WEBVTT

00:00:00.000 --> 00:00:03.149
What if the key to unlocking you know, the true

00:00:03.149 --> 00:00:05.849
potential of advanced AI? What if it isn't about

00:00:05.849 --> 00:00:08.369
paying more for the fastest model or even getting

00:00:08.369 --> 00:00:10.529
better hardware? What if the difference between

00:00:10.529 --> 00:00:13.169
just a generic answer and a really expert level

00:00:13.169 --> 00:00:16.469
insight rests on just a few specific, simple

00:00:16.469 --> 00:00:19.149
phrases? We're talking today about how to force

00:00:19.149 --> 00:00:22.329
the AI to slow down, to think deeply, and to

00:00:22.329 --> 00:00:25.250
actually give you its absolute best work. And

00:00:25.250 --> 00:00:27.089
this is the kicker, even if you're working entirely

00:00:27.089 --> 00:00:29.829
on the free plan. Welcome to the Deep Dive. Look,

00:00:29.910 --> 00:00:32.149
if you're like most people, You might be using

00:00:32.149 --> 00:00:34.630
something like ChatGPT the same way you'd use

00:00:34.630 --> 00:00:36.850
a slightly smarter Google. You ask a simple question,

00:00:36.950 --> 00:00:39.429
you get a decent but generic answer, and then

00:00:39.429 --> 00:00:41.409
you just move on. Yeah, it's transactional. Exactly.

00:00:41.729 --> 00:00:44.369
Today, we are going to completely transform that

00:00:44.369 --> 00:00:47.350
habit. We're doing a deep dive into 21 high -leverage

00:00:47.350 --> 00:00:51.049
techniques that turn this tool from a basic chatbot

00:00:51.049 --> 00:00:54.609
into a persistent, organized, and truly customized

00:00:54.609 --> 00:00:57.000
intellectual workspace. Our mission today is

00:00:57.000 --> 00:00:59.399
pretty straightforward. We want to help you skip

00:00:59.399 --> 00:01:01.740
past that initial learning curve and move from

00:01:01.740 --> 00:01:04.799
being a typical question and answer user into

00:01:04.799 --> 00:01:07.280
someone who really understands how to architect

00:01:07.280 --> 00:01:10.420
better results. We'll cover advanced prompt architecture,

00:01:10.780 --> 00:01:13.459
some secret organizational features, and how

00:01:13.459 --> 00:01:16.079
you can essentially build many applications without

00:01:16.079 --> 00:01:18.700
ever touching a line of code. That's a great

00:01:18.700 --> 00:01:21.099
roadmap. Let's start at the most fundamental

00:01:21.099 --> 00:01:25.519
level. How you structure your input and... maybe

00:01:25.519 --> 00:01:28.260
more critically, how you manage the AI's actual

00:01:28.260 --> 00:01:31.099
reasoning process. Okay, so the core frustration

00:01:31.099 --> 00:01:34.409
we always hear from people is just... They get

00:01:34.409 --> 00:01:36.390
these underwhelming results because their prompts

00:01:36.390 --> 00:01:38.790
are often too vague or too basic. Or just plain

00:01:38.790 --> 00:01:41.250
poor. Yeah. Right. And that's usually the case.

00:01:41.430 --> 00:01:44.189
It's rarely the model failing. It's almost always

00:01:44.189 --> 00:01:46.930
a lack of structured guidance. Now, there's an

00:01:46.930 --> 00:01:48.590
immediate shortcut for this. It's called prompt

00:01:48.590 --> 00:01:50.489
packs. Oh, I found these to be high leverage

00:01:50.489 --> 00:01:52.769
right away. These are basically custom prompts

00:01:52.769 --> 00:01:55.370
that OpenAI itself developed for specific professional

00:01:55.370 --> 00:01:57.129
roles. They have them for sales, engineering,

00:01:57.430 --> 00:02:00.049
finance, marketing. So if you're running a marketing

00:02:00.049 --> 00:02:03.819
strategy, you can grab a pack. with, say, 20

00:02:03.819 --> 00:02:06.540
or 30 highly structured prompts for campaign

00:02:06.540 --> 00:02:09.379
planning or market research. You just click a

00:02:09.379 --> 00:02:11.719
link, the prompt is preloaded with all the structure

00:02:11.719 --> 00:02:14.199
it needs, and it's just waiting for your details.

