WEBVTT 00:00:00.000 --> 00:00:03.779 Can an AI truly outperform a human in the chaotic, 00:00:03.799 --> 00:00:07.139 high stakes game of cryptocurrency? That's really 00:00:07.139 --> 00:00:09.000 the compelling question at the heart of our deep 00:00:09.000 --> 00:00:10.580 dive today. We're looking at this fascinating 00:00:10.580 --> 00:00:14.000 experiment building and then, you know, unleashing 00:00:14.000 --> 00:00:16.760 a custom AI agent into the world of meme coin 00:00:16.760 --> 00:00:19.839 trading on Solana, which is notoriously volatile. 00:00:20.059 --> 00:00:21.719 And what makes this really interesting is that 00:00:21.719 --> 00:00:23.219 they didn't just theorize, they actually built 00:00:23.219 --> 00:00:25.559 a tool and the results were, well, surprisingly 00:00:25.559 --> 00:00:28.199 clear. So this deep dive is based on Project 00:00:28.199 --> 00:00:31.140 Orca. AI is triumphant crypto trading. It's a 00:00:31.140 --> 00:00:32.659 fascinating account of pushing the boundaries 00:00:32.659 --> 00:00:35.100 of AI and finance. Our mission today is really 00:00:35.100 --> 00:00:38.039 to unpack exactly how this AI works, why it matters, 00:00:38.280 --> 00:00:39.840 and maybe what it means for anyone interested 00:00:39.840 --> 00:00:42.119 in navigating these complex fast -moving information 00:00:42.119 --> 00:00:44.960 environments. So let's dive in. Yeah, and you 00:00:44.960 --> 00:00:47.799 know, before we even get to the results, which 00:00:47.799 --> 00:00:50.859 are quite something, it's really crucial to understand 00:00:50.859 --> 00:00:53.479 why this whole thing is so significant. for us 00:00:53.479 --> 00:00:56.520 as humans, I mean, especially in these high speed, 00:00:56.740 --> 00:00:58.820 high stress trading environments like Solana 00:00:58.820 --> 00:01:02.700 meme coins, the challenges are just inherent 00:01:02.700 --> 00:01:06.430 and they often lead to these... common vicious 00:01:06.430 --> 00:01:09.109 cycles for retail traders. Oh, I see it constantly. 00:01:09.170 --> 00:01:12.290 The sheer speed of Solana mean coins just seems 00:01:12.290 --> 00:01:14.450 to turbocharge these human pitfalls, doesn't 00:01:14.450 --> 00:01:17.069 it? Like, imagine the FOMO when a coin jumps 00:01:17.069 --> 00:01:20.730 500 % in minutes or just the paralysis from seeing 00:01:20.730 --> 00:01:22.969 hundreds of new tokens launch every single day. 00:01:23.170 --> 00:01:25.129 Our brains just aren't built for that. Exactly. 00:01:25.150 --> 00:01:27.069 You get caught in that FOMO cycle, right? Buying 00:01:27.069 --> 00:01:29.810 a hyped coin, watching it moon, you know, skyrocket, 00:01:29.909 --> 00:01:31.590 then boom, you get rug pulled and you're left 00:01:31.590 --> 00:01:33.790 holding the bag because the devs just vanished 00:01:33.790 --> 00:01:37.200 with the money. Or there's the emotionally driven 00:01:37.200 --> 00:01:39.920 technical analysis cycle. People spend hours, 00:01:39.920 --> 00:01:42.120 you know, meticulously charting, only to let 00:01:42.120 --> 00:01:44.219 emotion completely override their own strategy. 