245
00:29:01,000 --> 00:29:06,000
Ridiculous right because you know one company controls a lot of these nodes.

246
00:29:06,000 --> 00:29:12,000
Yeah but I guess but in a really true scenario where you have thousands of thousands of nodes and truly distributed.

247
00:29:12,000 --> 00:29:14,000
I wonder how.

248
00:29:14,000 --> 00:29:20,000
Upgrades happen especially when you have to do security patches the really rapid security patches right.

249
00:29:20,000 --> 00:29:28,000
Yeah, so I wonder if there's like a model like I'm trying to remember the name of the protocol.

250
00:29:28,000 --> 00:29:32,000
I think you have hollow chain.

251
00:29:32,000 --> 00:29:37,000
Yeah, I am thinking of hollow chain.

252
00:29:37,000 --> 00:29:45,000
I think where the the sort of the date of the validation integrity of the blocks if you like a kind of push to app level on they.

253
00:29:45,000 --> 00:29:54,000
So I don't exactly what their upgrade model is but I presume is that most clients are obviously not keeping full state it's more of a bit torrent model where.

254
00:29:54,000 --> 00:30:03,000
It's just more about current version of protocol for two things that want to actually sort of be clients in the network so.

255
00:30:03,000 --> 00:30:08,000
There's probably a bunch of quite interesting different ways of cutting this problem.

256
00:30:08,000 --> 00:30:15,000
But I wonder if like you know the more fundamental thing that's maybe relevant for Cosmos is you know should we be tying.

257
00:30:15,000 --> 00:30:23,000
Upgrades to the concept of governance right because that there are situations in which a software upgrade might be conscientious and we've we've.

258
00:30:23,000 --> 00:30:25,000
We've certainly seen that in June.

259
00:30:25,000 --> 00:30:29,000
Oh right that's that's definitely happened over the last 10 weeks.

260
00:30:29,000 --> 00:30:42,000
But for like the two that we've had prop 20 prop prop 20 is funny because it's controversial prop 21 is literally a like oh minor bug let's knock out a release in a kind of.

261
00:30:42,000 --> 00:30:50,000
If you worked in industry and done continuous deployment or continuous delivery you would literally hit the button the second time ago oops.

262
00:30:50,000 --> 00:30:53,000
And then it would be out five minutes later right.

263
00:30:53,000 --> 00:31:06,000
It wouldn't be a thing and so there is like there is an interesting question I suppose given that we're seeing that governance in Cosmos is coming up against challenges whether or not.

264
00:31:06,000 --> 00:31:11,000
I suppose whether or not there there will be changes in that direction like are there types of.

265
00:31:11,000 --> 00:31:21,000
So something I think D me on the core team for Juno was talking about and I think Jack was talking about maybe on the podcast a while ago was like whether we should be having to speed.

266
00:31:21,000 --> 00:31:27,000
Governance for software upgrades like you say emergency is like one day.

267
00:31:27,000 --> 00:31:32,000
Regular is like five day and quorum is maybe different for those two.

268
00:31:32,000 --> 00:31:39,000
But then you know people try and sneak through things on the one they shouldn't on the so it becomes back it comes back to this like.

269
00:31:39,000 --> 00:31:48,000
How trustless are we and do we have to trust our validators and I think the answer is and I've been shitposting about this on Twitter for some time now is that.

270
00:31:48,000 --> 00:31:53,000
In my opinion proof of stake is actually just a distributed system it's not.

271
00:31:53,000 --> 00:32:00,000
For some definition of decentralized decentralized is a distributed system with multiple operators who don't necessarily trust each other.

272
00:32:00,000 --> 00:32:08,000
But in practice they actually have to so in practice you've got a hundred operators rather than one operator and that's only like one order of magnitude different right.

273
00:32:08,000 --> 00:32:11,000
So it's a good question.

274
00:32:11,000 --> 00:32:13,000
The.

275
00:32:13,000 --> 00:32:18,000
I don't I think the term trustless is fundamentally stupid.

276
00:32:18,000 --> 00:32:29,000
You trust something at the end you know beat at the cell level beat at whatever level there is something that you're trusting if you're running a Bitcoin node you're still trusting the CPU is going to act.

277
00:32:29,000 --> 00:32:33,000
So it's going to be good right you're trusting something.

278
00:32:33,000 --> 00:32:39,000
I think decentralization is about removing attack vectors right.

279
00:32:39,000 --> 00:32:53,000
When in considerable chain whole thing or double spending or you know verification right things of that nature but it's not removing trust entirely you can't like.

280
00:32:53,000 --> 00:32:55,000
Like you got a trust.

281
00:32:55,000 --> 00:32:59,000
Look at the access point let's talk about Kepler right.

282
00:32:59,000 --> 00:33:06,000
You got a trust that Kepler you're installing on your computer is valid and verified by someone else.

283
00:33:06,000 --> 00:33:09,000
You're trusting Google at that point.

284
00:33:09,000 --> 00:33:14,000
But you can't use it a decentralized network.

285
00:33:14,000 --> 00:33:19,000
Without some level of trust.

286
00:33:19,000 --> 00:33:24,000
The only question is how much of trust do we require right so something like.

287
00:33:24,000 --> 00:33:36,000
Chain upgrades right like I mean you're trusting validators to do the necessary diligence intelligence and you know review the code like I don't know if you guys do review the code but most of them don't.

288
00:33:36,000 --> 00:33:38,000
Right review the code.

289
00:33:38,000 --> 00:33:41,000
And you know.

290
00:33:41,000 --> 00:33:45,000
Literally like act on behalf of the people right.

291
00:33:45,000 --> 00:33:53,000
And you have to coordinate you have to learn who other validators I think there's no truly.

292
00:33:53,000 --> 00:33:59,000
Even as a Bitcoin like this is upgrade for Bitcoin there you're still trusting the core team.

293
00:33:59,000 --> 00:34:10,000
That they are acting on behalf of the people not everybody's being you know my grandma has Bitcoin she's not reading code.

294
00:34:10,000 --> 00:34:13,000
She running the client.

295
00:34:13,000 --> 00:34:20,000
Right like she's just trusting somebody to run the client for her.

296
00:34:20,000 --> 00:34:29,000
But that trust is inherently not dangerous because you know her trust in the client doesn't mean is removing the trust from the network.

297
00:34:29,000 --> 00:34:30,000
Yeah.

298
00:34:30,000 --> 00:34:35,000
Like so trust minimize is a better word than trust less.

299
00:34:35,000 --> 00:34:46,000
The minimal trust you you you need the lower friction that you get to interact between two parties in that particular network.

300
00:34:46,000 --> 00:34:50,000
The lower friction you have the higher scalability you have for a system.

301
00:34:50,000 --> 00:34:51,000
Right.

302
00:34:51,000 --> 00:34:57,000
So Bitcoin or decentralized systems is removed or lower the friction of interaction between two parties for whatever reason.

303
00:34:57,000 --> 00:35:03,000
We had kind of commerce in a car she now do deployments in Juneau actually run smart contracts.

304
00:35:03,000 --> 00:35:04,000
Right.

305
00:35:04,000 --> 00:35:10,000
You don't need a high level high degree of trust from the other party in order to function.

306
00:35:10,000 --> 00:35:11,000
Right.

307
00:35:11,000 --> 00:35:14,000
And that's what makes decentralized systems fascinating.

308
00:35:14,000 --> 00:35:16,000
Yeah.

309
00:35:16,000 --> 00:35:18,000
Whatever term on it right.

310
00:35:18,000 --> 00:35:20,000
Ultimately everything is a distributed system.

311
00:35:20,000 --> 00:35:21,000
Ultimately.

312
00:35:21,000 --> 00:35:22,000
Right.

313
00:35:22,000 --> 00:35:29,000
The only difference is how much how much trust you need to participate or run the network.

314
00:35:29,000 --> 00:35:30,000
Yeah.

315
00:35:30,000 --> 00:35:32,000
And when it comes to like yeah.

316
00:35:32,000 --> 00:35:38,000
I feel that upgrades should be.

317
00:35:38,000 --> 00:35:44,000
Frictionless the actual process of upgrading from one version to another migrating the data.

318
00:35:44,000 --> 00:35:46,000
I think that process should be fixed frictionless.

319
00:35:46,000 --> 00:35:50,000
But what goes in the upgrade should not be for us.

320
00:35:50,000 --> 00:35:52,000
That should be on debate.

321
00:35:52,000 --> 00:35:54,000
Absolutely needed.

322
00:35:54,000 --> 00:35:58,000
We all after you come to a consensus what goes to the upgrade.

323
00:35:58,000 --> 00:36:00,000
I think from there on it should be completely frictionless.

324
00:36:00,000 --> 00:36:09,000
All that means is we should be able to ship upgrades when we need and have enough time for the community to come to consensus.

325
00:36:09,000 --> 00:36:10,000
Right.

326
00:36:10,000 --> 00:36:11,000
Yeah.

327
00:36:11,000 --> 00:36:23,000
And the code that matters then there you know that you can you can look at it a different way which is maybe back to the like I said not an expert on Holochain but gather there are some similarities between their model and the bit torrent model.

328
00:36:23,000 --> 00:36:30,000
You know maybe we should be looking at a thing where you dirty secret most validators don't run an archive node.

329
00:36:30,000 --> 00:36:36,000
Most validators doing aggressively especially on large chains especially on smart contract chains.

330
00:36:36,000 --> 00:36:43,000
If you're part of a chain upgrade you know what the difference between prune or full load is.

331
00:36:43,000 --> 00:36:44,000
Yeah.

332
00:36:44,000 --> 00:36:45,000
Right.

333
00:36:45,000 --> 00:37:03,000
So you know there's a whole bunch of things here where you go OK well actually a lot of the complexities of managing that upgraded to do with the state and the validity of the state changes effectively and whether or not it's consensus breaking and what the next block off the upgrade is going to do.

334
00:37:03,000 --> 00:37:15,000
But keeping almost no state on the client would sort of make that whole process a lot easier but then you kind of get back into the well OK so where's the system integrity come from.

335
00:37:15,000 --> 00:37:26,000
And that's like OK so if we've got lots of peer to peer nodes that just going to interact with each other where's the archive of state.

336
00:37:26,000 --> 00:37:35,000
Then we've designed a new system that's not quite what we're currently working with right. So it's I don't know to me this is all very as a kind of like systems designer.

337
00:37:35,000 --> 00:37:37,000
I think this is all quite interesting.

338
00:37:37,000 --> 00:37:46,000
Now which gives us the archive archive now or something that gives us a whole new states and very fine what not.

339
00:37:46,000 --> 00:38:03,000
I mean I mean I did to be perfectly honest I mean I think you're you're semi joking about that. But I don't think it's necessarily a bad idea. I mean like there is an argument for doing something that is stateless with more of like a cell towers model where your archive nodes are effectively a cell towers going like.

340
00:38:03,000 --> 00:38:13,000
Yep. No you've just asked me about data integrity and I can tell you the hash that you've just given me is the one that the majority of nodes or a sub graph of nodes have given me and that seems reasonable.

341
00:38:13,000 --> 00:38:19,000
You know it's all about fast finality at the end of the day. It boils down to how fast you need your finality right.

342
00:38:19,000 --> 00:38:27,000
Someone has ideas actually good friend of mine. Dan Lynch is it's an up and coming Cosmos Tower.

343
00:38:27,000 --> 00:38:30,000
He is actually doing a lot of work for Osmosis now.

344
00:38:30,000 --> 00:38:42,000
But he had an idea for what was called but the idea was like we have a big challenge is bootstrapping and discovering RPC nodes and discovering state bootstrapping in general.

345
00:38:42,000 --> 00:38:44,000
It's a huge problem right.

346
00:38:44,000 --> 00:38:48,000
What's having a chain or a client as a matter of fact.

347
00:38:48,000 --> 00:38:55,000
And how do you know. Well first question you got to ask is how do you know what you're getting Israel or not.

348
00:38:55,000 --> 00:38:58,000
How do you trust that RPC server.

349
00:38:58,000 --> 00:39:04,000
But the only way to do that is like run by yourself from the code compound area area area.

350
00:39:04,000 --> 00:39:12,000
And I assume that GitHub is the official source code and it's not you know it doesn't have any exploits or not trying to steal your money right.

351
00:39:12,000 --> 00:39:21,000
So you're trusting GitHub again here and you're trusting that the canonical source code and you're trusting GitHub actually verified the authors are right.

352
00:39:21,000 --> 00:39:26,000
Anyway the the I think we need to we need to do much better.

353
00:39:26,000 --> 00:39:32,000
We need to do much better without GitHub involved necessarily right.

354
00:39:32,000 --> 00:39:51,000
And we need to do much better with bootstrapping and in a DAO idea seem to be very very attractive especially at DAO that is where the members are effectively validators right because they are most connected with with with managing multi managing multi or validating multiple

355
00:39:51,000 --> 00:40:08,000
chains. If they all come together and have some sort of like you know decentralized way to to sign off on these packages or bless the packages or bless the state and host the state and get some fees for it or something.

