WEBVTT

00:00:00.000 --> 00:00:02.500
Have you ever been scrolling through an article,

00:00:02.859 --> 00:00:05.299
maybe some kind of report, and you just get this

00:00:05.299 --> 00:00:09.539
feeling? Oh, yeah. That unmistakable, boring,

00:00:09.660 --> 00:00:14.039
repetitive text, it just feels cold, empty. Yeah,

00:00:14.060 --> 00:00:16.039
it's like it was written for a machine, you know,

00:00:16.079 --> 00:00:19.140
it uses all the same smart words like tapestry

00:00:19.140 --> 00:00:21.859
or vibrate. Real people just don't talk like

00:00:21.859 --> 00:00:23.539
that, especially when they're trying to explain

00:00:23.539 --> 00:00:25.679
something. Exactly. And that right there, that's

00:00:25.679 --> 00:00:28.399
the enemy. We're calling it Slop. It's basically

00:00:28.399 --> 00:00:30.600
junk food for the internet. Welcome to the Deep

00:00:30.600 --> 00:00:32.460
Dive. This whole conversation we're about to

00:00:32.460 --> 00:00:35.520
have is based on a guide with one mission, really,

00:00:35.880 --> 00:00:39.640
to help you humanize your AI output. We're here

00:00:39.640 --> 00:00:41.640
to help you cut through all that noise. We are

00:00:41.640 --> 00:00:44.399
unpacking what the guide calls the Slop Evader

00:00:44.399 --> 00:00:46.859
Mindset. I'm going to look at some specific prompt

00:00:46.859 --> 00:00:48.979
engineering techniques, but more importantly,

00:00:49.159 --> 00:00:52.560
a really strict four -step workflow to make sure

00:00:52.560 --> 00:00:55.159
everything you create feels like a real person

00:00:55.159 --> 00:00:57.390
actually wrote it. We'll get into which words

00:00:57.390 --> 00:00:59.710
you should just delete on site, how to talk to

00:00:59.710 --> 00:01:02.509
the AI like an assistant, not a magician, and

00:01:02.509 --> 00:01:04.870
how to use the past to clean up the future of

00:01:04.870 --> 00:01:06.829
your content. OK, so let's start with that core

00:01:06.829 --> 00:01:09.689
definition. What really is slop? Because it's

00:01:09.689 --> 00:01:12.790
more than just a few robotic words. It is. At

00:01:12.790 --> 00:01:16.030
its core, it's content made without any human

00:01:16.030 --> 00:01:20.209
care or perspective. The AI is trained to sound.

00:01:20.730 --> 00:01:22.989
agreeable and authoritative, so it defaults to

00:01:22.989 --> 00:01:26.370
this really formal wordy text. It uses delve

00:01:26.370 --> 00:01:29.370
when it should just say, look into. And if you

00:01:29.370 --> 00:01:31.370
publish that stuff, there are immediate consequences.

00:01:31.469 --> 00:01:33.390
Absolutely. The first problem is just the reader.

00:01:33.569 --> 00:01:35.109
They clicked your link because they wanted to

00:01:35.109 --> 00:01:37.849
hear from you. And if they hit a wall of boring,

00:01:38.049 --> 00:01:40.450
homogenized text, they're gone. They get bored.

00:01:40.569 --> 00:01:43.209
They get bored. And when they get bored, Google

00:01:43.209 --> 00:01:46.170
notices. That low engagement can really hurt

00:01:46.170 --> 00:01:48.569
your visibility. So this gets to the idea of

00:01:48.569 --> 00:01:50.730
who's in charge. I love them. analogy from the

00:01:50.730 --> 00:01:52.530
guide on this. You have to think of yourself

00:01:52.530 --> 00:01:55.530
as the chef. The AI can get you the ingredients,

00:01:55.629 --> 00:01:57.689
you know, it can chop the vegetables, but it

00:01:57.689 --> 00:02:00.109
can't cook the meal. Right. If you treat the

00:02:00.109 --> 00:02:02.709
AI like a magic button, you're just serving microwave

00:02:02.709 --> 00:02:05.950
food. A slop evader always treats the AI like

00:02:05.950 --> 00:02:08.349
a helper, a very powerful helper, but you're

00:02:08.349 --> 00:02:11.509
the boss. That is a perfect analogy. And this

00:02:11.509 --> 00:02:13.870
brings us to, I think, a vulnerable admission.