00:02:14.439 --> 00:02:18.039
It's an instant fix for prompt quality. So that

00:02:18.039 --> 00:02:20.840
handles the structure problem. But let's talk

00:02:20.840 --> 00:02:22.979
about the mechanics. This is the free hack that

00:02:22.979 --> 00:02:25.780
really pays off that hook we started with. Trigger

00:02:25.780 --> 00:02:28.219
words. Right. Most people in the free tier are

00:02:28.219 --> 00:02:30.909
just letting the model run in auto mode. And

00:02:30.909 --> 00:02:33.069
what that means is it's prioritizing speed and

00:02:33.069 --> 00:02:35.969
efficiency over deep thinking. It's taking shortcuts.

00:02:36.110 --> 00:02:38.090
It's giving you the fastest answer, not the best

00:02:38.090 --> 00:02:40.689
one. Exactly. We need to force some quality control.

00:02:40.889 --> 00:02:42.990
And you don't get the quality settings of a paid

00:02:42.990 --> 00:02:46.069
plan, but you can sort of reroute the prompt

00:02:46.069 --> 00:02:49.789
internally. By adding specific, simple phrases,

00:02:49.949 --> 00:02:52.610
these trigger words, you force the model to use

00:02:52.610 --> 00:02:54.889
its deeper, more careful reasoning models, even

00:02:54.889 --> 00:02:57.129
in the base version. Yeah, the coup is to convey

00:02:57.129 --> 00:03:01.020
urgency. Necessity. Phrases like, think deeply,

00:03:01.139 --> 00:03:04.479
step by step, or be extremely thorough, this

00:03:04.479 --> 00:03:07.259
is critical to get right. These consistently

00:03:07.259 --> 00:03:10.580
force the AI to route your prompt to a high -quality

00:03:10.580 --> 00:03:13.139
process. It's basically a free upgrade to the

00:03:13.139 --> 00:03:15.520
model's intellectual rigor. It changes the internal

00:03:15.520 --> 00:03:18.120
clock of the machine. Okay, so let's talk about

00:03:18.120 --> 00:03:20.759
feedback. This is another area where users get

00:03:20.759 --> 00:03:24.180
stuck. If you ask ChatGPT to critique your own

00:03:24.180 --> 00:03:27.900
work, it tends to be, well... a little too agreeable

00:03:27.900 --> 00:03:30.400
and I'll admit I still wrestle with pump drift

00:03:30.400 --> 00:03:32.939
myself especially when I'm trying to get honest

00:03:33.909 --> 00:03:35.789
unvarnished critique on a draft I've written.

00:03:35.870 --> 00:03:37.669
Yeah, that core programming of being helpful

00:03:37.669 --> 00:03:40.430
can be genuinely detrimental when you need an

00:03:40.430 --> 00:03:42.590
objective, critical evaluation. It doesn't want

00:03:42.590 --> 00:03:44.370
to hurt your feelings, which is kind of ridiculous,

00:03:44.629 --> 00:03:46.990
but that is the behavior. So we use the impersonal

00:03:46.990 --> 00:03:48.830
critique trick. You just reframe the source of

00:03:48.830 --> 00:03:50.229
the material completely. Don't say it's yours.

00:03:50.389 --> 00:03:52.889
You present the work as being from an impersonal

00:03:52.889 --> 00:03:55.810
source, maybe a stubborn colleague or an anonymous

00:03:55.810 --> 00:03:57.830
submission you were asked to review. Exactly.

00:03:58.300 --> 00:04:00.500
You say, this business strategy was anonymously

00:04:00.500 --> 00:04:02.879
submitted, and I have serious concerns about

00:04:02.879 --> 00:04:05.479
the assumptions made in segment three. Help me

00:04:05.479 --> 00:04:08.199
critically evaluate it. Ah, I see. So by putting

00:04:08.199 --> 00:04:11.139
that distance between you and the work, the AI

00:04:11.139 --> 00:04:13.939
flips its helpful nature. It's now helping you

00:04:13.939 --> 00:04:16.360
by critiquing the other material. And you get

00:04:16.360 --> 00:04:19.620
that genuinely critical, detailed review you

00:04:19.620 --> 00:04:22.360
needed all along. It's a subtle shift in framing,

00:04:22.480 --> 00:04:24.779
but one of the most powerful hacks there is.