356
00:40:08,000 --> 00:40:10,000
It's not a bad idea actually.

357
00:40:10,000 --> 00:40:14,000
It could be a chain by itself or it could be a DAO on June.

358
00:40:14,000 --> 00:40:23,000
I think somebody's somebody's trying to build a get chain in cosmos right. I think Paul could choose sent round a thing.

359
00:40:23,000 --> 00:40:29,000
Paul could choose an excellent valid by the way I'm not sure if they validate a cache but one of the pop up in there.

360
00:40:29,000 --> 00:40:45,000
Yeah, one of the absolute best. They were they were talking to us the other day about I forget the name of it but there's a chain launching in cosmos sort of with some get integration hooks and I think they want to do some of that stuff in the future whether or

361
00:40:45,000 --> 00:40:48,000
not it will pan out exactly as intended.

362
00:40:48,000 --> 00:40:50,000
I don't know.

363
00:40:50,000 --> 00:40:53,000
So, some specific questions on a cash.

364
00:40:53,000 --> 00:40:57,000
I keep saying a cash is because I'm pronouncing it the English word I can't.

365
00:40:57,000 --> 00:41:02,000
I know showing them whatever anyway, culturally why unfortunately.

366
00:41:02,000 --> 00:41:07,000
So, for those that may be less familiar with a cash.

367
00:41:07,000 --> 00:41:13,000
What what drove you to start in what was like your original vision.

368
00:41:13,000 --> 00:41:22,000
So, bit of a backstory my been a programmer for a little 25 years, and I founded a company called Angel hack.

369
00:41:22,000 --> 00:41:34,000
Angel hack is the largest hackathon accelerator in the world hackathon based accelerator, but it was really we introduced hackathons to the world hackathons before Angel hack or a way underground concept.

370
00:41:34,000 --> 00:41:40,000
And I helped launch several developer tooling companies I mean my primary focus.

371
00:41:40,000 --> 00:41:45,000
Over the last I would say 10 years I've been building and shipping developer tooling.

372
00:41:45,000 --> 00:41:50,000
And one of the biggest company help launch was fire base.

373
00:41:50,000 --> 00:41:58,000
Fire base became absolute absolute darling in developer tooling and got acquired by Google.

374
00:41:58,000 --> 00:42:13,000
And so my passion was always developing tools that are hyperscalable without the friction, because a lot of times you have distributed systems with hyperscalability, but that extremely hard to use.

375
00:42:13,000 --> 00:42:19,000
And vice versa right systems that are super simple are not very scalable.

376
00:42:19,000 --> 00:42:25,000
Right so the idea was how do you really create infrastructure that's easy to use at the same time.

377
00:42:25,000 --> 00:42:28,000
You know, easy to scale right.

378
00:42:28,000 --> 00:42:30,000
It's still a problem to date.

379
00:42:30,000 --> 00:42:33,000
And we.

380
00:42:33,000 --> 00:42:39,000
And I stumbled this 2013, two years into Angel hack.

381
00:42:39,000 --> 00:42:48,000
I really wanted to solve the deployment problem, right like so deploying something from your workstation to the cloud.

382
00:42:48,000 --> 00:42:57,000
Was extremely, you know, difficult, especially if you're trying to do if you're under duress at a hack.

383
00:42:57,000 --> 00:42:58,000
Right.

384
00:42:58,000 --> 00:43:03,000
It's easy to like do that on a hero cool.

385
00:43:03,000 --> 00:43:06,000
Now you have an effort finders or so what not.

386
00:43:06,000 --> 00:43:09,000
But those systems generally can scale right.

387
00:43:09,000 --> 00:43:14,000
Heracle literally wouldn't scale like just ask you to go somewhere else.

388
00:43:14,000 --> 00:43:18,000
And if you try to use the Amazon, and it's just a nightmare, right.

389
00:43:18,000 --> 00:43:27,000
And you have to do like, you know, share for a puppet or one of these like nasty things, which are very unpleasant to use tools.

390
00:43:27,000 --> 00:43:30,000
And so I discovered this thing called containers.

391
00:43:30,000 --> 00:43:33,000
And I fell in love with them.

392
00:43:33,000 --> 00:43:42,000
And I was I was like, and containers back then this is right before Docker.

393
00:43:42,000 --> 00:43:44,000
Came along with containers really simple.

394
00:43:44,000 --> 00:43:49,000
And, and I was like working on how do you make these containers more fault tolerant.

395
00:43:49,000 --> 00:43:54,000
And then I discovered Kubernetes in 2014.

396
00:43:54,000 --> 00:44:02,000
This is way before Kubernetes, it's right, you know, very, very early and spoken for a conference and whatnot.

397
00:44:02,000 --> 00:44:09,000
I started contributing to the project quite a bit and found the company overclock labs to take Kubernetes to market.

398
00:44:09,000 --> 00:44:16,000
And I love Kubernetes because it was on its Docker containers in general would solve the parody problem.

399
00:44:16,000 --> 00:44:21,000
You build once you test it and you can take that container and deploy it on a server.

400
00:44:21,000 --> 00:44:33,000
You don't have to worry about dependencies, right, which is most of the time the, you know, the most of the time where the issues happen and back then mutable infrastructure was a thing.

401
00:44:33,000 --> 00:44:35,000
I want to immutable infrastructure.

402
00:44:35,000 --> 00:44:36,000
Right.

403
00:44:36,000 --> 00:44:41,000
So, and containers solve a lot of these problems.

404
00:44:41,000 --> 00:44:45,000
Kubernetes all the orchestration and for tolerance problems.

405
00:44:45,000 --> 00:44:55,000
And, and that's, that's where I saw the light to build a truly scalable, easy to use infrastructure and Kubernetes manual is not easy.

406
00:44:55,000 --> 00:44:58,000
I guess it's still not easy to use, but it was horrible back then.

407
00:44:58,000 --> 00:45:03,000
I mean, I know documentation is very, very early days.

408
00:45:03,000 --> 00:45:14,000
And we built overclock labs today, Kubernetes market, but also and do so in a, in a, in a product that was high performance, edge capable scheduling.

409
00:45:14,000 --> 00:45:15,000
Right.

410
00:45:15,000 --> 00:45:27,000
So you could literally use overclock labs product to schedule to turn any data center, mostly bare metal data centers into, into cloud enabled products.

411
00:45:27,000 --> 00:45:28,000
Right.

412
00:45:28,000 --> 00:45:30,000
So, and federated them.

413
00:45:30,000 --> 00:45:45,000
So you can literally create this beautiful layer and that you can deploy to let's say 10 different regions in like three minutes or four minutes using all running communities and communities didn't have multi cloud coordination capability.

414
00:45:45,000 --> 00:45:52,000
So we wrote the, the, the control plane for the, for the multi cloud to make Kubernetes multi cloud in 2015.

415
00:45:52,000 --> 00:45:58,000
And it worked amazingly well.

416
00:45:58,000 --> 00:46:08,000
And so that got me to talk to the community's conference and the, and I met a lot of incredible people.

417
00:46:08,000 --> 00:46:21,000
A lot of the code I wrote for that became helm and, you know, helm came around way later and they used a lot of the code I wrote to do this deployments right to do this, you know, package management style deployments.

418
00:46:21,000 --> 00:46:29,000
And when we were deploying this, we discovered most compute sitting in this data centers we're deploying to was unused.

419
00:46:29,000 --> 00:46:33,000
The number was somewhere like 90% of the compute was not used.

420
00:46:33,000 --> 00:46:34,000
Right.

421
00:46:34,000 --> 00:46:43,000
So, and well, then then it becomes very obvious to create a marketplace to unlock this compute and make it available.

422
00:46:43,000 --> 00:46:47,000
This is my new high performance compute status centers with their metal storage.

423
00:46:47,000 --> 00:46:48,000
Right.

424
00:46:48,000 --> 00:46:57,000
And, and, and, and bare metal is a lot more efficient than the hypervisor in terms of like resource allocation.

425
00:46:57,000 --> 00:47:08,000
And it's a lot more faster because you don't have the double kernel problem you're going directly to the box right all the good stuff and from a security standpoint all that all the good stuff.

426
00:47:08,000 --> 00:47:18,000
And, and so when we decided to really build a marketplace we wanted to make sure it's open, because we're open source people.

427
00:47:18,000 --> 00:47:28,000
If you've been working on open so we didn't want to create a close source sort of like company and big problem with open source is well, it's not sustainable right as a business.

428
00:47:28,000 --> 00:47:37,000
And how do you really create a network or a, you know, or a product that can set self sustain itself.

429
00:47:37,000 --> 00:47:50,000
And, and we were also like being a edge and evil product we had to create a, you know, sort of like central as system, right because I'm a big fan of optimistic schedulers.

430
00:47:50,000 --> 00:48:05,000
And a big thing about optimistic schedules versus pessimistic schedulers is pessimistic schedulers need a control plane that can that, you know, controls the nodes with optimistic is self, you know, propagating right.

431
00:48:05,000 --> 00:48:08,000
It's slower but actually more, more, more.

432
00:48:08,000 --> 00:48:21,000
I'll put my hand up and say, I didn't, I didn't fully grasp what's going on why why why does the optimistic schedule and need a centralized router.

433
00:48:21,000 --> 00:48:33,000
So a optimistic scheduler would try to optimistically schedule and then fail and then, you know, fail to succeed and doesn't require a lock.

434
00:48:33,000 --> 00:48:40,000
A lock has to be stored. Why, because when you need a lock that means that lock has to be maintained by a system.

435
00:48:40,000 --> 00:48:44,000
And that system could go down like in pessimistic schedulers.

436
00:48:44,000 --> 00:48:48,000
So we wanted a lock less system.

437
00:48:48,000 --> 00:48:49,000
Right.

438
00:48:49,000 --> 00:48:52,000
That's where the optimistic scheduler design really shines.

439
00:48:52,000 --> 00:48:59,000
And we kind of built a optimistic scheduler at overclock that was amazing for scheduling the edge nodes.

440
00:48:59,000 --> 00:49:05,000
Right. So and these nodes were like self coordinate without a central control plane in a peer to peer matter.

441
00:49:05,000 --> 00:49:11,000
And they would use me with it CD those most popular data structure, you know, that store.

442
00:49:11,000 --> 00:49:16,000
And we kind of wrote like a bit torn style replication mechanism.

443
00:49:16,000 --> 00:49:21,000
So we kind of had like a quasi blockchain and a mercury integrity check.

444
00:49:21,000 --> 00:49:22,000
Right.

445
00:49:22,000 --> 00:49:27,000
So yeah, I had like a quasi blockchain without even realizing we both blockchains.

446
00:49:27,000 --> 00:49:30,000
This is in 20, 15, 2016 time frame.

447
00:49:30,000 --> 00:49:36,000
And then we were like, well, I mean, we had a blockchain without our tokens to, you know, to any permission blockchain.

448
00:49:36,000 --> 00:49:37,000
Right.

449
00:49:37,000 --> 00:49:43,000
And then we realized, well, we might actually buy the bullet and go full blockchain.

450
00:49:43,000 --> 00:49:44,000
Instead of.

451
00:49:44,000 --> 00:49:55,000
I think leads us on to the next kind of question area that we kind of had for you, which was like, this sounds like the original vision was was very, very strictly kind of going in this direction.

452
00:49:55,000 --> 00:49:58,000
You're already going with Docker and Kubernetes and whatnot.

453
00:49:58,000 --> 00:50:01,000
And it's like, okay, well, but how can we make this sustainable?

454
00:50:01,000 --> 00:50:07,000
So it's a combination of the economic model and the fact that you're already kind of already building a blockchain.

455
00:50:07,000 --> 00:50:16,000
So then the next question we were going to kind of ask about, you know, your involvement in the cosmos, which is what made you build on cosmos in the first place?

456
00:50:16,000 --> 00:50:21,000
And I what what technical benefits did did that bring and what challenges.

457
00:50:21,000 --> 00:50:29,000
So, so the decision was like, well, we want to do this, do we do a one or do we do an Ethereum Ethereum was the only chain.

458
00:50:29,000 --> 00:50:33,000
We actually prototyped in the Ethereum on Ethereum, Crypto Kitties launch and crash out chain.

459
00:50:33,000 --> 00:50:36,000
And we're like, well, we're not, this is not going to work.

460
00:50:36,000 --> 00:50:37,000
Right.

461
00:50:37,000 --> 00:50:44,000
Ethereum like Crypto Kitties literally took down the chain for a few remember that for a while.

462
00:50:44,000 --> 00:50:53,000
And we were not even able to do anything and we decided to go L one. And the question was to build our own chain from scratch.

463
00:50:53,000 --> 00:50:59,000
Or if we can find some help to do that quicker and then somebody introduced me to Jay Kwan.

464
00:50:59,000 --> 00:51:01,000
And this is 2018.

465
00:51:01,000 --> 00:51:03,000
Early this 2018.

466
00:51:03,000 --> 00:51:08,000
And they were like, Hey, there's this library called Tenderman people are building which does the consensus for you.

467
00:51:08,000 --> 00:51:11,000
And everything else you build on top of it.