00:01:44.299 --> 00:01:46.140 They ignore their stop losses. That happens all 00:01:46.140 --> 00:01:48.599 the time. And then, inevitably, they start blaming 00:01:48.599 --> 00:01:50.859 everyone else for the losses. And then there's 00:01:50.859 --> 00:01:53.680 the information overload cycle, joining dozens 00:01:53.680 --> 00:01:56.159 of alpha groups, getting totally conflicting 00:01:56.159 --> 00:01:58.920 signals, missing the entries. And then FOMO -ing 00:01:58.920 --> 00:02:01.819 in right at the top. Right at the very top. It's 00:02:01.819 --> 00:02:03.920 exhausting just describing it. It really is. 00:02:04.000 --> 00:02:06.500 It almost sounds like humans are just, well... 00:02:06.349 --> 00:02:08.729 fundamentally unsuited for this specific game. 00:02:09.310 --> 00:02:11.849 Our biology, our brains, they're just not wired 00:02:11.849 --> 00:02:14.889 to process and react in real time when narratives 00:02:14.889 --> 00:02:17.750 can flip every 15 minutes. We're limited by, 00:02:17.750 --> 00:02:20.189 you know, how much info we can process, cognitive 00:02:20.189 --> 00:02:23.889 biases, emotional responses, FUD, FOMO, revenge 00:02:23.889 --> 00:02:27.590 trading, our oda loop, observe, orient, decide, 00:02:27.810 --> 00:02:30.509 act. It takes minutes, sometimes hours. So how 00:02:30.509 --> 00:02:32.490 does an AI agent compare to that? Well, that's 00:02:32.490 --> 00:02:34.629 where the contrast becomes really stark. AI agents, 00:02:34.789 --> 00:02:36.550 they don't sleep. They don't get emotional. They 00:02:36.550 --> 00:02:39.009 don't revenge trade after a loss. They aren't 00:02:39.009 --> 00:02:42.030 swayed by FUD fear, uncertainty, doubt, or FOMO 00:02:42.030 --> 00:02:44.830 fear of missing out. They just scan, analyze, 00:02:45.090 --> 00:02:47.680 decide, and execute consistently. relentlessly, 00:02:47.900 --> 00:02:50.039 and their OD loops were talking milliseconds. 00:02:50.199 --> 00:02:53.479 Milliseconds, wow. So this led to the core question 00:02:53.479 --> 00:02:56.080 of the experiment, didn't it? Could an AI agent, 00:02:56.159 --> 00:02:58.180 with the exact same information we have access 00:02:58.180 --> 00:03:01.800 to, outperform the average human trader just 00:03:01.800 --> 00:03:04.699 by sticking to its strategy with, like, ruthless 00:03:04.699 --> 00:03:07.240 consistency? And the source hinted at the answer. 00:03:07.360 --> 00:03:09.240 Yeah, the short answer, according to them, was 00:03:09.240 --> 00:03:12.990 yes. And it wasn't even close. OK, that's a powerful 00:03:12.990 --> 00:03:15.849 statement. And it really explains why Project 00:03:15.849 --> 00:03:18.610 Orca caught our eye. It wasn't some black box 00:03:18.610 --> 00:03:20.789 thing you just throw money at. It was a custom 00:03:20.789 --> 00:03:25.949 GPT built on OpenAI's platform. That feels empowering 00:03:25.949 --> 00:03:28.689 somehow. It suggests that anyone with a solid 00:03:28.689 --> 00:03:30.830 idea can potentially build a specialized agent, 00:03:31.030 --> 00:03:33.889 not just coding wizards. Exactly. So let's break 00:03:33.889 --> 00:03:35.629 down the four core components. This is essentially 00:03:35.629 --> 00:03:38.009 the blueprint they used, right? Right. The first 00:03:38.009 --> 00:03:39.949 part, and you could argue it's the most critical, 00:03:40.129 --> 00:03:42.210 is the instructions. They call it the soul of 00:03:42.210 --> 00:03:44.849 the GPT. This is where you define its personality, 00:03:45.069 --> 00:03:47.729 its role, rules, goals, all in just plain English. 00:03:47.789 --> 00:03:50.650 You're basically programming it with words. OK. 00:03:50.650 --> 00:03:53.270 And for Orca, that core prompt was super specific. 