00:02:14.270 --> 00:02:16.090
The reason people keep the slop in the first

00:02:16.090 --> 00:02:18.699
place. Yeah, that's a tough one. We keep it because

00:02:18.699 --> 00:02:20.680
it sounds professional. We think that formal

00:02:20.680 --> 00:02:24.039
language means we're an expert. It's true. I

00:02:24.039 --> 00:02:26.719
still find myself fighting that instinct. Early

00:02:26.719 --> 00:02:29.219
in my career, I'd always use utilize instead

00:02:29.219 --> 00:02:30.860
of use, because I thought simple words sounded

00:02:30.860 --> 00:02:34.460
too casual. But the guide is right. Simple words

00:02:34.460 --> 00:02:37.800
actually show clarity and confidence. It's about

00:02:37.800 --> 00:02:40.919
stopping the need to delve into complex transformations

00:02:40.919 --> 00:02:44.620
and just look into changes. Exactly. It invites

00:02:44.620 --> 00:02:46.860
the reader in instead of pushing them away. You

00:02:46.860 --> 00:02:49.680
have to actively hunt down words like optimize,

00:02:50.060 --> 00:02:53.379
comprehensive, synergy, harness. They're all

00:02:53.379 --> 00:02:55.919
red flags that the AI is just padding for word

00:02:55.919 --> 00:02:58.680
count. So if we really commit to this, to being

00:02:58.680 --> 00:03:01.120
the chef and cutting out all that filler, how

00:03:01.120 --> 00:03:03.520
does this directly affect my website's visibility

00:03:03.520 --> 00:03:06.280
and health? Slop hurts ranking because board

00:03:06.280 --> 00:03:08.860
readers quickly click away. OK, let's get into

00:03:08.860 --> 00:03:11.900
the mechanics. Because knowing what slop is,

00:03:12.120 --> 00:03:14.840
doesn't help unless we know how to prompt the

00:03:14.840 --> 00:03:17.680
AI correctly. It seems like it's not enough to

00:03:17.680 --> 00:03:21.020
just have a prompt. You need to mix in real human

00:03:21.020 --> 00:03:23.539
data. Exactly. And that brings us to the grade

00:03:23.539 --> 00:03:26.680
five student rule. It's really powerful. When

00:03:26.680 --> 00:03:28.840
you tell the AI, explain this to me like I'm

00:03:28.840 --> 00:03:31.599
10 years old. You're not asking for childish

00:03:31.599 --> 00:03:34.500
content. You're just forcing it to use core concepts,

00:03:34.699 --> 00:03:36.819
right? Can't hide. Precisely. It can't hide behind

00:03:36.819 --> 00:03:39.659
jargon. It has to use analogies in simple language,

00:03:39.800 --> 00:03:42.500
which proves it actually understands the concept.

00:03:42.860 --> 00:03:44.960
It's a great defense against the AI just repeating

00:03:44.960 --> 00:03:47.539
stuff it's seen. And you have to be so specific.

00:03:47.699 --> 00:03:49.759
The guide has a great example prompt, something

00:03:49.759 --> 00:03:52.520
like, write about fixing a flat bicycle tire.