468
00:51:11,000 --> 00:51:14,000
And we liked it because number one was Golang.

469
00:51:14,000 --> 00:51:16,000
It's easy for us to go pick it up.

470
00:51:16,000 --> 00:51:19,000
And we were looking at the development on Tenderman.

471
00:51:19,000 --> 00:51:23,000
It was how to high cadence on GitHub is very important to us.

472
00:51:23,000 --> 00:51:36,000
And number three, and the fact that it's Golang gave us confidence that we can actually go build on something that's incomplete and actually, you know, contribute and build together if you have to.

473
00:51:36,000 --> 00:51:38,000
And the consensus.

474
00:51:38,000 --> 00:51:41,000
For those at the back, Tenderman is Golang.

475
00:51:41,000 --> 00:51:49,000
The Cosmos SDK is Golang Docker and a lot of the tooling around Docker containerization Kubernetes.

476
00:51:49,000 --> 00:52:00,000
These large scale, maybe inverted commas enterprise software suites for doing orchestration on a massive scale at like data center level.

477
00:52:00,000 --> 00:52:02,000
A lot of these tools are written in Golang.

478
00:52:02,000 --> 00:52:05,000
So obviously that Gregory talked about his background in that area.

479
00:52:05,000 --> 00:52:10,000
So that's why you when you see the goal, the kind of go light bulb go on there.

480
00:52:10,000 --> 00:52:17,000
That's why I've been doing go for since 2014, officially full time go.

481
00:52:17,000 --> 00:52:19,000
If you look at my GitHub, you'll see a lot of history.

482
00:52:19,000 --> 00:52:20,000
I've contributed.

483
00:52:20,000 --> 00:52:23,000
My libraries are used by Kubernetes.

484
00:52:23,000 --> 00:52:31,000
I was early contributed Terraform Docker just that era of 2013 2014 infrastructure software.

485
00:52:31,000 --> 00:52:34,000
That's that's my background, right.

486
00:52:34,000 --> 00:52:38,000
And go is very comfortable for us.

487
00:52:38,000 --> 00:52:44,000
We try to pick a language or a framework usually in the language that we're comfortable with so we can actually make contributions.

488
00:52:44,000 --> 00:52:49,000
And Tenderman wasn't stable for a long time.

489
00:52:49,000 --> 00:52:51,000
And it didn't have a module.

490
00:52:51,000 --> 00:52:54,000
So we ended up writing all the modules ourselves.

491
00:52:54,000 --> 00:53:01,000
And then Cosmos SDK came in 2019, I think.

492
00:53:01,000 --> 00:53:08,000
Cosmos was developing parallelly to our to our, you know, project.

493
00:53:08,000 --> 00:53:14,000
And then it we drew apart, right, because we had our own one we were developing.

494
00:53:14,000 --> 00:53:17,000
We didn't really, you know, Cosmos was not ready.

495
00:53:17,000 --> 00:53:25,000
So we ended up just and the lot of the modules we wrote, of course, in a small team, we're not going to be able to have the same level of contributions.

496
00:53:25,000 --> 00:53:28,000
The core ignite team did.

497
00:53:28,000 --> 00:53:37,000
And in 2019, we decided to bite the bullet and switch to Cosmos SDK and throw away a lot of our code.

498
00:53:37,000 --> 00:53:43,000
Adam, a co-founder did this passive PR somewhere.

499
00:53:43,000 --> 00:53:48,000
Somewhere in the GitHub's, but we deleted like 80,000 lines of code.

500
00:53:48,000 --> 00:53:55,000
Because that was the kind of like it was amazing because you know, it's like deleting code.

501
00:53:55,000 --> 00:54:01,000
And like the leading code means less attack surface.

502
00:54:01,000 --> 00:54:06,000
You're really, you're really exchanging code more than deleting, right?

503
00:54:06,000 --> 00:54:13,000
Something that's more reviewed, more PR review, that's more ideally more stable, ideally more maintained, right?

504
00:54:13,000 --> 00:54:19,000
Like so quote unquote, well maintained, right?

505
00:54:19,000 --> 00:54:25,000
But there's always we write the least amount of code as possible.

506
00:54:25,000 --> 00:54:27,000
That's my policy, right?

507
00:54:27,000 --> 00:54:30,000
Although we end up, we end up writing a lot of code.

508
00:54:30,000 --> 00:54:36,000
We have the saying that hey, months of coding can save hours of planning.

509
00:54:36,000 --> 00:54:37,000
Yeah, yeah.

510
00:54:37,000 --> 00:54:45,000
But anyway, you want to write less code and SDK significantly reduced our amount of code that we need to write.

511
00:54:45,000 --> 00:54:52,000
And now it's a thriving ecosystems amazing to see the modules that are developing and we wrote a few more modules we contributed to and whatnot.

512
00:54:52,000 --> 00:54:58,000
So this is a long winded history as to how we started and why we went down the rabbit hole.

513
00:54:58,000 --> 00:55:04,000
And believe me, when we were doing Cosmos, nobody, there are not too many teams that are doing Cosmos, very few actually.

514
00:55:04,000 --> 00:55:07,000
Yeah, yeah, it was one of the first projects, right?

515
00:55:07,000 --> 00:55:10,000
I would assume in terms of larger projects that came off the hub.

516
00:55:10,000 --> 00:55:13,000
Yeah, we're the first IBC chain, right?

517
00:55:13,000 --> 00:55:16,000
Like after the hub, the first IBC chain.

518
00:55:16,000 --> 00:55:24,000
So when Osmos is launched, they launched with Akash as the first token besides Osmo and Ion.

519
00:55:24,000 --> 00:55:29,000
Like pool number three and pool number four is Akash, right?

520
00:55:29,000 --> 00:55:32,000
So and Sunny is also a good friend.

521
00:55:32,000 --> 00:55:38,000
But but yeah, being early in the ecosystem, you could get a lot of help from the Cosmos or G's.

522
00:55:38,000 --> 00:55:54,000
Sunny was very, very helpful and we wrote, you know, I've read almost every paper of Cosmos and, you know, like, you know, a lot of the economics were influenced by Sunny's work and Jack was very helpful.

523
00:55:54,000 --> 00:56:00,000
Jack worked for Akash for like six months, I believe, helped us really, you know, get in there.

524
00:56:00,000 --> 00:56:16,000
And it's fun to be early because you get access to all these amazing people, you know, working like, you know, on the team for even like, what's the name of the most founder for their work for like a few weeks at Akash as well.

525
00:56:16,000 --> 00:56:20,000
So we had a lot of like Cosmos or G's working at Akash at some point.

526
00:56:20,000 --> 00:56:23,000
So they have like the biggest knowledge base, right?

527
00:56:23,000 --> 00:56:25,000
For the thing because they wrote it.

528
00:56:25,000 --> 00:56:29,000
So yeah, you probably have the best knowledge base.

529
00:56:29,000 --> 00:56:32,000
So that's the that's the history of Akash.

530
00:56:32,000 --> 00:56:35,000
Where do you see the project going, you know, in like five years?

531
00:56:35,000 --> 00:56:37,000
Like what's your vision now?

532
00:56:37,000 --> 00:56:43,000
Before we get into that, the fray, I sort of have an interim question between those two.

533
00:56:43,000 --> 00:56:46,000
And that is, you recently upgraded.

534
00:56:46,000 --> 00:56:54,000
And I didn't go into the test net and I'm not a validator on the Akash main net.

535
00:56:54,000 --> 00:57:02,000
I wasn't inside that, but I like resigned from that somewhat lately because I just don't have not been a main net validator there.

536
00:57:02,000 --> 00:57:05,000
I don't really have the additional time to put to that.

537
00:57:05,000 --> 00:57:11,000
But my question is, so it was a major upgrade this one, I believe.

538
00:57:11,000 --> 00:57:18,000
And I'm just curious as to what feature upgrades there was, like what additional functionality did this upgrade bring?

539
00:57:18,000 --> 00:57:21,000
Or was it just more of a background upgrade?

540
00:57:21,000 --> 00:57:24,000
No, this added storage.

541
00:57:24,000 --> 00:57:37,000
Well, so for context, cloud being able to compete or having complement or having feature match with Amazon is a long game.

542
00:57:37,000 --> 00:57:47,000
You're not when Akash launch, you know, we didn't plan on having all the features Amazon has on day one, right?

543
00:57:47,000 --> 00:57:52,000
It's a long roadmap for us to get there to even match all the features.

544
00:57:52,000 --> 00:58:03,000
So initially, when we launched the Akash platform, what a year ago, we wanted to perfect the container orchestration, fmr container orchestration.

545
00:58:03,000 --> 00:58:11,000
Because once you add state to it, you have a second variable that is extremely hard to debug to stabilize and whatnot.

546
00:58:11,000 --> 00:58:28,000
So we want to keep the features that small enough so that our teams can, you know, have focus and we have less things to worry about and less things to validate.

547
00:58:28,000 --> 00:58:40,000
So we put up our plan saying that we're going to do fmr containers without state, then we're going to do state, then we're going to add compute capability, this GPU is in bandwidth and whatnot.

548
00:58:40,000 --> 00:58:48,000
Then that will give us a foundation to actually build a real cloud on top of it, the true serverless cloud, the way we're going, right?

549
00:58:48,000 --> 00:59:07,000
What a true serverless cloud will look like essentially is, you know, writing serverless code, be it no code, and being able to literally orchestrate across different services that are, you know, offered to you in an open market and in a competitive manner.

550
00:59:07,000 --> 00:59:14,000
And, you know, an ecosystem where there are a lot of players and benefiting from each other, right?

551
00:59:14,000 --> 00:59:19,000
The idea is to remove a need for a centralized gatekeeper to all these technologies.

552
00:59:19,000 --> 00:59:24,000
Today, if you want access to any good technology, I mean, you got to be on the cloud, right?

553
00:59:24,000 --> 00:59:38,000
So where we are going is first solve the fundamentals and then we get into the actual services, right? I want to be able to get Postgres, manage Postgres on a cloud, right?

554
00:59:38,000 --> 00:59:44,000
And I would want that Postgres to come from the creator of Post, creators of Postgres, not from Amazon.

555
00:59:44,000 --> 00:59:49,000
Because Amazon is literally, is widely labeling it and adding a premium right?

556
00:59:49,000 --> 00:59:55,000
The creators don't make any money, right? The operator, the people that contribute to the network don't make any money.

557
00:59:55,000 --> 00:59:58,000
So our goal of Vision of Akash is to change that.

558
00:59:58,000 --> 01:00:09,000
So this Mainnet 3 is essentially the first step in us upgrading capabilities and adding elastic storage, essentially, right?

559
01:00:09,000 --> 01:00:10,000
Right.

560
01:00:10,000 --> 01:00:19,000
So now we have unbounded storage. You can attach how much ever that provider is capable of providing you in terms of storage.

561
01:00:19,000 --> 01:00:23,000
There is no limit. And this is network attached storage. Now we have NVME storage.

562
01:00:23,000 --> 01:00:27,000
You have all kinds of high performance storage that you can get access to.

563
01:00:27,000 --> 01:00:36,000
And we are seeing a lot of activity from Chia, Project Chia is one of these really good projects that is storage heavy, essentially.

564
01:00:36,000 --> 01:00:42,000
The consensus algorithm is storage based. So for them, they can get specialized storage on the current cloud today.

565
01:00:42,000 --> 01:00:45,000
They're able to come to Akash and get that high performance storage.

566
01:00:45,000 --> 01:00:48,000
They want something like a million IOSs on them, right?

567
01:00:48,000 --> 01:01:00,000
So Akash is opening up and in the future, if you look at what's happening with MyTea, which is an app that started by Sahel, who's a known founder, a lot of new technology.

568
01:01:00,000 --> 01:01:06,000
And for MyTea, they need their own data centers because they can get the performance on the current cloud.

569
01:01:06,000 --> 01:01:13,000
So the way Akash is going, Akash want to be in a position to serve these new founders with specialized requirements.

570
01:01:13,000 --> 01:01:28,000
So they should be able to put an RFP and be like, hey, anyone that can provide this kind of capacity, talk to me, I'll get you, come to Akash and I'll be able to consume you in a peer-to-peer manner.

571
01:01:28,000 --> 01:01:41,000
So really the goal of Akash is to open up the market to anybody and anyone that's capable of serving or providing compute to anyone that's capable of consuming compute.

572
01:01:41,000 --> 01:01:47,000
And serving compute is not a complicated thing. Running a data center is not as complicated as writing software.

573
01:01:47,000 --> 01:01:49,000
I mean, it is being standardized.

574
01:01:49,000 --> 01:01:51,000
That's a pretty hot take.

575
01:01:51,000 --> 01:01:53,000
I mean, it's got a lot of work.

576
01:01:53,000 --> 01:02:03,000
I mean, writing software is a lot more heavier than maintaining, you know, changing your disks, literally.

577
01:02:03,000 --> 01:02:06,000
Like the 8.4 million data centers in the world.

578
01:02:06,000 --> 01:02:08,000
Right.

579
01:02:08,000 --> 01:02:13,000
And again, the modern software is designed to be fault tolerant.