00:03:53.330 --> 00:03:56.949 It was like, you are Orca, an elite decentralized 00:03:56.949 --> 00:03:59.849 finance analyst. You communicate clearly, concisely, 00:03:59.849 --> 00:04:02.810 and with data -driven reasoning. Every recommendation 00:04:02.810 --> 00:04:05.569 must have rationale, data, risk -reward analysis. 00:04:06.050 --> 00:04:08.250 Emotion is something you analyze in others, not 00:04:08.250 --> 00:04:11.039 something you experience. That last bit is key. 00:04:11.259 --> 00:04:13.280 Really emphasizes the objectivity, doesn't it? 00:04:13.300 --> 00:04:15.199 Yeah. And its core directives were just as clear. 00:04:15.539 --> 00:04:18.240 Mission, find and analyze Solana meme coins. 00:04:18.680 --> 00:04:21.560 Selection criteria, focus on low cap, new tokens, 00:04:21.759 --> 00:04:23.920 lock liquidity, healthy holder distribution, 00:04:24.100 --> 00:04:26.579 specifically a Janini coefficient below 0 .8. 00:04:26.759 --> 00:04:29.040 And the Gigani coefficient, that's about avoiding 00:04:29.040 --> 00:04:31.019 coins concentrated in just a few hands, right? 00:04:31.079 --> 00:04:33.620 Yeah. Less risk of manipulation. Precisely. It 00:04:33.620 --> 00:04:37.319 also had to scan for narratives, AI, DEPIN, GameFi 00:04:37.319 --> 00:04:40.329 themes. Risk management was... baked in, minimum 00:04:40.329 --> 00:04:44.069 40 % profit target, hard 15 % stop loss, and 00:04:44.069 --> 00:04:46.230 a crucial forbidden rule. Avoid anything looking 00:04:46.230 --> 00:04:48.310 like a honey pot where you can buy but not sell 00:04:48.310 --> 00:04:50.730 or having dodgy developer activity. Okay, so 00:04:50.730 --> 00:04:53.009 it had its strategy, its rule book. Once the 00:04:53.009 --> 00:04:55.750 soul is set, you give it knowledge, its own sort 00:04:55.750 --> 00:04:58.329 of private library. This involves uploading your 00:04:58.329 --> 00:05:01.769 own documents, right? PDFs, text files, turning 00:05:01.769 --> 00:05:05.310 it from a generalist AI into a specialist. Exactly. 00:05:05.389 --> 00:05:08.069 And for Orca, the knowledge base was really insightful. 00:05:08.529 --> 00:05:11.470 It included a PDF report on major meme coin rug 00:05:11.470 --> 00:05:14.529 pulls from 2024 detailing the early warning signs. 00:05:14.589 --> 00:05:16.569 Oh, that's smart. Learning from history. Yeah, 00:05:16.569 --> 00:05:19.529 and a text file with lists of reputable developer 00:05:19.529 --> 00:05:22.670 wallets versus known blacklisted ones. And crucially, 00:05:22.829 --> 00:05:24.560 Instru - professional docs on how to analyze 00:05:24.560 --> 00:05:26.720 smart contracts on Solscan, the Solana block 00:05:26.720 --> 00:05:29.399 explorer to spot malicious functions. So it knew 00:05:29.399 --> 00:05:31.600 what bad looked like based on real examples. 00:05:31.660 --> 00:05:33.420 Right. It wasn't just looking for good. It was 00:05:33.420 --> 00:05:36.459 actively avoiding known bad patterns. OK, component 00:05:36.459 --> 00:05:38.800 three is capabilities. These are the built -in 00:05:38.800 --> 00:05:41.259 open AI tools you switch on or off. That's right. 00:05:41.379 --> 00:05:43.620 Based on what your AI needs to do. For Orca, 00:05:43.839 --> 00:05:46.220 they enabled web browse. Absolutely critical 00:05:46.220 --> 00:05:49.279 for real time data, news, sentiment. The GPT 00:05:49.279 --> 00:05:51.839 would be blind otherwise. Makes sense. And advanced 00:05:51.839 --> 00:05:55.160 data analysis. or code interpreter. Super useful 00:05:55.160 --> 00:05:58.000 for crunching complex data like transaction histories. 