00:04:24.939 --> 00:04:27.139
Okay, that covers getting honest critique. But

00:04:27.139 --> 00:04:28.819
what about when you're just starting out? You

00:04:28.819 --> 00:04:30.540
know, you need to gather context for a brand

00:04:30.540 --> 00:04:33.339
new request. We often spend ages writing up a

00:04:33.339 --> 00:04:35.680
huge context document. Right. Stop struggling

00:04:35.680 --> 00:04:38.279
to provide all the context yourself. Just use

00:04:38.279 --> 00:04:40.839
the hack of asking the AI to ask you clarifying

00:04:40.839 --> 00:04:43.579
questions. If you just say, create a weekly newsletter

00:04:43.579 --> 00:04:46.920
plan, that's way too vague. But if you tell it,

00:04:46.939 --> 00:04:49.120
before starting, ask me the top five clarifying

00:04:49.120 --> 00:04:51.060
questions you need to tailor this to my audience

00:04:51.060 --> 00:04:54.779
and my three -month goals, the output is completely

00:04:54.779 --> 00:04:57.980
different. So the AI, now in its role as a sharp,

00:04:58.060 --> 00:05:00.560
strategic partner, it will generate questions

00:05:00.560 --> 00:05:02.379
you probably hadn't even thought of yourself.

00:05:02.819 --> 00:05:05.220
And once you answer them, its advice becomes

00:05:05.220 --> 00:05:07.980
surgical. And finally, after all this work, you

00:05:07.980 --> 00:05:10.139
went back and forth, I don't know, 15 times to

00:05:10.139 --> 00:05:12.639
get the perfect response. Don't lose that effort.

00:05:13.160 --> 00:05:15.639
This is where prompt discolation comes in. Right.

00:05:15.740 --> 00:05:18.000
That long, iterative process you just went through

00:05:18.000 --> 00:05:20.240
contains all the perfect, optimized guidance

00:05:20.240 --> 00:05:23.259
you gave the AI over several steps. So you just

00:05:23.259 --> 00:05:26.019
asked the AI to synthesize that entire conversation

00:05:26.019 --> 00:05:29.620
into one single reusable prompt that would have

00:05:29.620 --> 00:05:32.240
produced that same result on the first try. It

00:05:32.240 --> 00:05:34.139
saves you a massive amount of time on future

00:05:34.139 --> 00:05:36.420
projects. So let's focus just on the free tier

00:05:36.420 --> 00:05:39.699
user for a moment. How can they ensure high quality

00:05:39.699 --> 00:05:42.860
output every single time they hit enter? By consistently

00:05:42.860 --> 00:05:46.250
using trigger words. You force the AI to employ

00:05:46.250 --> 00:05:48.689
those deeper, higher -quality reasoning models.

00:05:48.949 --> 00:05:51.029
Okay, so we have to address the idea of persistence.

00:05:51.290 --> 00:05:54.170
If you are still using ChatGPT as just a series

00:05:54.170 --> 00:05:56.850
of isolated, disposable chats, you're missing

00:05:56.850 --> 00:05:59.250
the single biggest shift in its utility, projects.

00:05:59.709 --> 00:06:02.029
Projects are the ultimate organization engine.

00:06:02.189 --> 00:06:05.069
They just recently became a core, free feature,

00:06:05.170 --> 00:06:07.810
and they move the AI from being a simple chat

00:06:07.810 --> 00:06:10.889
tool to a persistent, organized, and systematic

00:06:10.889 --> 00:06:14.910
workspace. Yeah, think of them as organized folders

00:06:14.910 --> 00:06:16.750
for your intellectual life. You could have a

00:06:16.750 --> 00:06:18.490
folder for fitness tracking, one for finance

00:06:18.490 --> 00:06:20.670
planning, another one dedicated to your long

00:06:20.670 --> 00:06:22.589
-form writing. But it's so much more than just

00:06:22.589 --> 00:06:26.269
folders. The real power is in the memory that's

00:06:26.269 --> 00:06:28.970
tied to that specific project folder. You can

00:06:28.970 --> 00:06:31.240
upload reference files. Things like your brand

00:06:31.240 --> 00:06:33.939
style guides, your company's SOPs or your full

00:06:33.939 --> 00:06:37.019
portfolio history. And these files serve as permanent