580
01:02:13,000 --> 01:02:16,000
It's designed to be interpret tolerant.

581
01:02:16,000 --> 01:02:18,000
It should be if it's not.

582
01:02:18,000 --> 01:02:26,000
Yeah, really interesting to hear you talk about, you know, sort of serverless technologies and make a kind of on-demand compute element.

583
01:02:26,000 --> 01:02:46,000
Because obviously, so pre coming to work in crypto, a lot of what I did as an independent consultant and we did as a company was trying to get enterprise-sized companies to ditch running metal.

584
01:02:46,000 --> 01:02:51,000
And if they weren't in the cloud, moving to cloud, and if they were already in the cloud, move to serverless.

585
01:02:51,000 --> 01:02:57,000
And if, you know, sometimes jumping the whole way from bare metal to serverless, right?

586
01:02:57,000 --> 01:03:12,000
Because fundamentally, for most kind of consumer grade applications where it's, you know, e-commerce or some kind of thing that can be served headlessly, some kind of content website, some other bits and bobs, maybe a next JS site,

587
01:03:12,000 --> 01:03:19,000
maybe, you know, some React or Gatsby or, you know, whichever kind of, whichever front-end it is.

588
01:03:19,000 --> 01:03:23,000
It's not particularly heavyweight and the API requirements on it aren't particularly heavyweight.

589
01:03:23,000 --> 01:03:30,000
And you don't need lots of big heavy servers sitting around, you know, to serve a few web requests with a TypeScript API, do you know what I mean?

590
01:03:30,000 --> 01:03:32,000
Yeah.

591
01:03:32,000 --> 01:03:46,000
And so, you know, serverless is a great paradigm for that because even on AWS, you can, you know, have 80 serverless functions in a test environment, a prod environment and have like quite significant traffic and pay like $200 a month.

592
01:03:46,000 --> 01:03:52,000
You know, I've been in that scenario where we had a large team sort of running this project and people were like, oh, what is this going to cost us?

593
01:03:52,000 --> 01:03:53,000
What is this going to cost us?

594
01:03:53,000 --> 01:04:03,000
And I was like, I guarantee you our monthly cost for the serverless function invocation is going to be less than one day of one developer's time.

595
01:04:03,000 --> 01:04:04,000
And I was right by a good margin as well.

596
01:04:04,000 --> 01:04:06,000
But you look at that and you go, well, hang on.

597
01:04:06,000 --> 01:04:11,000
Like you said earlier, there's a lot of unused compute sitting around.

598
01:04:11,000 --> 01:04:30,000
What if you can take that model of on demand, pay paper, micro second or whatever it is that AWS and Google do, you know, for their on demand compute for serverless functions and then also be spending at the rate of unserved effectively compute.

599
01:04:30,000 --> 01:04:32,000
It could get an order of magnitude cheaper, right?

600
01:04:32,000 --> 01:04:41,000
And for somebody that's kind of like a serverless advocate, where I'm like, I only really want to pay for what I'm using. And I think that that's what people should do.

601
01:04:41,000 --> 01:04:43,000
And it's an obvious place to have a market.

602
01:04:43,000 --> 01:04:45,000
That's, I think, a really exciting idea.

603
01:04:45,000 --> 01:04:48,000
Like if you can get it to that level of granularity, right?

604
01:04:48,000 --> 01:04:49,000
That's where we're going.

605
01:04:49,000 --> 01:04:51,000
That's where that's the direction we're heading towards.

606
01:04:51,000 --> 01:04:57,000
And I think like once there's so much opportunity and so many problems right now to solve.

607
01:04:57,000 --> 01:05:08,000
I mean, just before we can get to this true no code serverless, just write a function and run it, plug in in a low latency environment, you know, you know, run all that stuff.

608
01:05:08,000 --> 01:05:09,000
We need to solve bandwidth.

609
01:05:09,000 --> 01:05:10,000
We need to solve storage.

610
01:05:10,000 --> 01:05:17,000
We need to solve like, like high percent high performance compute HPC computing like with GPs and whatnot.

611
01:05:17,000 --> 01:05:22,000
Unless we can solve these problems in a high distributed and decentralized way.

612
01:05:22,000 --> 01:05:26,000
Then we're going to see the true benefits of serverless come to fruition, right?

613
01:05:26,000 --> 01:05:36,000
Like today in pricing, cloud has this course screen pricing model, which results in partial usage risk.

614
01:05:36,000 --> 01:05:42,000
So like you pay essentially one hour, one hour is a minimum.

615
01:05:42,000 --> 01:05:48,000
And get a standard tier model run the serverless pricing on the program with traditional instance pricing on demand pricing.

616
01:05:48,000 --> 01:05:51,000
You get this tier you pick from.

617
01:05:51,000 --> 01:05:55,000
And you and you have this time slots you pay with.

618
01:05:55,000 --> 01:05:59,000
And that's the course screen pricing model.

619
01:05:59,000 --> 01:06:05,000
And we all sort of suboptimal for running anything like a function.

620
01:06:05,000 --> 01:06:06,000
Right.

621
01:06:06,000 --> 01:06:15,000
And the pricing is not and most of the time and almost all all the time, the requirements for a function vary.

622
01:06:15,000 --> 01:06:17,000
It's not going to be like 16 gigs of memory.

623
01:06:17,000 --> 01:06:20,000
It's going to be probably like, I don't know, 12.1 gigs of memory.

624
01:06:20,000 --> 01:06:21,000
Yeah.

625
01:06:21,000 --> 01:06:23,000
It's so fine green.

626
01:06:23,000 --> 01:06:29,000
And it's very important for the future of the cloud to have a fine green model and the cost rewards you find a model.

627
01:06:29,000 --> 01:06:31,000
Provide it because very challenging.

628
01:06:31,000 --> 01:06:39,000
How do you price this on the on the backing and a lot of the work we're doing right now is like, how do you optimize, you know, your profit and loss.

629
01:06:39,000 --> 01:06:51,000
But so we have taken fundamentally solid primitives with pricing, solid primitives with distributed design that will bake into that will come of relevance.

630
01:06:51,000 --> 01:06:59,000
Once we have a serverless working because serverless will translate directly one to one down to instant pricing with our fine green model.

631
01:06:59,000 --> 01:07:02,000
And then you're going to see a lot cheaper serverless function.

632
01:07:02,000 --> 01:07:04,000
So this is not affordable today on the cloud.

633
01:07:04,000 --> 01:07:05,000
Right.

634
01:07:05,000 --> 01:07:06,000
Yeah, it's okay.

635
01:07:06,000 --> 01:07:09,000
If you're running something with a lower scale, but try to scale that up and it's scary.

636
01:07:09,000 --> 01:07:12,000
Some of these bills are especially in Amazon.

637
01:07:12,000 --> 01:07:13,000
Right.

638
01:07:13,000 --> 01:07:14,000
The Fright knows about that.

639
01:07:14,000 --> 01:07:15,000
It's scary.

640
01:07:15,000 --> 01:07:16,000
Right.

641
01:07:16,000 --> 01:07:17,000
More do you scale.

642
01:07:17,000 --> 01:07:20,000
I mean, I'm talking about like, I mean, I think, I think my, my, my valid.

643
01:07:20,000 --> 01:07:29,000
My, my validator bills are validator bills as a company are because we are running on AWS.

644
01:07:29,000 --> 01:07:47,000
It's the environment that we're most familiar with are substantially more expensive than the large scale serverless deployment with, you know, Aurora RDS in three AZs fully backed up.

645
01:07:47,000 --> 01:07:59,000
You know, we are talking deployments where everything's terraformed, everything's automated, do maybe a minute, two minutes tops of DBIops per year, that kind of level of automation.

646
01:07:59,000 --> 01:08:10,000
And, you know, that's that entire setup, I think, even on a per month per year basis was probably RDS was the only thing that was particularly expensive in that model.

647
01:08:10,000 --> 01:08:25,000
And serverless execution, like I say, we were paying under $400 a month, but we very heavily optimized and very, very heavily knew what the billing pattern was.

648
01:08:25,000 --> 01:08:26,000
So free.

649
01:08:26,000 --> 01:08:37,000
So like, it's an open secret of a CNAWS that you get more, I can't remember which slider you get, whether it's for the CPU or for the RAM, but if you maxed it out, it makes it gives you, yeah, it gives you more virtual cores.

650
01:08:37,000 --> 01:08:41,000
So it makes it go quicker depending on what language you're using.

651
01:08:41,000 --> 01:08:46,000
Yeah, so sometimes it's better to bash the slider all the way and it actually executes quicker and cost you less.

652
01:08:46,000 --> 01:08:52,000
So we kind of obviously optimized, we tried a few different settings to work out which one was the best one.

653
01:08:52,000 --> 01:08:57,000
So, I mean, you're right there, there's obviously an angle for optimization, right?

654
01:08:57,000 --> 01:09:07,000
But yeah, it's exciting to think of like how much more accessible that could get and especially, you know, again, how much better the tool that you get around it.

655
01:09:07,000 --> 01:09:18,000
So we're at about an hour and 10, which is when we usually start thinking about putting the fray to bed, given our time zone here.

656
01:09:18,000 --> 01:09:26,000
So we've had a couple of questions for you, Greg, which I like to pick up on before we start wrapping up.

657
01:09:26,000 --> 01:09:28,000
I have a lot more questions.

658
01:09:28,000 --> 01:09:33,000
We, well, yes, I was going to do the viewer. Do you want to do the viewer ones now?

659
01:09:33,000 --> 01:09:35,000
It's only 10.

660
01:09:35,000 --> 01:09:38,000
Well, from, so, well, do we put those in the spreadsheet?

661
01:09:38,000 --> 01:09:48,000
But in any case, so, Todd is, and this is a brief answer to this one, I know, Todd has asked if Akash has a container registry.

662
01:09:48,000 --> 01:09:59,000
There is one in the works called Open Registry, it's funded by the Akash Foundation, the other CAB actually community awards board.

663
01:09:59,000 --> 01:10:04,000
It's ready to go or very close to completion.

664
01:10:04,000 --> 01:10:09,000
And the idea is to have the Open Registry, and it's Akash Optimist running Akash.

665
01:10:09,000 --> 01:10:19,000
So to use that as official container registry for Akash, and right now we had it still depend on the Docker Hub or whatever container registry as external to Akash.

666
01:10:19,000 --> 01:10:23,000
But yes, there is one in the works and it's almost complete.

667
01:10:23,000 --> 01:10:33,000
Okay, that's pretty awesome because I think at the moment, like most people just use GitHub container registry and Docker Hub, right?

668
01:10:33,000 --> 01:10:43,000
So one thing that I know that people have been asking for for a long time is HTTPS capabilities.

669
01:10:43,000 --> 01:10:51,000
I think most people at the moment just use Cloudflare to proxy to the servers.

670
01:10:51,000 --> 01:11:01,000
So is that something that is sort of on the back burner because of the work around with Cloudflare or is that something that's actively been worked on?

671
01:11:01,000 --> 01:11:05,000
It's on the roadmap. It's actually being worked on right now.

672
01:11:05,000 --> 01:11:13,000
Is it something that's a, is there much higher priorities than that or is that like one of the higher priorities?

673
01:11:13,000 --> 01:11:20,000
It's a think schedule for either next quarter, if I have to.

674
01:11:20,000 --> 01:11:21,000
Okay.

675
01:11:21,000 --> 01:11:23,000
It's a priority.

676
01:11:23,000 --> 01:11:32,000
But definitely on the lower set of priority because people use Cloudflare and Cloudflare gives you a lot more functionality.

677
01:11:32,000 --> 01:11:41,000
functionality beyond just HTTPS, right? It gives you DDoS protection, which is a pretty helpful feature.

678
01:11:41,000 --> 01:11:48,000
But again, but it introduces centralization, right? So we are aware of what's going on.

679
01:11:48,000 --> 01:11:59,000
But yeah, it's in the works. Hopefully they'll be ready in another three months, but we do have GPUs as a top tier priority, right?

680
01:11:59,000 --> 01:12:07,000
And when you talk about like running validators, Minakash costs 90% cheaper now than Amazon roughly for the high end boxes, right?

681
01:12:07,000 --> 01:12:09,000
80 to 90% cheaper.

682
01:12:09,000 --> 01:12:10,000
That's crazy.

683
01:12:10,000 --> 01:12:19,000
And it's also something called Omnibus, Cosmos Omnibus that gives you a lot of fault tolerance scripts and whatnot.

684
01:12:19,000 --> 01:12:28,000
Mind your culture is very, very self-hosted model and doesn't have all the bills and vessels that you get with Amazon, especially fault tolerance and like auto scaling or whatnot.

685
01:12:28,000 --> 01:12:33,000
So you have to end up writing a lot of these things yourself.

686
01:12:33,000 --> 01:12:42,000
So like if you want auto scale, you monitor on chain, if you're monitoring performance and because offline, maybe you have a smart contract in Juneau that will do it.

687
01:12:42,000 --> 01:12:52,000
So there's still a lot of work that needs to be done to host, but the advantage you get is like a very fine-grained pricing model that's 80% cheaper.