00:05:58.399 --> 00:06:01.040 They obviously disabled daily three image generation. 00:06:01.420 --> 00:06:03.560 Orca wasn't making memes, it was trading them. 00:06:04.259 --> 00:06:06.259 Right. No need for pretty pictures. Definitely 00:06:06.259 --> 00:06:09.100 not. And the fourth component, actions. This 00:06:09.100 --> 00:06:11.629 is where it gets... really powerful. This is 00:06:11.629 --> 00:06:15.189 the hands. This lets the GPT actually connect 00:06:15.189 --> 00:06:17.970 to the outside world and do things using API's 00:06:17.970 --> 00:06:20.310 application programming interfaces. Nah. So it's 00:06:20.310 --> 00:06:22.730 not just thinking, it's acting. It had specific 00:06:22.730 --> 00:06:26.170 actions then. Yes. It had a get token data action 00:06:26.170 --> 00:06:28.550 connecting to the BirdEye API. That's a real 00:06:28.550 --> 00:06:31.389 time crypto data feed for instant price liquidity 00:06:31.389 --> 00:06:34.149 volume. It also had get security score connecting 00:06:34.149 --> 00:06:36.829 to go plus security API for checking contracts, 00:06:36.990 --> 00:06:39.720 looking for honeypots, vulnerabilities. And the 00:06:39.720 --> 00:06:43.199 big one, execute trade. This sent requests to 00:06:43.199 --> 00:06:45.459 a secure intermediary server to actually place 00:06:45.459 --> 00:06:48.500 buy or sell orders on a day X, like radium. So 00:06:48.500 --> 00:06:51.459 you combine these four, the instructions, the 00:06:51.459 --> 00:06:54.379 soul, the knowledge, the library, the capabilities, 00:06:54.660 --> 00:06:57.699 the tools, and the actions, the hands. And you 00:06:57.699 --> 00:07:01.100 get Project Orca, an AI agent with a strategic 00:07:01.100 --> 00:07:04.220 brain, deep expertise, analytical tools, and 00:07:04.220 --> 00:07:06.839 the autonomy to act in the market. Right. And 00:07:06.839 --> 00:07:09.399 with that foundation, The real experiment started. 00:07:09.740 --> 00:07:12.319 They funded a phantom wallet, what was it, $250 00:07:12.319 --> 00:07:15.759 USDC? Yep, $250. And connected it to the server 00:07:15.759 --> 00:07:17.420 Orca Control. The source mentioned it didn't 00:07:17.420 --> 00:07:19.639 feel like just a script running. Yeah. Or like 00:07:19.639 --> 00:07:21.500 a ready trader. Yeah, like it already understood 00:07:21.500 --> 00:07:23.579 the game. OK, this is where it gets really interesting. 00:07:23.639 --> 00:07:26.199 Let's talk about Orca in action. Specific trades. 00:07:26.360 --> 00:07:29.240 Let's do it. First one, a token called PXLcat. 00:07:29.480 --> 00:07:31.620 Orca flagged it almost immediately after launch. 00:07:32.240 --> 00:07:34.540 Low cap, pixel art tat theme, pretty standard 00:07:34.540 --> 00:07:36.939 meme coin stuff. But Orca's reasoning was sharp. 00:07:37.100 --> 00:07:40.639 Liquidity, 42Ks, burned meaning locked, safer. 00:07:41.319 --> 00:07:44.620 Genie 0 .72, so healthy holder spread. Organic 00:07:44.620 --> 00:07:47.500 X engagement, not bots. Contract clean, entry 00:07:47.500 --> 00:07:49.699 criteria met. Clear and concise. So what action 00:07:49.699 --> 00:07:52.980 did it take? Executed a $50 buy, set its profit 00:07:52.980 --> 00:07:56.519 target at plus 45%, stop loss at necks of 15%. 