00:06:37.019 --> 00:06:39.800
context. And that's the key. It is. Crucially,

00:06:39.860 --> 00:06:41.939
every chat you start with in that project folder

00:06:41.939 --> 00:06:44.379
automatically references those files. You never

00:06:44.379 --> 00:06:46.920
have to repeat that core information again. It's

00:06:46.920 --> 00:06:48.899
like assigning a dedicated staff member to that

00:06:48.899 --> 00:06:51.029
specific topic. and that leads right into the

00:06:51.029 --> 00:06:53.449
second layer here which is custom instructions

00:06:53.449 --> 00:06:56.910
per project unlike the universal custom instructions

00:06:56.910 --> 00:06:59.629
that apply to everything these settings only

00:06:59.629 --> 00:07:02.689
apply within that specific folder so this lets

00:07:02.689 --> 00:07:05.629
you define a highly specific role and tone just

00:07:05.629 --> 00:07:07.889
for that project for example in your writing

00:07:07.889 --> 00:07:11.149
project the instruction could be act as a ruthless

00:07:11.149 --> 00:07:14.170
copy editor with zero tolerance for passive voice

00:07:14.170 --> 00:07:17.069
right and to set this up really efficiently don't

00:07:17.069 --> 00:07:19.689
just name the project writing Name it after the

00:07:19.689 --> 00:07:22.490
outcome, like weekly newsletter engine. Then

00:07:22.490 --> 00:07:24.470
you upload those style guides. And then here's

00:07:24.470 --> 00:07:26.589
a pro tip. Instead of writing the custom instructions

00:07:26.589 --> 00:07:29.670
yourself, ask ChatGPT to give you a series of

00:07:29.670 --> 00:07:31.689
detailed questions to gather context about the

00:07:31.689 --> 00:07:34.310
project. Answer those questions and then ask

00:07:34.310 --> 00:07:36.529
the AI to rewrite your answers into its official.

00:07:36.910 --> 00:07:38.970
custom instructions. Wow. So you're optimizing

00:07:38.970 --> 00:07:41.410
the brain of your project using the tool itself.

00:07:41.610 --> 00:07:43.949
You got it. That is a fundamental workflow optimization.

00:07:44.509 --> 00:07:46.550
This whole structure, though, it connects really

00:07:46.550 --> 00:07:49.430
deeply to the idea of system prompts. We should

00:07:49.430 --> 00:07:51.230
probably clarify the difference between custom

00:07:51.230 --> 00:07:53.569
instructions and system prompts. Good point.

00:07:53.990 --> 00:07:56.509
So custom instructions are the permanent rules

00:07:56.509 --> 00:07:59.569
for the project folder. A system prompt is a

00:07:59.569 --> 00:08:01.750
temporary instruction that says the model's behavior

00:08:01.750 --> 00:08:04.810
before a specific task begins, just within a

00:08:04.810 --> 00:08:07.660
single chat thread. Most users only write the

00:08:07.660 --> 00:08:11.079
user prompt, you know, analyze this data. The

00:08:11.079 --> 00:08:13.959
power user operates on the system layer. You

00:08:13.959 --> 00:08:15.980
write the system prompt first, which dictates

00:08:15.980 --> 00:08:18.160
the lens through which the analysis will be performed.

00:08:18.420 --> 00:08:20.540
Exactly. For example, instead of just asking

00:08:20.540 --> 00:08:22.540
for a competitive analysis, you start with a

00:08:22.540 --> 00:08:25.500
system prompt like. You are a senior, highly

00:08:25.500 --> 00:08:27.939
skeptical financial analyst who is deeply detail

00:08:27.939 --> 00:08:31.279
oriented and prioritizes real world market constraints

00:08:31.279 --> 00:08:34.340
over theoretical optimism. That is fantastic.