00:03:52.699 --> 00:03:55.879
Use short sentences. Do not use words like comprehensive,

00:03:56.240 --> 00:03:59.020
optimize, or utilize. Talk like a mechanic helping

00:03:59.020 --> 00:04:01.979
a best friend. Start with, first, flip your bike

00:04:01.979 --> 00:04:04.699
upside down. See? That prompt tells it what to

00:04:04.699 --> 00:04:07.219
write, who to be, who it's talking to, and what

00:04:07.219 --> 00:04:09.120
words are off limits. You're modeling the voice

00:04:09.120 --> 00:04:11.919
before it even starts writing. And even then,

00:04:12.159 --> 00:04:14.810
you can't fully trust the output. Yeah. Which

00:04:14.810 --> 00:04:17.189
brings us to this time machine idea. Right. So

00:04:17.189 --> 00:04:19.589
after the AI gives you that first draft, the

00:04:19.589 --> 00:04:21.550
Slopivator assumes it's still got some sludge

00:04:21.550 --> 00:04:23.970
in it. The time machine concept is where you

00:04:23.970 --> 00:04:27.009
copy an AI phrase like, uh, the best method is

00:04:27.009 --> 00:04:29.949
to securely attach the patch and then you search

00:04:29.949 --> 00:04:33.069
for it on Reddit or Quora. But how do you do

00:04:33.069 --> 00:04:35.149
that? How do you filter by date? Because if you

00:04:35.149 --> 00:04:37.290
just search now, you're just going to get more

00:04:37.290 --> 00:04:40.089
AI slop. That's the key. In a Google search,

00:04:40.110 --> 00:04:43.480
you go to tools. then custom range, and you set

00:04:43.480 --> 00:04:48.040
the to date as November 30th, 2022. That date

00:04:48.040 --> 00:04:51.060
is so important because it's right before the

00:04:51.060 --> 00:04:53.660
big public LLM started flooding the internet

00:04:53.660 --> 00:04:55.800
with their particular style of writing. So that's

00:04:55.800 --> 00:04:58.720
why. The articles from before that are 100 %

00:04:58.720 --> 00:05:00.779
human written. It's a pure sample. It's a pure

00:05:00.779 --> 00:05:03.240
sample. It shows you how real people use slang,

00:05:03.399 --> 00:05:05.160
how they joke, how they explain things with a

00:05:05.160 --> 00:05:07.560
bit of frustration or humor. Those are the human

00:05:07.560 --> 00:05:10.040
ingredients you need to find to fix the AI's

00:05:10.040 --> 00:05:11.879
text. And you have to layer in your own personal

00:05:11.879 --> 00:05:14.360
stories too. Oh, absolutely. Think about another

00:05:14.360 --> 00:05:16.740
prompt like writing about putting salt in your

00:05:16.740 --> 00:05:19.139
coffee instead of sugar. If you just ask the

00:05:19.139 --> 00:05:21.100
AI, it'll probably say something like a common

00:05:21.100 --> 00:05:25.160
morning ritual. Ah. You have to forbid that cliché.

00:05:25.180 --> 00:05:27.939
Instead, tell it to describe the gritty physical

00:05:27.939 --> 00:05:31.360
sensation of that first horrible taste. That's

00:05:31.360 --> 00:05:35.180
real. So why is checking old Reddit posts from

00:05:35.180 --> 00:05:39.019
that pre -AI slop era so critical? Old posts

00:05:39.019 --> 00:05:42.259
capture authentic slang and real humor before

00:05:42.259 --> 00:05:45.360
AI homogenized online language. OK, so moving

00:05:45.360 --> 00:05:48.939
from theory to actually doing it. Creating a

00:05:48.939 --> 00:05:52.110
long 2 ,000 -word article without slop needs

00:05:52.110 --> 00:05:54.730
a real system. It's not one click. Let's walk

00:05:54.730 --> 00:05:57.230
through the four steps. Right. Step one is create

00:05:57.230 --> 00:05:59.490
a human -first outline. This is where I think

00:05:59.490 --> 00:06:01.709
most people get it wrong. They ask the AI for

00:06:01.709 --> 00:06:03.670
an outline, and it spits out generic headings.