00:07:56.519 --> 00:07:59.120 The token picked up steam, price went up, and 00:07:59.120 --> 00:08:01.899 Orca automatically exited at plus 52 % profit. 00:08:02.060 --> 00:08:05.160 A nice $26 game. OK, solid first rate, finding 00:08:05.160 --> 00:08:07.420 a gem early. Did it just stick to that formula, 00:08:07.459 --> 00:08:09.339 or did it adapt? What was next? It definitely 00:08:09.339 --> 00:08:12.519 adapted. Next significant one was Neuro. Orca 00:08:12.519 --> 00:08:15.060 picked this because it fit the AI narrative directive, 00:08:15.360 --> 00:08:18.279 a hot sector at the time. Ah, using its instructions 00:08:18.279 --> 00:08:21.120 about themes. Exactly, its reasoning. Volume 00:08:21.120 --> 00:08:24.459 up 300 % in the last hour. On -chain data shows 00:08:24.459 --> 00:08:26.360 lots of new wallets making small buys indicates 00:08:26.360 --> 00:08:29.240 retail interest, not whales manipulating. Higher 00:08:29.240 --> 00:08:32.070 risk, but upside aligned. So it laddered its 00:08:32.070 --> 00:08:34.669 entry, bought $25 straight away, placed another 00:08:34.669 --> 00:08:37.090 limit order a bit lower. Smart roof management. 00:08:37.110 --> 00:08:39.029 How did that play out? It was volatile. Dipped 00:08:39.029 --> 00:08:40.590 pretty close to the stop loss, actually. But 00:08:40.590 --> 00:08:42.549 then there was this fake partnership announcement, 00:08:42.730 --> 00:08:45.090 which pumped the price. Orga took some profits 00:08:45.090 --> 00:08:47.490 at plus 30 % and let the rest run, exiting at 00:08:47.490 --> 00:08:50.750 plus 68%. Another win, despite the volatility. 00:08:51.049 --> 00:08:53.149 Interesting. And you mentioned sometimes what 00:08:53.149 --> 00:08:55.370 an AI doesn't do is just as important. Tell me 00:08:55.370 --> 00:08:57.190 about the rejected trade, Monsoon. Oh, great. 00:08:57.269 --> 00:08:59.690 Monsoon was trending hard on DexGreener. Looked 00:08:59.690 --> 00:09:02.149 like an obvious play. the kind humans might FOMO 00:09:02.149 --> 00:09:04.549 into. Yeah, easy to get sucked in. But Orca's 00:09:04.549 --> 00:09:06.929 rejection rationale was spot on thanks to its 00:09:06.929 --> 00:09:09.769 rules and knowledge. It basically said, analyzing 00:09:09.769 --> 00:09:12.830 Monsoon, price looks good, but contract analysis 00:09:12.830 --> 00:09:15.190 shows mint authority is still enabled and held 00:09:15.190 --> 00:09:18.590 by the deployer. Over 55 % supply held by top 00:09:18.590 --> 00:09:22.309 10 wallets. Risk of rug pull or dev dump, extremely 00:09:22.309 --> 00:09:25.870 high. Avoiding. Wow. And what happened with Monsoon? 00:09:26.129 --> 00:09:28.269 Two days later, the creator minted millions of 00:09:28.269 --> 00:09:30.960 new tokens, dumped them on the market, Price 00:09:30.960 --> 00:09:33.460 crashed to basically zero or could dodge the 00:09:33.460 --> 00:09:36.210 bullet. A big one. A huge one. It's hard to imagine 00:09:36.210 --> 00:09:38.230 a human consistently catching that kind of trap 00:09:38.230 --> 00:09:39.669 in the heat of the moment. Yeah, absolutely. 00:09:39.809 --> 00:09:41.990 So after about 48 hours, what did the scoreboard 00:09:41.990 --> 00:09:43.549 look like? What were the overall results? Yeah, 00:09:43.570 --> 00:09:45.789 the results were pretty stark. Really showed 00:09:45.789 --> 00:09:48.090 what it could do in a short burst. Initial capital, 00:09:48.409 --> 00:09:53.730 $250. Final value, $685 .21. OK. Net profit, 00:09:54.070 --> 00:09:58.750 $435 .21. That's a return on investment of 174 00:09:58.750 --> 00:10:03.679 .08%. 