00:08:34.620 --> 00:08:36.620
It changes the entire reasoning frame. Every

00:08:36.620 --> 00:08:39.080
single response after that follows the tone and

00:08:39.080 --> 00:08:41.360
the decision logic you set. You're not fixing

00:08:41.360 --> 00:08:43.639
bad answers with longer questions. You're fixing

00:08:43.639 --> 00:08:45.720
them by intentionally changing how the model

00:08:45.720 --> 00:08:48.379
thinks. And just a minor hack here, because these

00:08:48.379 --> 00:08:50.740
persistent threads can sometimes get lost in

00:08:50.740 --> 00:08:52.879
your chat history. If you remember the topic,

00:08:53.059 --> 00:08:55.539
you can ask a current chat to recall memories

00:08:55.539 --> 00:08:59.419
from a past conversation. It's a small but useful

00:08:59.419 --> 00:09:02.639
feature when you need that lost context. So if

00:09:02.639 --> 00:09:04.960
the listener only takes one thing from this segment,

00:09:05.100 --> 00:09:07.759
what's the core advantage of setting up projects

00:09:07.759 --> 00:09:10.860
with custom instructions? It ensures consistent

00:09:10.860 --> 00:09:13.740
behavior and a defined role across every single

00:09:13.740 --> 00:09:16.539
conversation related to that specific outcome.

00:09:16.909 --> 00:09:19.029
All right, let's pivot now to functionality that

00:09:19.029 --> 00:09:21.470
really speeds up your workflow, things that eliminate

00:09:21.470 --> 00:09:24.909
manual steps. If you are doing any kind of heavy

00:09:24.909 --> 00:09:26.870
writing or structured editing, you have to know

00:09:26.870 --> 00:09:30.210
about Canvas. Canvas is essentially a specialized

00:09:30.210 --> 00:09:32.470
document editor right inside the AI environment.

00:09:32.809 --> 00:09:35.190
You ask the AI to open a document in Canvas,

00:09:35.350 --> 00:09:37.389
and you can instantly start targeting specific

00:09:37.389 --> 00:09:39.950
sections for high -level changes. This is great

00:09:39.950 --> 00:09:42.149
for granular work. You can manually edit the

00:09:42.149 --> 00:09:44.370
text, or you can just highlight a paragraph and

00:09:44.370 --> 00:09:46.490
prompt for a targeted change like... convert

00:09:46.490 --> 00:09:48.830
this paragraph into a detailed timeline table

00:09:48.830 --> 00:09:51.590
with three columns. And this is the part that

00:09:51.590 --> 00:09:54.429
saves so much time. The formatting is preserved

00:09:54.429 --> 00:09:57.110
perfectly when you export the file. You get a

00:09:57.110 --> 00:10:00.409
clean PDF, a Microsoft Word document, or a markdown

00:10:00.409 --> 00:10:02.769
file without losing any of the structure or having

00:10:02.769 --> 00:10:05.529
to repaste and reformat everything. But the real

00:10:05.529 --> 00:10:08.070
shift in leverage, I think, comes with the ability

00:10:08.070 --> 00:10:11.009
to build many apps without needing any coding

00:10:11.009 --> 00:10:13.789
skills at all. And you can integrate this right

00:10:13.789 --> 00:10:16.429
into your projects. This is truly mind -boggling.

00:10:16.529 --> 00:10:18.929
You take that workout tracker example. You just

00:10:18.929 --> 00:10:21.000
describe what you want. You say... Build me a

00:10:21.000 --> 00:10:23.039
fully functional workout tracker website with

00:10:23.039 --> 00:10:25.720
session history, analytics, and an adjustable

00:10:25.720 --> 00:10:27.679
rest timer. And without writing a single piece

00:10:27.679 --> 00:10:30.600
of code or defining the interface, ChatGPT will

00:10:30.600 --> 00:10:32.740
write the necessary code, create a functional

00:10:32.740 --> 00:10:35.019
interface, and include the adjustable timers

00:10:35.019 --> 00:10:37.440
and analytics. You get a usable tool right there.

00:10:37.620 --> 00:10:41.590
Whoa. Imagine scaling that, building customized

00:10:41.590 --> 00:10:44.169
internal tools for completer business processes,

00:10:44.309 --> 00:10:46.529
like a specialized inventory management system

00:10:46.529 --> 00:10:49.149
or a dynamic sales calculator for your team.