688
01:12:52,000 --> 01:12:55,000
And soon enough, we're going to talk about Juneau.

689
01:12:55,000 --> 01:13:04,000
We're going to, in the next upgrade, which is much sooner, within the next month at least, we're going to enable interchange accounts.

690
01:13:04,000 --> 01:13:10,000
So with that, you'll be able to use your Juneau account to deploy directly on Akash.

691
01:13:10,000 --> 01:13:13,000
You don't need an Akash wallet.

692
01:13:13,000 --> 01:13:17,000
And that Juneau account will own the deployment.

693
01:13:17,000 --> 01:13:25,000
It's great when you have a DAO, right, because DAO can own the web app or whatever you're using.

694
01:13:25,000 --> 01:13:33,000
Say if you're on a matrix server with element front end to have your own Slack that's completely sovereign to you and fully decentralized that you own completely.

695
01:13:33,000 --> 01:13:37,000
And you don't want to deal with a credit card or put it on Amazon.

696
01:13:37,000 --> 01:13:44,000
You can use your Juneau account, fund the deployment using your Juneau account directly on Akash.

697
01:13:44,000 --> 01:13:48,000
We're actually waiting for that functionality.

698
01:13:48,000 --> 01:14:06,000
So there's a project that I'm working on with some other very smart people called Howl, which because of my pronunciation, H-O-W-L, Howl, not Howl9000 is what a lot of people think.

699
01:14:06,000 --> 01:14:14,000
But it's sort of governed by a main DAO. And that DAO is the goal.

700
01:14:14,000 --> 01:14:24,000
We're kind of hoping that that feature will be ready for us to kick the tires of it to get the HowlDAO.

701
01:14:24,000 --> 01:14:34,000
HowlDAO with the Howl to actually host itself on Akash, if you like, and pay for its own hosting out of its development fund.

702
01:14:34,000 --> 01:14:39,000
Because it's only the main front end is only sort of React.

703
01:14:39,000 --> 01:14:46,000
There's not a hugely expensive service to run. So if we can work out a way of putting on Akash, it's going to cost even less.

704
01:14:46,000 --> 01:14:49,000
And it can sort of pay for itself and support itself.

705
01:14:49,000 --> 01:14:51,000
And that is kind of really interesting to us.

706
01:14:51,000 --> 01:14:59,000
So there's an idea of making the actual separating out the running of the service from the development of it.

707
01:14:59,000 --> 01:15:04,000
Maybe all the way back to what we were talking about with the Validator Set and Core Teams and how much trust you have.

708
01:15:04,000 --> 01:15:12,000
As soon as you give the community the ability or the DAO the ability to run its own code, it can kind of turn around and say,

709
01:15:12,000 --> 01:15:15,000
well, we don't like the direction developers have taken.

710
01:15:15,000 --> 01:15:18,000
Might take it a different direction actually, lads.

711
01:15:18,000 --> 01:15:22,000
And maybe we're rugging ourselves by looking at doing that.

712
01:15:22,000 --> 01:15:26,000
But I think it's like an interesting development of how you can do things, right?

713
01:15:26,000 --> 01:15:32,000
Yeah, it's great, right? I mean, like, it's, yeah, I mean, I think that should be the way to go.

714
01:15:32,000 --> 01:15:35,000
I mean, you should, your member should control your code.

715
01:15:35,000 --> 01:15:39,000
And what goes on to the, in that should be voted upon using governance.

716
01:15:39,000 --> 01:15:42,000
That's, yeah, I think that's really exciting.

717
01:15:42,000 --> 01:15:49,000
And I hope that the DAO DAO tooling advances to a point where you can make deployments through voting.

718
01:15:49,000 --> 01:15:51,000
That'll be pretty awesome.

719
01:15:51,000 --> 01:15:53,000
Yeah, I think that's being discussed right now.

720
01:15:53,000 --> 01:15:58,000
There's a PR somewhere with, with,

721
01:15:58,000 --> 01:15:59,000
Yeah, yeah.

722
01:15:59,000 --> 01:16:04,000
I think interchange accounts is part of the IBC 3.1, is that?

723
01:16:04,000 --> 01:16:06,000
Don't remember correctly, but.

724
01:16:06,000 --> 01:16:14,000
Yeah, so it has to go, I think from Juno's point of view, the interchange accounts is dependent on an upstream.

725
01:16:14,000 --> 01:16:17,000
It going into, fix going into Wasm.

726
01:16:17,000 --> 01:16:19,000
Sorry, not Wasm.

727
01:16:19,000 --> 01:16:23,000
So X, X Wasm, the Cosmos module.

728
01:16:23,000 --> 01:16:26,000
But I think that's going to happen pretty soon.

729
01:16:26,000 --> 01:16:29,000
Like I gather Ethan and Simon were talking about it.

730
01:16:29,000 --> 01:16:34,000
So it is soon to be M.

731
01:16:34,000 --> 01:16:43,000
So Greg, you did mention Cosmos Omnibus, which is mostly maintained, I think by Tom.

732
01:16:43,000 --> 01:16:44,000
Tom, you on?

733
01:16:44,000 --> 01:16:45,000
Yeah.

734
01:16:45,000 --> 01:16:46,000
Yeah.

735
01:16:46,000 --> 01:16:55,000
So thank you everyone, make sure you delegate to him on, certainly on a cash and the other chains on his, lately been contributing a lot and starting up.

736
01:16:55,000 --> 01:17:01,000
Validators, but he's been working on Cosmos Omnibus for a very long time.

737
01:17:01,000 --> 01:17:05,000
And I don't know how he keeps so much of that up to date.

738
01:17:05,000 --> 01:17:10,000
He's all over most chains and there's a couple of other contributors in there as well.

739
01:17:10,000 --> 01:17:12,000
I'm also becoming a Dow.

740
01:17:12,000 --> 01:17:25,000
I think Omnibus has a sort of like the best scenario because they maintain the snapshots, they maintain the code, they maintain the scalability and fault tolerance scripts.

741
01:17:25,000 --> 01:17:26,000
They maintain all that.

742
01:17:26,000 --> 01:17:30,000
I think that should be like, that could be a good source of truth.

743
01:17:30,000 --> 01:17:33,000
So I think one of the problems at the LB, awesome.

744
01:17:33,000 --> 01:17:36,000
Are you saying they're already maintaining snapshots?

745
01:17:36,000 --> 01:17:37,000
They are.

746
01:17:37,000 --> 01:17:38,000
Okay.

747
01:17:38,000 --> 01:17:39,000
That's awesome.

748
01:17:39,000 --> 01:17:42,000
I think that's awesome.

749
01:17:42,000 --> 01:17:49,000
I know that the storage was always a problem there.

750
01:17:49,000 --> 01:17:59,000
So now that there's persistent storage now, right, then that might be bowed a lot better now for management of those nodes.

751
01:17:59,000 --> 01:18:01,000
I'm not sure if there's an integration there yet.

752
01:18:01,000 --> 01:18:02,000
But certainly.

753
01:18:02,000 --> 01:18:04,000
There's also a discussion to run.

754
01:18:04,000 --> 01:18:11,000
So I think we're using either storage or for your archives or storage or or SIA.

755
01:18:11,000 --> 01:18:21,000
There's even a discussion to run a local gateway servers in local Akash providers.

756
01:18:21,000 --> 01:18:23,000
We're running the gateway servers for these storage network.

757
01:18:23,000 --> 01:18:28,000
So your downloads fees are going to be really, really low.

758
01:18:28,000 --> 01:18:31,000
I mean, high, high, and show it.

759
01:18:31,000 --> 01:18:36,000
So historically, the keys were kept on S3.

760
01:18:36,000 --> 01:18:42,000
But these days, I think leaning towards you can still use that same.

761
01:18:42,000 --> 01:19:00,000
Well, the same contain like the same code base, I think with Cosmos omnibus with very little variation to that to actually use like a like a soft sign from an external server like TM KMS, which sort of eliminates your

762
01:19:00,000 --> 01:19:06,000
key storage issues for that, which I think is worth a look.

763
01:19:06,000 --> 01:19:15,000
But one one other question I had outside of that was it's quite difficult to manage your deployment in terms of updating it.

764
01:19:15,000 --> 01:19:24,000
So one thing that we do at the moment is on Juno, we've deployed the Juno network dot IO website on Akash.

765
01:19:24,000 --> 01:19:41,000
And every time so it's quite fluid that website in updates for the ecosystem page where various projects add their information and then we update the container and then update a cache.

766
01:19:41,000 --> 01:19:55,000
So there's no way that I know of at the moment on Akash and I'm a cache guru to have a system where you can automatically update from a GitHub commit to update your instance.

767
01:19:55,000 --> 01:20:07,000
So basically now at the moment the way I've got it set up is if you make a commit or when you tag the same as on Cosmos omnibus, it will create a new container.

768
01:20:07,000 --> 01:20:13,000
And then you have to manually go and like redeploy the instance.

769
01:20:13,000 --> 01:20:21,000
Has anyone worked on like some sort of update from commit situation for React apps?

770
01:20:21,000 --> 01:20:24,000
We do that internally for Akash website.

771
01:20:24,000 --> 01:20:29,000
We actually have a get action that will handle the whole workflow for you.

772
01:20:29,000 --> 01:20:32,000
We don't even touch the deployments, but we're going to open source that way.

773
01:20:32,000 --> 01:20:40,000
There's some amount of cleanup that's happening and will open source will get enable deployments onto Akash.

774
01:20:40,000 --> 01:20:44,000
The two sets of tools that you will be looking at.

775
01:20:44,000 --> 01:20:45,000
One is a build time.

776
01:20:45,000 --> 01:20:47,000
Another one is deploy time.

777
01:20:47,000 --> 01:20:53,000
So the build time is building the container and deploy time is taking the container and deploying it.

778
01:20:53,000 --> 01:20:57,000
So for build time, if you have a Docker file, great.

779
01:20:57,000 --> 01:21:00,000
Use a Docker file for your build.

780
01:21:00,000 --> 01:21:04,000
If not, we can use something called build packs.io.

781
01:21:04,000 --> 01:21:12,000
These are literally Herico build packs that are community managed and that are amazing.

782
01:21:12,000 --> 01:21:17,000
It gives you Herico like capability, like get pushed like animals or the dependencies and all that good stuff right for you.

783
01:21:17,000 --> 01:21:24,000
So we have a fresh project that you haven't built a Docker container yet and building container is a nightmare.

784
01:21:24,000 --> 01:21:26,000
Build packs is a great way to do it.

785
01:21:26,000 --> 01:21:37,000
I have lots of examples on how to use build packs and do the deploy time using a GitHub action to take your container.

786
01:21:37,000 --> 01:21:46,000
Or you can use a GitHub action to invoke the build phase as well.

787
01:21:46,000 --> 01:21:56,000
But effectively, creating a sdl file around the container with that particular version, updating it completely on the background using GitHub actions.

788
01:21:56,000 --> 01:22:04,000
We have that in the open source site very soon to deal with a lot of these issues because I agree 100% with you.

789
01:22:04,000 --> 01:22:06,000
That's exciting to me.

790
01:22:06,000 --> 01:22:14,000
I'm not sure if any of these other guys are actually working with the cache deployments, but that really excites me.

791
01:22:14,000 --> 01:22:23,000
That's one of my biggest issues with using a cache is if you have a lot of deployments keeping them up to date with,

792
01:22:23,000 --> 01:22:30,000
especially when you work with a decentralized team like we do on Juno,

793
01:22:30,000 --> 01:22:39,000
if they go and you either have to get set up an alert that someone's tagged something new or go in and look at the commits and see what they've done

794
01:22:39,000 --> 01:22:48,000
and then tag something new when you think it's far enough along and then go and do the update.

795
01:22:48,000 --> 01:22:53,000
I use the cache analytics which makes it a lot easier than using ZLi.

796
01:22:53,000 --> 01:22:55,000
Yeah, cache analytics is awesome.

797
01:22:55,000 --> 01:22:59,000
We're also working collaboration features.

798
01:22:59,000 --> 01:23:06,000
A lot of new features that are coming in, authentication, using a cache login essentially and effective login mechanism.

799
01:23:06,000 --> 01:23:09,000
That'll open up collaboration.

800
01:23:09,000 --> 01:23:17,000
You can effectively build apps on top of a cache using a cache as an identity provider.

801
01:23:17,000 --> 01:23:24,000
You can add collaboration and you can have an external state on top of a cache, something similar to Terraform,

802
01:23:24,000 --> 01:23:27,000
where you can have metadata for applications.

803
01:23:27,000 --> 01:23:31,000
Metadata is important because it's very hard to identify applications right now.

804
01:23:31,000 --> 01:23:35,000
You just get a big long string.

805
01:23:35,000 --> 01:23:39,000
Challenging decentralized systems is just an experience.

806
01:23:39,000 --> 01:23:45,000
It's just going to be so decentralized native that you have to rethink on how these applications work.

807
01:23:45,000 --> 01:23:50,000
I'll be making a video announcement soon about our new head of product.