174 % in 48 hours. Yep. Total trades were 00:10:03.679 --> 00:10:08.480 9, win rate was 77 .8%, 7 wins, 2 losses. Best 00:10:08.480 --> 00:10:12.559 trade was QWELP, up 112 .4%. Worst was ZIVP, 00:10:12.879 --> 00:10:15.600 hit the stop loss at negative 15%, exactly as 00:10:15.600 --> 00:10:17.320 programmed. OK, impressive numbers. Now they 00:10:17.320 --> 00:10:19.440 did acknowledge it wasn't flawless, right? Some 00:10:19.440 --> 00:10:21.299 minor issues. Yeah, they mentioned one trade 00:10:21.299 --> 00:10:23.580 had a high slippage, cost a bit of profit, and 00:10:23.580 --> 00:10:25.340 it missed some potential entries late at night 00:10:25.340 --> 00:10:28.820 due to API lag sometimes. So not perfect. Still. 00:10:29.139 --> 00:10:31.600 But the conclusion was undeniable, right? It 00:10:31.600 --> 00:10:34.259 massively outperformed what manual trading likely 00:10:34.259 --> 00:10:36.279 would have achieved in the same time frame. And 00:10:36.279 --> 00:10:39.350 here's the kicker. with almost zero human effort 00:10:39.350 --> 00:10:42.220 after the initial setup. Right. Even if a human 00:10:42.220 --> 00:10:44.919 team worked around the clock, matching that output, 00:10:45.100 --> 00:10:46.940 let alone beating it, would be incredibly tough. 00:10:47.480 --> 00:10:49.639 It traded while they slept, analyzed while they 00:10:49.639 --> 00:10:51.539 ate lunch, stuck to the plan when emotions would 00:10:51.539 --> 00:10:54.299 make a human flinch. The consistency and scale 00:10:54.299 --> 00:10:56.360 are the real takeaways. That consistency, yeah, 00:10:56.419 --> 00:10:59.120 it paints a really clear picture of maybe a new 00:10:59.120 --> 00:11:01.740 crypto paradigm emerging. Retail traders, let's 00:11:01.740 --> 00:11:04.120 face it, are often burned out, trying to manually 00:11:04.120 --> 00:11:07.039 track thousands of daily mean coin launches on 00:11:07.039 --> 00:11:09.460 Solana. It's not just hard, it's impossible. 00:11:09.659 --> 00:11:12.580 It really is. The game is changing. Agents like 00:11:12.580 --> 00:11:14.419 Orca, they feel like a new kind of participant. 00:11:14.720 --> 00:11:17.120 Never tired, never emotional, always executing 00:11:17.120 --> 00:11:20.000 the plan. It feels like more than just, you know, 00:11:20.200 --> 00:11:23.360 convenience. Oh, absolutely. Think of these AI 00:11:23.360 --> 00:11:25.460 agents as an extension of yourself, really, like 00:11:25.460 --> 00:11:28.139 an upgrade for your, let's be honest, slow human 00:11:28.139 --> 00:11:32.450 brain. Fair enough. It gives you 247 execution 00:11:32.450 --> 00:11:35.649 without burnout. You get personalized risk settings 00:11:35.649 --> 00:11:38.350 that it follows absolutely. Instant reaction 00:11:38.350 --> 00:11:40.950 to market shifts. It builds an on -chain memory, 00:11:41.250 --> 00:11:43.809 potentially getting smarter over time. And crucially, 00:11:44.190 --> 00:11:46.950 it shields you from your own emotional interference. 00:11:47.110 --> 00:11:49.549 So for you listening, the key thing to grasp 00:11:49.549 --> 00:11:53.559 is maybe. You're not working against these AI 00:11:53.559 --> 00:11:55.539 agents. You're working with them. They don't 00:11:55.539 --> 00:11:58.080 replace your strategies. They scale them. You 00:11:58.080 --> 00:12:00.100 design the playbook. They run it flawlessly. 