00:10:49.250 --> 00:10:51.529
That's not just leverage, that's real automation.

00:10:51.889 --> 00:10:54.789
Moving on to just pure speed gains, formatting

00:10:54.789 --> 00:10:58.259
and file extraction. We spend so much time copying

00:10:58.259 --> 00:11:01.059
tables out of PDFs and pasting them into spreadsheets.

00:11:01.100 --> 00:11:04.080
Just stop. Yeah, stop doing that. You can now

00:11:04.080 --> 00:11:07.539
request direct exports into specific downloadable

00:11:07.539 --> 00:11:11.419
file types right from the chat. CSV, JSON, HTML,

00:11:11.840 --> 00:11:15.039
PowerPoint, even SVG. It eliminates all that

00:11:15.039 --> 00:11:17.529
annoying manual reformatting. And the extraction

00:11:17.529 --> 00:11:20.350
is incredibly powerful. You can upload a 50 -page

00:11:20.350 --> 00:11:23.409
PDF with four different tables and just ask for

00:11:23.409 --> 00:11:25.830
the output as a downloadable CSV. Or upload a

00:11:25.830 --> 00:11:28.110
PDF with a bunch of images and ask the AI to

00:11:28.110 --> 00:11:30.029
extract all of them into a single file. It's

00:11:30.029 --> 00:11:32.730
instantaneous format conversion. Right. And finally,

00:11:32.750 --> 00:11:34.690
we should just briefly cover app connectors.

00:11:35.090 --> 00:11:37.409
If you enable them in your settings, though they

00:11:37.409 --> 00:11:39.429
often require a paid tier for full functionality,

00:11:39.809 --> 00:11:42.309
they let the AI interact directly with external

00:11:42.309 --> 00:11:44.690
software. The Canva integration is the clearest

00:11:44.690 --> 00:11:47.779
example of this. You can prompt, create an engaging

00:11:47.779 --> 00:11:50.740
pitch deck for my business idea. ChatGPT generates

00:11:50.740 --> 00:11:53.500
the core copy, uses the connector to create visuals

00:11:53.500 --> 00:11:55.879
based on your prompt, and then gives you an editable

00:11:55.879 --> 00:11:58.000
design link you open directly in Canva. It's

00:11:58.000 --> 00:12:00.919
integrated workflow automation. So beyond just

00:12:00.919 --> 00:12:03.500
asking simple questions, how do these tools really

00:12:03.500 --> 00:12:05.840
save us the most time? They eliminate manual

00:12:05.840 --> 00:12:09.059
reformatting and allow for complex visual workflow

00:12:09.059 --> 00:12:12.929
automation that bypasses manual coding. All right.

00:12:12.950 --> 00:12:14.850
Let's quickly wrap up with a few different interaction

00:12:14.850 --> 00:12:17.169
modes and how you can access specialized expertise.

00:12:17.730 --> 00:12:20.629
First, two quick keyboard shortcuts for speed

00:12:20.629 --> 00:12:23.049
are worth remembering. Commander Control plus

00:12:23.049 --> 00:12:25.710
Shift plus O gets you a new chat instance immediately.

00:12:26.049 --> 00:12:28.889
And Commander Control plus Shift plus S opens

00:12:28.889 --> 00:12:31.350
and closes the sidebar. Right. Simple time savers.

00:12:31.470 --> 00:12:34.070
Then we have these incredible high utility features,

00:12:34.409 --> 00:12:37.210
voice and vision modes. Voice mode is just fantastic

00:12:37.210 --> 00:12:39.820
for role playing or things way beyond. basic

00:12:39.820 --> 00:12:42.360
conversation. You can use it to practice a difficult

00:12:42.360 --> 00:12:45.139
conversation or run through a complex salary

00:12:45.139 --> 00:12:48.720
negotiation with real -time feedback or use it

00:12:48.720 --> 00:12:50.399
for specialized language learning where you get

00:12:50.399 --> 00:12:52.620
immediate conversational correction. It completely

00:12:52.620 --> 00:12:55.620
changes the dynamic. And paired with voice is

00:12:55.620 --> 00:12:58.629
mobile vision. Using your phone's camera in real

00:12:58.629 --> 00:13:01.009
time opens up a whole new world of possibilities.