808
01:23:50,000 --> 01:23:56,000
He was leading Terraform at HashiCore.

809
01:23:56,000 --> 01:24:05,000
We're bringing a lot of former Cloud people into the mix, people that actually built Terraform.

810
01:24:05,000 --> 01:24:12,000
One of the beautiful things about Terraform I liked was it has great automation, but I hated state synchronization.

811
01:24:12,000 --> 01:24:19,000
I believe they have like cloud state synchronization now, but back in the day we used to manually ship the state if you wanted to collaborate.

812
01:24:19,000 --> 01:24:28,000
I think the cloud product has done quite a bit and we're going to take a lot of lessons from the Terraform cloud to add collaboration

813
01:24:28,000 --> 01:24:37,000
in a fully decentralized way, and great for teams that are collaborating across the pond and whatnot.

814
01:24:37,000 --> 01:24:48,000
There's a lot of deployment side of things from a user experience side of things that we're focusing heavily on.

815
01:24:48,000 --> 01:24:57,000
While we build the capabilities on the side, our latest hires are all focused on making the product simple.

816
01:24:57,000 --> 01:25:03,000
I hate to use the command line. There are so many steps right now.

817
01:25:03,000 --> 01:25:07,000
There are literally 14 steps to deploy something on a cache.

818
01:25:07,000 --> 01:25:16,000
If you do it manually, it's a lot of work to like, and then maintain it as well, and you have to keep a lot of information to be able to maintain your deployments.

819
01:25:16,000 --> 01:25:29,000
The metadata is hard. I literally have like a make file that I had on my command center, and the make file also has this gigantic state file.

820
01:25:29,000 --> 01:25:38,000
My make files look like it's embarrassing to open source my make files because they're just crying right now.

821
01:25:38,000 --> 01:25:46,000
It's just a flat file system to do metadata and whatnot, but we learned a lot.

822
01:25:46,000 --> 01:26:04,000
It was important for us to not work on workflows and really focus on functionality and stabilizing functionality because if we dictate a workflow, then we limit ourselves in terms of creativity.

823
01:26:04,000 --> 01:26:10,000
We're not going to let people come up with, right? Like, Akashlytics wouldn't exist if we made deployments super simple.

824
01:26:10,000 --> 01:26:20,000
And in terms of Akashlytics, it has done a phenomenal job, much better job than what we could, with almost no resources.

825
01:26:20,000 --> 01:26:29,000
And that's the kind of commitment we want people to have towards building Akash, and now they're actually doing really well.

826
01:26:29,000 --> 01:26:42,000
And we learned a lot from them, right? That's how to deploy. And of course, we're hiring people to make things simpler, but simplicity, removing friction is a large, large part of our work right now.

827
01:26:42,000 --> 01:26:53,000
Even with a new command line that we will be coming up with, a deploy command line that's outside of the actual command line that makes, that has opinions, right?

828
01:26:53,000 --> 01:27:00,000
Because just to provide things, you've got to opinion. Yeah. Well, that's all pretty exciting.

829
01:27:00,000 --> 01:27:05,000
And so it sounds like you're still, you know, there's a long roadmap and we're moving through it.

830
01:27:05,000 --> 01:27:17,000
I think it's going to, as things get easier and more inclusive in features, I think the like take up of users will come along with it.

831
01:27:17,000 --> 01:27:26,000
Once, I think, yeah, for a lot of people, I think once it becomes just easy to use, it'll become prolific.

832
01:27:26,000 --> 01:27:34,000
If people can just like, you know, here's the thing, click, here's the GitHub repo, click the button, it's done, it's up.

833
01:27:34,000 --> 01:27:41,000
And, you know, we want five instance instances with a load balance area, just go, doot, doot, doot, all over.

834
01:27:41,000 --> 01:27:49,000
And like, and a good command line experience, I think good command line experience is so integral to high quality developers.

835
01:27:49,000 --> 01:27:55,000
Web UI is great, but I think web UI, for me, especially, it's a secondary in extension, right?

836
01:27:55,000 --> 01:28:00,000
I use web UI to get a bigger picture, but I prefer command line always.

837
01:28:00,000 --> 01:28:10,000
Web UI requires like a certain amount of centralization as well in the workflow.

838
01:28:10,000 --> 01:28:13,000
Oh, I bought some GUI, like a cashlet is a GUI, right?

839
01:28:13,000 --> 01:28:19,000
Yeah, but it's, you download it to your thing, it's not dependent on like some web server somewhere.

840
01:28:19,000 --> 01:28:23,000
It's like a local, but is this not a web deployment of that, is it?

841
01:28:23,000 --> 01:28:26,000
It's a web-based, it's an electronic app, basically.

842
01:28:26,000 --> 01:28:27,000
Okay, right.

843
01:28:27,000 --> 01:28:30,000
But the point is a GUI, right? Forget that.

844
01:28:30,000 --> 01:28:33,000
A GUI is definitely a complementary, right?

845
01:28:33,000 --> 01:28:41,000
You'll be able to do, like, I think you should be able to do things very quickly using command line

846
01:28:41,000 --> 01:28:46,000
and you got to use a GUI if you need to see a bigger or compare things, right?

847
01:28:46,000 --> 01:28:49,000
Like logs, for example.

848
01:28:49,000 --> 01:28:52,000
I'm a big grep person, right?

849
01:28:52,000 --> 01:28:56,000
I want my greps, I want my filtering capabilities, like right on the command line, right?

850
01:28:56,000 --> 01:29:02,000
But if I do want to share the logs, I should be able to do that using a GUI, right?

851
01:29:02,000 --> 01:29:05,000
Or if I want to collaborate with other people, I want to be able to use GUI.

852
01:29:05,000 --> 01:29:11,000
Maybe I just want to deploy once and just click a button I want to use a GUI, but, you know, things like that, right?

853
01:29:11,000 --> 01:29:14,000
But if I want to do anything advanced, I'd rather prefer a command line.

854
01:29:14,000 --> 01:29:17,000
A command line, a good command line because it's part of the workflow, right?

855
01:29:17,000 --> 01:29:20,000
Like you're developing, I use Vim, my workflow.

856
01:29:20,000 --> 01:29:27,000
I use Vim to write code and I should be able to exit that Vim or stay in that context and deploy immediately from that Vim, right?

857
01:29:27,000 --> 01:29:34,000
Go to a website and open that because that's a whole different experience and a whole different context you're switching.

858
01:29:34,000 --> 01:29:38,000
Well, it sounds like your hire from Terraform is in the same similar mindset, right?

859
01:29:38,000 --> 01:29:41,000
Terraform absolutely follows that same type of structure where?

860
01:29:41,000 --> 01:29:42,000
Command line first.

861
01:29:42,000 --> 01:29:47,000
Command line first and then GUI, there is some GUI to be able to support that and there's some third party tools that help support that,

862
01:29:47,000 --> 01:29:50,000
but that command line straightforward, reusable, right?

863
01:29:50,000 --> 01:29:52,000
I love Hatchiko tool in general.

864
01:29:52,000 --> 01:30:00,000
The way Mishashi Moto had a lot of influence in my way of designing command lines, right?

865
01:30:00,000 --> 01:30:11,000
Like, a Hatch command line is a crime because we literally took the Cosmology StK modules because it gives you all these like Cobra modules that you just plug in.

866
01:30:11,000 --> 01:30:12,000
Right.

867
01:30:12,000 --> 01:30:13,000
And you get all the functionality.

868
01:30:13,000 --> 01:30:14,000
So we didn't want to like mess with that.

869
01:30:14,000 --> 01:30:19,000
And so we couldn't do anything beyond that we want to see in a command line.

870
01:30:19,000 --> 01:30:33,000
So, but a command line we're building right now from scratch that doesn't inherit the SDK functions, but essentially is designed to do deployments only and not anything else.

871
01:30:33,000 --> 01:30:37,000
The other answer functionality that you get with your chain management or not, right?

872
01:30:37,000 --> 01:30:47,000
So deployment first, a very Hatchiko inspired, you know, look and feel or experience where we're hiding people from Hatchiko and help us to do this stuff.

873
01:30:47,000 --> 01:31:00,000
Yep. But also, if you know any good command line developers, they're very, very hard to find by the way, like they're like almost UX people for the command line, you know, if you know anybody out there that's going to command line, please come talk to me.

874
01:31:00,000 --> 01:31:01,000
I love command.

875
01:31:01,000 --> 01:31:02,000
Look at my GitHub.

876
01:31:02,000 --> 01:31:11,000
Personally, any repository that I have over a thousand stars is all command line works are written.

877
01:31:11,000 --> 01:31:19,000
And also Mitch Mitch Hatchimori uses my command line libraries, which is, which I take with, you know, which is my biggest flex.

878
01:31:19,000 --> 01:31:24,000
So because I consider him to be the master of command lines, right?

879
01:31:24,000 --> 01:31:26,000
So I love command lines.

880
01:31:26,000 --> 01:31:35,000
And I will support your command line, you know, aspirations, like we wanted to kick ass command line come talk to me.

881
01:31:35,000 --> 01:31:38,000
Right. So, yeah.

882
01:31:38,000 --> 01:31:41,000
All right. Well, thanks for indulging me in my questions.

883
01:31:41,000 --> 01:31:47,000
I guess now we should let the fray wrap up is probably getting very, very late over there now.

884
01:31:47,000 --> 01:31:50,000
It is now very late here.

885
01:31:50,000 --> 01:32:04,000
So yeah, we always do we always do around Robin on the final question, Greg, just to just to wrap up, which is what are you most excited about this week in the cosmos coming up?

886
01:32:04,000 --> 01:32:09,000
Well, I've been so ingrained in like a cash test net.

887
01:32:09,000 --> 01:32:17,000
This week, actually, we have quite a few new providers coming out in a car or providing persistent storage.

888
01:32:17,000 --> 01:32:21,000
And we are having torching.

889
01:32:21,000 --> 01:32:25,000
Validated around our car should be the first real big validator.

890
01:32:25,000 --> 01:32:41,000
And it only has 21 validators. And they've been very kind enough to to delegate foundation delegations to to running on a cash because, as you know, a lot of the torching nodes around on AWS and GCP, which they're not happy about.

891
01:32:41,000 --> 01:32:43,000
And they're getting a lot of flack in the community.

892
01:32:43,000 --> 01:32:48,000
So I'll be working on my first torching first validator.

893
01:32:48,000 --> 01:32:52,000
I'll be running in production myself.

894
01:32:52,000 --> 01:33:01,000
I mean, a lot of people run validators on a cash and all the proceeds that we get from validating torching, we go back to the community.

895
01:33:01,000 --> 01:33:10,000
So we'll probably put that on a community pool and the community decide if they want the tokens to be distributed to themselves or burned or whatever.

896
01:33:10,000 --> 01:33:20,000
So it'll be fully community run, you know, running on a cash, a very publicly run validating and see how that happens.

897
01:33:20,000 --> 01:33:23,000
And I'll be joining the validator set, of course.

898
01:33:23,000 --> 01:33:24,000
Nice.

899
01:33:24,000 --> 01:33:26,000
Awesome.

900
01:33:26,000 --> 01:33:27,000
Right.

901
01:33:27,000 --> 01:33:28,000
Setting.

902
01:33:28,000 --> 01:33:36,000
Yeah, your prop 21, of course, that's, that's, is the most exciting.

903
01:33:36,000 --> 01:33:38,000
Is it more upgrades?

904
01:33:38,000 --> 01:33:41,000
It's getting up at 3am.

905
01:33:41,000 --> 01:33:43,000
It's most something. I'm not sure.

906
01:33:43,000 --> 01:33:56,000
I feel like I'm so disconnected these days from community at large and what's exciting was not, especially the main net and the upgrades and all these crazy things happening in a cash.

907
01:33:56,000 --> 01:34:04,000
I think being, being like, you know, such an integral part of a big, big project sort of limits your scope to what you can pay attention to.

908
01:34:04,000 --> 01:34:06,000
Sure. Yeah.

909
01:34:06,000 --> 01:34:08,000
Yeah.

910
01:34:08,000 --> 01:34:09,000
Okay.

911
01:34:09,000 --> 01:34:10,000
Sorry, the fray.

912
01:34:10,000 --> 01:34:12,000
What do you think of me?

913
01:34:12,000 --> 01:34:18,000
I'm most excited about getting my taxes done.

914
01:34:18,000 --> 01:34:23,000
It's five, five days or something beyond the end of the tax year.

915
01:34:23,000 --> 01:34:31,000
And everybody involved is going, Hey, man, where are the spreadsheets and saying, Hey, there's been 10 weeks of drama.

916
01:34:31,000 --> 01:34:36,000
And I am, I am also the book company bookkeeper.

917
01:34:36,000 --> 01:34:41,000
And we're three months behind nobody buys that they say, well, get it done.

918
01:34:41,000 --> 01:34:45,000
So I'm looking forward to that being done so I can get back to writing some code.

919
01:34:45,000 --> 01:34:47,000
That would be pretty cool.