00:12:00.659 --> 00:12:02.659 Exactly. The human role shifts, doesn't it? From 00:12:02.659 --> 00:12:04.539 being the person clicking the buy -sell buttons 00:12:04.539 --> 00:12:06.799 to being the architect of the strategy, the designer 00:12:06.799 --> 00:12:09.720 of the prompt. The skill isn't reading one 15 00:12:09.720 --> 00:12:12.480 -minute chart anymore. It's writing the instructions 00:12:12.480 --> 00:12:15.779 that let an AI interpret thousands of them simultaneously. 00:12:16.179 --> 00:12:18.639 A fundamental shift. And the source makes it 00:12:18.639 --> 00:12:20.679 clear this is just the beginning, right? They 00:12:20.679 --> 00:12:23.159 prove viability, but there's more to explore. 00:12:23.360 --> 00:12:25.179 For sure. They talked about future directions 00:12:25.179 --> 00:12:27.580 like running multiple wallets, maybe with competing 00:12:27.580 --> 00:12:29.960 strategies, one for scalping, one for swings, 00:12:30.600 --> 00:12:33.379 using more detailed prompts to test how nuanced 00:12:33.379 --> 00:12:36.500 the AI's understanding can get. Integrating more 00:12:36.500 --> 00:12:39.899 diverse data, maybe sentiment from obscure forums 00:12:39.899 --> 00:12:42.330 or something. Oh, interesting. And tracking long 00:12:42.330 --> 00:12:45.350 -term performance against human traders. But 00:12:45.350 --> 00:12:47.830 the bigger picture connecting it all seems to 00:12:47.830 --> 00:12:51.470 be this idea of humans and machines teaming up. 00:12:51.710 --> 00:12:54.129 Teaming up to fight the scams, the noise, the 00:12:54.129 --> 00:12:56.730 manipulation in these really wild market spaces. 00:12:57.090 --> 00:12:59.340 That makes sense. But it definitely raises some 00:12:59.340 --> 00:13:01.480 big questions, too, doesn't it? Like, what happens 00:13:01.480 --> 00:13:03.419 when everyone has an AI trading agent? Does it 00:13:03.419 --> 00:13:06.159 just become an AI arms race, fastest algorithm 00:13:06.159 --> 00:13:08.480 wins? That's a valid concern. And could it trigger 00:13:08.480 --> 00:13:10.500 unintended consequences, like flash crashes, 00:13:10.659 --> 00:13:12.820 if thousands of AIs react identically to the 00:13:12.820 --> 00:13:16.019 same signal? These are critical, ethical, and 00:13:16.019 --> 00:13:18.360 systemic questions we'll have to grapple with 00:13:18.360 --> 00:13:20.879 as this tech becomes more widespread. No easy 00:13:20.879 --> 00:13:23.799 answers there yet. But the source material itself 00:13:23.799 --> 00:13:25.740 offered a really compelling conclusion, I thought. 00:13:25.799 --> 00:13:28.100 It said something like, We don't need to beat 00:13:28.100 --> 00:13:30.460 the machines, we need to trade with them. The 00:13:30.460 --> 00:13:33.799 revolution won't be hard -coded, it will be prompted. 00:13:33.980 --> 00:13:36.080 It will be prompted. That's a powerful thought. 00:13:36.320 --> 00:13:38.740 It really reframes the whole relationship between 00:13:38.740 --> 00:13:42.539 us and AI in this space. So for you, our listener, 00:13:42.700 --> 00:13:44.399 maybe the question to think about after this 00:13:44.399 --> 00:13:47.379 deep dive is, how will you adapt? How will you 00:13:47.379 --> 00:13:50.220 change your approach to information, to decision 00:13:50.220 --> 00:13:52.580 making, in a world that's increasingly powered 00:13:52.580 --> 00:13:54.419 by intelligent agents like these?