00:06:04.129 --> 00:06:06.250
Introduction, challenges, solutions. Yeah, that's

00:06:06.250 --> 00:06:08.170
useless. You have to stop asking the AI for a

00:06:08.170 --> 00:06:09.870
long article because it will just fill the space

00:06:09.870 --> 00:06:12.209
with junk to hit your word count. Instead, you

00:06:12.209 --> 00:06:14.410
the chef write down five to ten main questions

00:06:14.410 --> 00:06:16.730
you want to answer. And these have to come from

00:06:16.730 --> 00:06:19.310
real life, like the guides example, writing about

00:06:19.310 --> 00:06:22.019
a homestay business in California. A generic

00:06:22.019 --> 00:06:24.720
outline is no good. Your question should be things

00:06:24.720 --> 00:06:27.540
like, which area had the worst utility bills?

00:06:27.920 --> 00:06:30.800
Or, how did I find that amazing street food cart?

00:06:31.079 --> 00:06:35.060
Exactly. Then you move to step two. Use the SlopEvator

00:06:35.060 --> 00:06:38.019
tool for research, which is really the pre -AI

00:06:38.019 --> 00:06:40.560
check. You're back in the time machine. You use

00:06:40.560 --> 00:06:42.980
those date filters on Google, maybe search Stack

00:06:42.980 --> 00:06:46.139
Exchange to find pure human information from

00:06:46.139 --> 00:06:49.199
before November 2022. And we're looking for texture

00:06:49.199 --> 00:06:51.480
here. What kind of high value details? You're

00:06:51.480 --> 00:06:53.980
looking for complaints, brand names, little specific

00:06:53.980 --> 00:06:56.899
details. If a Reddit post from 2019 mentions

00:06:56.899 --> 00:06:59.000
a specific coffee shop in that neighborhood,

00:06:59.579 --> 00:07:02.040
that's gold. An AI would never find that, but

00:07:02.040 --> 00:07:04.180
it makes your writing feel instantly real. OK,

00:07:04.259 --> 00:07:06.899
so step three is command the AI to write in small

00:07:06.899 --> 00:07:09.180
bytes. This feels like the most important part

00:07:09.180 --> 00:07:12.060
for avoiding that drift. It is. The AI's brain

00:07:12.060 --> 00:07:14.639
just works better on one small task at a time.

00:07:14.920 --> 00:07:17.319
It has a smaller context window to worry about.

00:07:17.439 --> 00:07:20.199
If you ask for 2 ,000 words by word 1000, it's

00:07:20.199 --> 00:07:21.959
already forgetting the voice you told it to use

00:07:21.959 --> 00:07:24.199
at the start. So we feed it our research, our

00:07:24.199 --> 00:07:27.720
notes, and we just ask for part one. Why does

00:07:27.720 --> 00:07:30.920
keeping it under, say, 300 words make such a

00:07:30.920 --> 00:07:33.160
big difference? It just manages that context

00:07:33.160 --> 00:07:35.759
so much better. It forces the AI to be concise

00:07:35.759 --> 00:07:38.379
and stick to the rules you set. Your prompt should

00:07:38.379 --> 00:07:41.839
say, use I and you, be friendly, avoid words

00:07:41.839 --> 00:07:45.100
like tailored, and keep it under 300 words. Then

00:07:45.100 --> 00:07:47.339
you immediately review that chunk for any slop

00:07:47.339 --> 00:07:50.000
and fix it on the spot. And then finally, step

00:07:50.000 --> 00:07:54.300
four, the human polish and bridge building. You've

00:07:54.300 --> 00:07:55.980
got all these perfect little sections. Now you

00:07:55.980 --> 00:07:58.139
have to glue them together and you have to avoid

00:07:58.139 --> 00:08:01.259
the robotic glue. Exactly. You delete those cold

00:08:01.259 --> 00:08:04.339
transition words. Moreover, furthermore, and

00:08:04.339 --> 00:08:07.240
the absolute worst in conclusion, They just sound

00:08:07.240 --> 00:08:10.180
like a machine. Instead, you use human bridges.