920
01:34:47,000 --> 01:34:52,000
Mine then two weeks ago, April 15th, you played into that line here, right?

921
01:34:52,000 --> 01:34:54,000
Yeah, April 30th here.

922
01:34:54,000 --> 01:34:57,000
So, so yeah, too far they don't have to be filed for ages.

923
01:34:57,000 --> 01:35:05,000
It's just that your accountants, as soon as they, as soon as they crack it out, then their job's done, then they can invoice you and you're done.

924
01:35:05,000 --> 01:35:06,000
I know, right?

925
01:35:06,000 --> 01:35:07,000
It's invoice me.

926
01:35:07,000 --> 01:35:08,000
Right.

927
01:35:08,000 --> 01:35:15,000
Especially in our case, because the crypto is so complicated, they, they're basically like, you provide us the spreadsheets, then we will work it out.

928
01:35:15,000 --> 01:35:21,000
And I'm like, I'm pretty sure the harder bit is the bit that we're doing, not the bit that you're doing, but you know,

929
01:35:21,000 --> 01:35:26,000
Nevertheless, I'm looking forward to getting out of the way.

930
01:35:26,000 --> 01:35:32,000
There's quite a few exciting test nets coming up actually, which the defund one, which is like a ETF,

931
01:35:32,000 --> 01:35:35,000
ETF and the Cosmos space, I think is really interesting.

932
01:35:35,000 --> 01:35:38,000
A dropping to a cash holders, I hear.

933
01:35:38,000 --> 01:35:39,000
And there you go.

934
01:35:39,000 --> 01:35:40,000
Yeah, nice.

935
01:35:40,000 --> 01:35:46,000
So I'm quite interested in getting involved from that test net, maybe if I have time and there's,

936
01:35:46,000 --> 01:35:51,000
Oh, there's something, there's one other that I think a couple of the other folks on this call.

937
01:35:51,000 --> 01:35:56,000
There's one other interesting project that I saw this week and was like, well, that's kind of, oh, also Quicksilver is,

938
01:35:56,000 --> 01:35:59,000
I think just test that start, the Quicksilver.

939
01:35:59,000 --> 01:36:01,000
That's quite exciting.

940
01:36:01,000 --> 01:36:05,000
I'm back to your point about Tom from EcoStake being a great person.

941
01:36:05,000 --> 01:36:12,000
The Cosmos Interchain UK meetup, he's talking at the one in June, as is Joe from Quicksilver.

942
01:36:12,000 --> 01:36:20,000
So nice very good as well as cool guy, cool projects, which one am I talking about could be both.

943
01:36:20,000 --> 01:36:22,000
Talk about meetups.

944
01:36:22,000 --> 01:36:29,000
I go and hack with JunoFox quite a bit here in San Francisco.

945
01:36:29,000 --> 01:36:31,000
Yeah, because you know me out.

946
01:36:31,000 --> 01:36:39,000
I hacked in there, but every Saturday they come together at a cafe and then they hack together.

947
01:36:39,000 --> 01:36:41,000
It felt very pure.

948
01:36:41,000 --> 01:36:42,000
What was that?

949
01:36:42,000 --> 01:36:46,000
Are you talking about like Jake and Noah and Zeke?

950
01:36:46,000 --> 01:36:55,000
Jake, I only saw Jake and Dan Lynch is a good friend and Dan and Jake and there are a bunch of other people.

951
01:36:55,000 --> 01:36:59,000
I saw like 10 to 15 people all building cool things.

952
01:36:59,000 --> 01:37:05,000
I used to do that quite a lot in 2010, 2010, 2011 timeframe during my angel hack days.

953
01:37:05,000 --> 01:37:12,000
And like the mission coffee shop hacking on a Saturday is a very pure endeavor.

954
01:37:12,000 --> 01:37:20,000
I think we're going to, we're actually opening our office here, a big one, hopefully in the area, in the mission area.

955
01:37:20,000 --> 01:37:28,000
It's pretty much the same thing. The idea is to invite Cosmos contributors or Cosmos.

956
01:37:28,000 --> 01:37:33,000
Anything if you're working in a Cosmos ecosystem, doesn't matter what project, you should be able to come and hack with us.

957
01:37:33,000 --> 01:37:43,000
So I'm going to make it a thing and I'm going to make it very accessible to ecosystem projects.

958
01:37:43,000 --> 01:37:49,000
Need some hot offices for us bloody validated plebs when we're getting there.

959
01:37:49,000 --> 01:37:53,000
If you're a San Francisco, you're a validator. You can use our office.

960
01:37:53,000 --> 01:37:55,000
Is there a patent that you can use?

961
01:37:55,000 --> 01:37:57,000
Yeah.

962
01:37:57,000 --> 01:38:01,000
So I think, yeah, like offices are sort of becoming back in vogue now, right?

963
01:38:01,000 --> 01:38:04,000
It's fun to hack with people.

964
01:38:04,000 --> 01:38:06,000
Really fun to hack with people.

965
01:38:06,000 --> 01:38:17,000
So something interesting that the Daudau guys do is that they just sometimes jump into like the discord stream.

966
01:38:17,000 --> 01:38:19,000
And, you know, it's not organized or anything.

967
01:38:19,000 --> 01:38:25,000
It's just a couple of them all want to work on something and they just jump in there and stream it publicly.

968
01:38:25,000 --> 01:38:28,000
You can go in there and watch what they're doing and chat to them.

969
01:38:28,000 --> 01:38:30,000
It's really fun experience for me anyway.

970
01:38:30,000 --> 01:38:39,000
I'm not like a massive coder and I learn a lot every time I go in there and those guys are always willing to help me understand what they're doing, which is kind of cool.

971
01:38:39,000 --> 01:38:42,000
Probably counterproductive for them.

972
01:38:42,000 --> 01:38:45,000
I don't think I can code while streaming.

973
01:38:45,000 --> 01:38:50,000
I mean, it's really cool to collaborate anyway.

974
01:38:50,000 --> 01:38:51,000
Sorry.

975
01:38:51,000 --> 01:38:53,000
It's whether you can talk at the same time, right?

976
01:38:53,000 --> 01:38:57,000
It's kind of like the whole pair programming thing.

977
01:38:57,000 --> 01:39:04,000
I find I've literally, I will not take a job if they say pair programming, you got to do it.

978
01:39:04,000 --> 01:39:06,000
It's sometimes fun.

979
01:39:06,000 --> 01:39:07,000
I'll give you that.

980
01:39:07,000 --> 01:39:18,000
And pair programming, especially if your peers are at the same level as you, it's not fun when your peer is not at the same level, either higher level or lower level.

981
01:39:18,000 --> 01:39:26,000
But if you're the same level in terms of experience and seniority, it's a lot of fun because you're like actually exchanging ideas.

982
01:39:26,000 --> 01:39:32,000
And, you know, and one person is researching and other person is coding, you know, or you switch it up.

983
01:39:32,000 --> 01:39:33,000
It's actually a lot of fun.

984
01:39:33,000 --> 01:39:35,000
I got a lot of work done that way.

985
01:39:35,000 --> 01:39:39,000
It used to be a thing back in the day, but, you know, it used to be a religion back in the day.

986
01:39:39,000 --> 01:39:43,000
I hated that, but I like it, bearing with like my peers.

987
01:39:43,000 --> 01:39:54,000
But when I do a mob programming, which is just another like another escalation of this stupidity in my opinion, but that's, that's, I've given enough hot takes today.

988
01:39:54,000 --> 01:39:58,000
In this instance, they were working in like UIs.

989
01:39:58,000 --> 01:40:06,000
So I guess they were just, you know, talking about how it looks and then just changing stuff and talking about how it looks like iterating.

990
01:40:06,000 --> 01:40:09,000
It's like voluntary pairing or mobbing, isn't it?

991
01:40:09,000 --> 01:40:23,000
Like, I mean, I think the problem, I think the problem is that you, if you work in the industry and you do this stuff long enough, anything that people spontaneously do, because they have a couple of people who they know and trust and they like and they want to do work in that way.

992
01:40:23,000 --> 01:40:36,000
They'll work that way and be incredibly productive because you can't replicate that experience of working with people you like and trust and respect and work well with in the way that you want to work with them.

993
01:40:36,000 --> 01:40:51,000
And then people see that and they go, ah, the thing that must be good about that is that there were four people on the stream, not, oh, they're motivated to build this thing because it's a lot more interesting in building another fucking web shop or whatever it is you do for a big company, right?

994
01:40:51,000 --> 01:40:54,000
And so they go, right, thou shalt work in fours.

995
01:40:54,000 --> 01:40:59,000
And if you're a developer in particularly like, so my background is functional programming.

996
01:40:59,000 --> 01:41:11,000
So I am already a spiky character on some teams because my design patterns, the way I do stuff, the way I think about problems is very different and opinionated.

997
01:41:11,000 --> 01:41:14,000
And also because I've quite often been like a leader or whatever as well.

998
01:41:14,000 --> 01:41:19,000
So I can be a little bit like, have you thought about this and maybe we're going to do it this way.

999
01:41:19,000 --> 01:41:21,000
Use common lisp.

1000
01:41:21,000 --> 01:41:28,000
I mean, well, I mean, as you can see from my name on the thing, I am actually a lisp and I've spent the most of my career.

1001
01:41:28,000 --> 01:41:36,000
But you know, that said, like you end up in this situation where you're like, OK, well, you don't want to.

1002
01:41:36,000 --> 01:41:43,000
The reason the reason that the solution that you come up with here might be better is actually because of the clarity of your thinking.

1003
01:41:43,000 --> 01:41:45,000
It's not because of like working in a for or whatever.

1004
01:41:45,000 --> 01:41:51,000
And again, like it can be quite difficult to communicate other than with a working solution why an approach is better.

1005
01:41:51,000 --> 01:42:02,000
Like a really good example of this would be things like in reduction, reduce elimination of intermediate collections if you're using map iteration and reduce.

1006
01:42:02,000 --> 01:42:11,000
So if you're used to functional programming, you have the idea that you and Russ Russ has this so relevant smart contracts is that you have lazy evaluation, right?

1007
01:42:11,000 --> 01:42:16,000
So, you know, just like this idea ended up in Apache Spark, right?

1008
01:42:16,000 --> 01:42:22,000
It's the foundation of modern day processing, which is that you lay out everything you want to happen and it doesn't happen until you fold.

1009
01:42:22,000 --> 01:42:23,000
Right.

1010
01:42:23,000 --> 01:42:32,000
And you do like a right folder or a left folder or whatever, but like a reduction operation or an action or a collect or, you know, however, your language does this, you know.

1011
01:42:32,000 --> 01:42:41,000
And there's a whole bunch of different ways you can approach any given problem and it's quite hard sometimes to communicate though.

1012
01:42:41,000 --> 01:42:43,000
I don't know what I agree with.

1013
01:42:43,000 --> 01:42:45,000
I mean, totally.

1014
01:42:45,000 --> 01:42:50,000
Like when you're doing stuff like that, you get so ingrained into that.

1015
01:42:50,000 --> 01:42:54,000
The map reduces like to especially with functional, you multiple levels deep.

1016
01:42:54,000 --> 01:42:56,000
You're like five levels deep.

1017
01:42:56,000 --> 01:42:59,000
It's hard to talk and hard to explain.

1018
01:42:59,000 --> 01:43:00,000
Right.

1019
01:43:00,000 --> 01:43:02,000
So, exactly.

1020
01:43:02,000 --> 01:43:08,000
It's like whether you can stream and talk and also communicate with somebody else when you're filling your idea space with these things.

1021
01:43:08,000 --> 01:43:25,000
And I think some also, I mean, I think in the philosophy of programming languages, some programming languages are geared towards lone wolf programming because they give you much more expressivity, much more terseness at the expense of being less readable.

1022
01:43:25,000 --> 01:43:30,000
I mean, nobody's going to argue Haskell is more readable than go, right?

1023
01:43:30,000 --> 01:43:41,000
But you can probably write a 40,000 line go program in like 40 lines of Haskell in some cases.

1024
01:43:41,000 --> 01:43:54,000
Like if you like pure script is an example where if you see like a React application written in pure script, it can be tiny by comparison to the React equivalent and Clojure script has that property as well.

1025
01:43:54,000 --> 01:44:00,000
You can have a multiple less lines and you would in JavaScript.

1026
01:44:00,000 --> 01:44:01,000
Right.

1027
01:44:01,000 --> 01:44:05,000
But it comes at the cost of these languages.

1028
01:44:05,000 --> 01:44:16,000
They have extra inherent complexity because you can you have all these additional language structures like juxt and you know, whatever that are just quite alien.

1029
01:44:16,000 --> 01:44:18,000
You're writing calculus.

1030
01:44:18,000 --> 01:44:20,000
Right.

1031
01:44:20,000 --> 01:44:22,000
Sorry.

1032
01:44:22,000 --> 01:44:24,000
I was going to say something.

1033
01:44:24,000 --> 01:44:35,000
The house coffee is a place that yeah, I was going to say I think Alan is one of the gents who are frequent horse from the dad out guys.

1034
01:44:35,000 --> 01:44:37,000
Sorry, that was a bit of an interruption.

1035
01:44:37,000 --> 01:44:38,000
It's a really cool group.

1036
01:44:38,000 --> 01:44:43,000
I feel like I'm just going to highly recommend the hack.

1037
01:44:43,000 --> 01:44:45,000
So sorry, the phrase.

1038
01:44:45,000 --> 01:44:50,000
Did you have any more to add or I should not be adding anymore.

1039
01:44:50,000 --> 01:44:57,000
I mean, any more to the functional programming fire or the what's called fire the programming fire.

1040
01:44:57,000 --> 01:45:05,000
So I'm going to shut up and because I mean, say you're like literally an hour.

1041
01:45:05,000 --> 01:45:07,000
I was always late in show.

1042
01:45:07,000 --> 01:45:09,000
It turns extremely technical from the fray.

1043
01:45:09,000 --> 01:45:10,000
Right.

1044
01:45:10,000 --> 01:45:11,000
Like I started getting to some real.

1045
01:45:11,000 --> 01:45:13,000
It's like a symptom of him being tired.

1046
01:45:13,000 --> 01:45:17,000
I think it is like all all the filters go off.

1047
01:45:17,000 --> 01:45:18,000
I just start talking code.

1048
01:45:18,000 --> 01:45:21,000
Like if we let you go, you just start speaking assembly.

1049
01:45:21,000 --> 01:45:27,000
Yeah, my best work like traditionally like at the point is you know, you have like 90 minute sleep cycles.

1050
01:45:27,000 --> 01:45:28,000
Yeah.

1051
01:45:28,000 --> 01:45:29,000
I have my most productive period.

1052
01:45:29,000 --> 01:45:35,000
I think when I'm just on the crest down towards literally falling asleep in my chair.

1053
01:45:35,000 --> 01:45:38,000
So, so I think it's probably that like this time in the morning.

1054
01:45:38,000 --> 01:45:40,000
In fact, this is exact.

1055
01:45:40,000 --> 01:45:44,000
You know what it's one of my sleep cycles hits almost exactly at midnight.

1056
01:45:44,000 --> 01:45:50,000
So I'm actually probably about 10 minutes after the end of a fatigue point.

1057
01:45:50,000 --> 01:45:52,000
So I'm probably coming back up.

1058
01:45:52,000 --> 01:45:54,000
There you go.

1059
01:45:54,000 --> 01:45:57,000
I think my most productive time is fast becoming 3am.

1060
01:45:57,000 --> 01:46:07,000
I've started just waking up at like 2 30 and just rising like a friggin like a vampire out of my casket.

1061
01:46:07,000 --> 01:46:10,000
Just like directly straight up.

1062
01:46:10,000 --> 01:46:12,000
Did I miss an upgrade?

1063
01:46:12,000 --> 01:46:17,000
Okay, so you said what are you excited about next week in cosmos this week and moving all my stuff to a cache.

1064
01:46:17,000 --> 01:46:18,000
That's number one.

1065
01:46:18,000 --> 01:46:20,000
Anything that's not a cash is going to a cash.

1066
01:46:20,000 --> 01:46:22,000
I really appreciate this.

1067
01:46:22,000 --> 01:46:23,000
Hold the phone.

1068
01:46:23,000 --> 01:46:27,000
That was a bit of a great what's going to a cash.

1069
01:46:27,000 --> 01:46:35,000
Anything that I have like in terms of web hosting or yes, anything outside of well right now valid values.

1070
01:46:35,000 --> 01:46:41,000
I need really need to look more at that because we run a pretty large colo structure where we use some cloud services and things like that.

1071
01:46:41,000 --> 01:46:48,000
But I haven't really got to that point about running nodes where we run a tremendous number of non-validator nodes, right?

1072
01:46:48,000 --> 01:46:54,000
And to be able to reduce the cost of that and be able to be able to have there's they are femoral right now.

1073
01:46:54,000 --> 01:47:04,000
Those are things that they have uptime but are not exactly necessary in terms of in sorts of state right now, which I know is coming.

1074
01:47:04,000 --> 01:47:06,000
But right.

1075
01:47:06,000 --> 01:47:18,000
So having been a user of cosmos omnibus for regular nodes, it is very good.

1076
01:47:18,000 --> 01:47:26,000
Even without like, you know, if you're coming in from a say a poker to snapshot or one of their main team snapshots.

1077
01:47:26,000 --> 01:47:30,000
I'm not sure how soon they are how, you know, small they are.

1078
01:47:30,000 --> 01:47:36,000
But man, it's Europe, you can point stuff to it and Europe.

1079
01:47:36,000 --> 01:47:42,000
You just want to, you know, tune your resources so they don't get nuked when someone starts hitting them on the RPC port.

1080
01:47:42,000 --> 01:47:58,000
But I think you'll have a very good time if you want to try and move some of your resource into a cache and I'd be happy to like give you a bloody pointer on omnibus, which is super easy to use anyway.

1081
01:47:58,000 --> 01:48:11,000
Fun fact made the PR on that for the Juno edition when Juno first started and Adam the Liborism, which wasn't working on omnibus at the time.

1082
01:48:11,000 --> 01:48:13,000
That's my contribution.

1083
01:48:13,000 --> 01:48:15,000
Delegate the king notes.

1084
01:48:15,000 --> 01:48:27,000
Isn't and you surfer, if you're moving your stuff to a cache, and if you want to delete if you want to validate an acacia, but not only that actor said, we'll do the delegation because one of the requirements is like,

1085
01:48:27,000 --> 01:48:31,000
if you're running on a cache, if you're a validator, you got a higher chance of delegation.

1086
01:48:31,000 --> 01:48:32,000
So talk to me.

1087
01:48:32,000 --> 01:48:34,000
All right, I like that.

1088
01:48:34,000 --> 01:48:46,000
I think away from all that, like the, you know, we're all in the cosmos and we have a tremendous amount of positive energy towards projects that are going on.

1089
01:48:46,000 --> 01:48:56,000
And I think the projects to me personally as a, as somebody who runs a small business within the cosmos, an independent validator, ones that make the most that that called me the most are extremely,

1090
01:48:56,000 --> 01:49:05,000
are extremely effective in utilitarian projects that use blockchain as a method of incentive.

1091
01:49:05,000 --> 01:49:13,000
And that's kind of how I got involved in blockchain and not so much just around the defi structure of things, but how do I build?

1092
01:49:13,000 --> 01:49:30,000
It's impossible to build a future vision around a really grandiose structure without either an absolute shit ton of cash or a way to be able to incentivize tens or thousands or hundreds of thousands or millions of people to be able to buy in that vision.

1093
01:49:30,000 --> 01:49:43,000
And I think blockchain and I think the cosmos, especially with low gas fees and a low barrier of entry and everything else is the right tool set to build a tremendous amount of incentivizing around specific ideas.

1094
01:49:43,000 --> 01:49:46,000
And this is a great idea to build incentivizing around.

1095
01:49:46,000 --> 01:49:48,000
And it's just a worthy cause to be able to do that.

1096
01:49:48,000 --> 01:49:52,000
And so I didn't understand, I didn't really know the history and going all the way back to 2013.

1097
01:49:52,000 --> 01:49:56,000
I think it's fascinating around how this vision, it wasn't, it wasn't a cosmos based vision, right?

1098
01:49:56,000 --> 01:50:10,000
It was a vision around how to be able to solve a very complex and fairly scalable project, a problem, and then really kind of fate brought you together with a tool set that allows you to get a partial of the way there, right?

1099
01:50:10,000 --> 01:50:11,000
Yeah.

1100
01:50:11,000 --> 01:50:14,000
So anything to do to support that, man, that's awesome.

1101
01:50:14,000 --> 01:50:17,000
Like that, that's something that's a worthwhile thing.

1102
01:50:17,000 --> 01:50:20,000
So that's, that's one thing this week.

1103
01:50:20,000 --> 01:50:25,000
Otherwise, oh, let's see.

1104
01:50:25,000 --> 01:50:28,000
But that was a lot.

1105
01:50:28,000 --> 01:50:33,000
I think outside of that, we have, um, no, I think that's, I think that's it for me actually this way.

1106
01:50:33,000 --> 01:50:37,000
That's a great goal for this week and something I'm excited about, about making some changes too.

1107
01:50:37,000 --> 01:50:38,000
Nice.

1108
01:50:38,000 --> 01:50:40,000
She won't see.

1109
01:50:40,000 --> 01:50:44,000
I'm excited about the secret upgrade next week.

1110
01:50:44,000 --> 01:50:47,000
And especially with the soft, that's going to be joining us next week.

1111
01:50:47,000 --> 01:50:51,000
More or less during the upgrade, I'm really curious of how it's going to go down.

1112
01:50:51,000 --> 01:50:58,000
Is he going to folding out because there's not great to manage who knows, but I'm excited for one way or another.

1113
01:50:58,000 --> 01:51:05,000
It'll be, that'll be, I think mine and blockpains first secret network upgrade.

1114
01:51:05,000 --> 01:51:10,000
So I love the names they come up with for those upgrades, by the way.

1115
01:51:10,000 --> 01:51:17,000
Yeah, no, they put, they put some good effort into the marketing side of it for sure.

1116
01:51:17,000 --> 01:51:20,000
So I think that leaves this with me.

1117
01:51:20,000 --> 01:51:24,000
I'm excited about like upgrading some security around my nodes.

1118
01:51:24,000 --> 01:51:36,000
So we're moving entirely to TM KMS as an interim measure until we're more comfortable with Horcrux and deploying Horcrux.

1119
01:51:36,000 --> 01:51:44,000
I'm a little bit reluctant on Horcrux at the moment from reading an issue that someone else had where they double signed.

1120
01:51:44,000 --> 01:51:51,000
So making sure that we're outside of that narrative when we are ready to deploy Horcrux.

1121
01:51:51,000 --> 01:52:01,000
But I think that TM KMS is a good interim step to make sure that our nodes are completely key free in that area.

1122
01:52:01,000 --> 01:52:05,000
All our signing material is completely firewalled off and secure.

1123
01:52:05,000 --> 01:52:08,000
So TM KMS is remote signer, right?

1124
01:52:08,000 --> 01:52:09,000
Sorry.

1125
01:52:09,000 --> 01:52:12,000
Is TM KMS a remote signer?

1126
01:52:12,000 --> 01:52:14,000
Remote signer, yeah. So that's what I was saying.

1127
01:52:14,000 --> 01:52:19,000
I think TM KMS is actually a good fit for Cosmos Omnibus.

1128
01:52:19,000 --> 01:52:29,000
Because at the moment, you pretty much have to keep your signing material on S3 storage the way it's set up, which is a little bit.

1129
01:52:29,000 --> 01:52:35,000
It's probably, wow, I mean, it adds a vector that people could get to.

1130
01:52:35,000 --> 01:52:41,000
It's just another thing to manage in terms of security, I suppose, if you're trying to run an actual validator.

1131
01:52:41,000 --> 01:52:51,000
But like I was saying, the way Cosmos Omnibus is set up, you would have to rejigger some of the scripting and possibly, well, actually, that's not entirely true.

1132
01:52:51,000 --> 01:52:58,000
Because you can actually set environment variables for your config files.

1133
01:52:58,000 --> 01:53:02,000
So it probably would then, in that case, be very easy.

1134
01:53:02,000 --> 01:53:15,000
It'd be nice if you could create some automation to link the deployment address that you're assigned after you deploy to automatically update your signer.

1135
01:53:15,000 --> 01:53:17,000
That'd be fun.

1136
01:53:17,000 --> 01:53:24,000
We've been looking at TM KMS. I think we should use TM KMS. It's amazing.

1137
01:53:24,000 --> 01:53:33,000
Well, I mean, really, it is already an option for anyone who wants to use Omnibus because of the environment variable configuration.

1138
01:53:33,000 --> 01:53:35,000
So it's just another line.

1139
01:53:35,000 --> 01:53:40,000
Well, it's two lines, one to turn off your local key.

1140
01:53:40,000 --> 01:53:44,000
And then you just wouldn't add TM KMS.

1141
01:53:44,000 --> 01:53:53,000
Sorry, you just wouldn't add S3 and then point towards your, well, allow your TM KMS IP.

1142
01:53:53,000 --> 01:53:59,000
So which kind of, I think, were, no, yeah, something that worked just fine.

1143
01:53:59,000 --> 01:54:06,000
But just linking that on the TM KMS side of things, I think, in an automated way would be pretty handy.

1144
01:54:06,000 --> 01:54:10,000
Yeah, we should use that as some examples there, for sure.

1145
01:54:10,000 --> 01:54:12,000
Yeah, for sure.

1146
01:54:12,000 --> 01:54:14,000
Anyway, that wraps it for me.

1147
01:54:14,000 --> 01:54:18,000
Did you want to walk us out, Afra?

1148
01:54:18,000 --> 01:54:22,000
You haven't cut me off yet. It's amazing.

1149
01:54:22,000 --> 01:54:30,000
Yeah, no, I've spent a few weeks kind of basically trolling you too aggressively and I kind of feel bad for it.