00:08:10.579 --> 00:08:11.920
Things that sound like you're actually talking

00:08:11.920 --> 00:08:14.620
to the reader, but here's the catch. Or, another

00:08:14.620 --> 00:08:16.699
thing I noticed was... Yeah, or, so what does

00:08:16.699 --> 00:08:18.800
this actually mean for you? Right, and the final

00:08:18.800 --> 00:08:20.839
test is to read the whole thing out loud. If

00:08:20.839 --> 00:08:23.259
you stumble, if you feel bored reading your own

00:08:23.259 --> 00:08:25.480
work, there's still flop in there. Fix it. So

00:08:25.480 --> 00:08:28.379
what's the ultimate risk of just asking the AI

00:08:28.379 --> 00:08:30.819
for the full 2 ,000 words at once? Asking for

00:08:30.819 --> 00:08:33.240
too much length at once makes the AI fill the

00:08:33.240 --> 00:08:36.289
space with empty low -value slop. Even with a

00:08:36.289 --> 00:08:38.250
perfect workflow, there are still some common

00:08:38.250 --> 00:08:40.570
traps people fall into. Things that can make

00:08:40.570 --> 00:08:42.909
your writing actually seem unprofessional or

00:08:42.909 --> 00:08:45.610
shallow. The first trap is confusing simple with

00:08:45.610 --> 00:08:49.070
lazy. Simple writing is about taking a really

00:08:49.070 --> 00:08:52.190
hard idea and making it easy to understand. It's

00:08:52.190 --> 00:08:54.789
not about deleting the idea itself. Right. You

00:08:54.789 --> 00:08:56.309
still have to do deep research. You still need

00:08:56.309 --> 00:08:58.970
to read the expert blogs. But you use simple

00:08:58.970 --> 00:09:01.409
metaphors to explain the how and the why. So

00:09:01.409 --> 00:09:04.289
instead of saying AI optimizes complex workloads,

00:09:04.710 --> 00:09:06.850
you say something like, AI is like a very fast

00:09:06.850 --> 00:09:08.549
assistant. It handles boring stuff. So you get

00:09:08.549 --> 00:09:10.409
two extra hours to grab coffee with friends.

00:09:10.710 --> 00:09:13.610
The value is clear. Trap number two is forgetting

00:09:13.610 --> 00:09:17.090
to use I and you. Mm -hmm. to that passive language,

00:09:17.149 --> 00:09:19.850
right? It is believe that or report show. It

00:09:19.850 --> 00:09:22.490
sounds like a dusty old textbook. You have to

00:09:22.490 --> 00:09:25.090
be confident. Use I to share your own experience

00:09:25.090 --> 00:09:27.629
and you to talk directly to the reader. When

00:09:27.629 --> 00:09:30.230
I write about tech, I always say, I tried the

00:09:30.230 --> 00:09:32.029
setup and I found that you might see an error

00:09:32.029 --> 00:09:34.070
here. Let me show you how to fix it. That makes

00:09:34.070 --> 00:09:35.590
the reader feel like you're right there with

00:09:35.590 --> 00:09:37.629
them. You know, I still wrestle with prompt drift

00:09:37.629 --> 00:09:40.860
myself. Just. Keeping that consistent voice across

00:09:40.860 --> 00:09:44.460
a long piece, it's hard. It takes constant work.

00:09:44.779 --> 00:09:47.419
And that leads to trap three, being too afraid

00:09:47.419 --> 00:09:51.279
of technical words. In some topics, crypto, AI,

00:09:51.720 --> 00:09:54.580
engineering, you can't avoid them. So you keep

00:09:54.580 --> 00:09:56.379
the technical term, but you define it right away.

00:09:56.620 --> 00:09:59.159
Precisely. You keep the term, but you add a simple

00:09:59.159 --> 00:10:01.789
explanation immediately after it. For example,

00:10:02.129 --> 00:10:04.809
we used a large language model, or LLM, which

00:10:04.809 --> 00:10:06.950
is basically a statistical parrot trained on

00:10:06.950 --> 00:10:09.250
huge amounts of text. You show you're an expert,

00:10:09.389 --> 00:10:12.610
but you don't leave anyone behind. Whoa. Imagine

00:10:12.610 --> 00:10:15.509
scaling this kind of deep human analysis across

00:10:15.509 --> 00:10:17.789
a huge project like a whole book or a thesis,

00:10:18.309 --> 00:10:20.929
injecting that much real -world value into every

00:10:20.929 --> 00:10:23.470
single chapter. That's powerful. It's all about

00:10:23.470 --> 00:10:26.070
proving your expertise through action. Don't

00:10:26.070 --> 00:10:28.750
just say automation saves time, tell the story.

00:10:29.129 --> 00:10:31.169
I used to spend two hours cleaning data every

00:10:31.169 --> 00:10:34.190
Tuesday. Now I press one button. I use that time

00:10:34.190 --> 00:10:36.789
to go for a walk. You have to describe the actual

00:10:36.789 --> 00:10:39.070
buttons you click. That proves you know what

00:10:39.070 --> 00:10:41.190
you're talking about. And at the end of the day,

00:10:41.629 --> 00:10:44.610
that simple, helpful content is what Google wants

00:10:44.610 --> 00:10:47.769
to see. People stay on the page longer, and that

00:10:47.769 --> 00:10:50.240
helps you rank higher. So for specialized articles,

00:10:50.279 --> 00:10:52.919
should I completely remove all technical jargon

00:10:52.919 --> 00:10:56.940
to achieve this slop -avator style? Keep necessary

00:10:56.940 --> 00:10:59.460
jargon, but immediately follow it with a plain

00:10:59.460 --> 00:11:02.100
English definition. Okay, let's summarize the

00:11:02.100 --> 00:11:03.740
big ideas for you, the listener, who's ready

00:11:03.740 --> 00:11:05.820
to start this journey. I think there are three

00:11:05.820 --> 00:11:08.240
main takeaways. First, just keep it simple. That

00:11:08.240 --> 00:11:10.759
grade five student rule is your North Star. If

00:11:10.759 --> 00:11:12.559
a 10 -year -old can't understand the sentence,

00:11:12.740 --> 00:11:15.159
it's too complicated. Delete the scary words.

00:11:15.440 --> 00:11:17.980
Second, tell your own stories. The AI has all

00:11:17.980 --> 00:11:20.019
the facts, but you're the only one with your

00:11:20.019 --> 00:11:22.580
life experiences. The frustration with the utility

00:11:22.580 --> 00:11:25.080
bill, the joy of finding a great coffee shop,

00:11:25.940 --> 00:11:29.100
those details build trust. And third, follow

00:11:29.100 --> 00:11:32.820
the process. Don't ask for 2 ,000 words. Use

00:11:32.820 --> 00:11:35.159
the four -step workflow and always remember the

00:11:35.159 --> 00:11:38.460
time machine. Ground your writing in real pre

00:11:38.460 --> 00:11:41.059
-AI human language. And the final piece of advice

00:11:41.059 --> 00:11:44.190
is, don't be afraid to make mistakes. Every time

00:11:44.190 --> 00:11:46.450
you find a robotic sentence and fix it, that's

00:11:46.450 --> 00:11:48.970
a win. The reader can feel that effort. At the

00:11:48.970 --> 00:11:51.250
end of the day, AI is just a tool, a powerful

00:11:51.250 --> 00:11:53.730
one, sure. But the heart of the story has to

00:11:53.730 --> 00:11:57.049
be yours. Your simple, real voice is your most

00:11:57.049 --> 00:11:59.710
valuable asset in a world that is just full of

00:11:59.710 --> 00:12:02.129
boring slop. So we encourage you to start your

00:12:02.129 --> 00:12:04.409
slop evader journey today, one sentence at a

00:12:04.409 --> 00:12:06.289
time. Thank you so much for sharing your sources

00:12:06.289 --> 00:12:08.210
and your time with us for this deep dive.