00:13:01.230 --> 00:13:03.830
You can point your camera at, say, a complicated

00:13:03.830 --> 00:13:06.009
piece of equipment you're trying to set up, a

00:13:06.009 --> 00:13:09.389
new router maybe, and ask the AI for guidance

00:13:09.389 --> 00:13:12.649
on which port is which. Or point it at complex

00:13:12.649 --> 00:13:15.009
homework diagrams and ask for immediate visual

00:13:15.009 --> 00:13:17.350
help. Now, on to the specialized assistants,

00:13:17.809 --> 00:13:21.190
custom GPTs. And while building your own requires

00:13:21.190 --> 00:13:24.440
a paid plan, remember this. Using the thousands

00:13:24.440 --> 00:13:27.379
of existing pre -built GPTs is completely free.

00:13:27.720 --> 00:13:30.019
You just find them via the Explore tab. Yeah,

00:13:30.059 --> 00:13:31.679
and these aren't just chats. They're specialized

00:13:31.679 --> 00:13:33.899
versions of the AI, often trained on specific

00:13:33.899 --> 00:13:37.220
documentation or optimized for a certain complex

00:13:37.220 --> 00:13:40.620
task, like the NAN workflow assistant or a focused

00:13:40.620 --> 00:13:43.159
research generator. And the crucial, really advanced

00:13:43.159 --> 00:13:46.179
hack here is how you can summon GPTs. You don't

00:13:46.179 --> 00:13:48.200
have to leave your current ongoing conversation

00:13:48.200 --> 00:13:50.240
to use them. You just type the at symbol and

00:13:50.240 --> 00:13:52.399
start typing the GPT name you want to use. This

00:13:52.399 --> 00:13:54.679
is a massive workflow change because it allows

00:13:54.679 --> 00:13:57.320
you to stack functionality. You could be deep

00:13:57.320 --> 00:14:00.299
in a technical draft in one chat, summon a specialized

00:14:00.299 --> 00:14:03.240
editing GPT to review just a single paragraph,

00:14:03.480 --> 00:14:06.120
and then seamlessly continue your original conversation.

00:14:06.559 --> 00:14:10.500
No losing context, no opening new tabs. And we've

00:14:10.500 --> 00:14:13.200
saved the last couple of points for the brief

00:14:13.200 --> 00:14:15.399
paid features summary because they do represent

00:14:15.399 --> 00:14:17.600
significant leverage if you decide to take the

00:14:17.600 --> 00:14:20.399
plunge. Right. The data analysis feature, for

00:14:20.399 --> 00:14:23.059
those on the paid plan, is essentially a full

00:14:23.059 --> 00:14:25.759
-fledged data scientist in a box. You upload

00:14:25.759 --> 00:14:28.720
large data sets, CSVs, Excel files, whatever.

00:14:28.919 --> 00:14:32.220
And the AI processes the data, identifies patterns,

00:14:32.340 --> 00:14:34.519
and generates multiple visualization options

00:14:34.519 --> 00:14:36.820
like interactive charts and graphs. And of course,

00:14:36.840 --> 00:14:39.399
the ability to build custom GPTs yourself is

00:14:39.399 --> 00:14:42.000
the ultimate time saver. You create reusable,

00:14:42.179 --> 00:14:44.620
personalized AI workers tailored to specific,

00:14:44.759 --> 00:14:47.500
highly repetitive tasks like creating a unique

00:14:47.500 --> 00:14:49.940
invoice based on your internal formatting just

00:14:49.940 --> 00:14:51.779
by interacting with the builder using natural

00:14:51.779 --> 00:14:54.320
language. So thinking about maximizing flow.

00:14:54.669 --> 00:14:56.009
What's the most important takeaway for people

00:14:56.009 --> 00:14:58.529
regarding these specialized GPTs? Remember that

00:14:58.529 --> 00:15:00.590
you can stack the expertise of those pre -built

00:15:00.590 --> 00:15:02.529
specialized assistants right into your current

00:15:02.529 --> 00:15:05.659
workflow using the simple at symbol. You know,

00:15:05.679 --> 00:15:07.480
if you begin implementing even five of these

00:15:07.480 --> 00:15:09.720
techniques today, you will immediately be operating

00:15:09.720 --> 00:15:11.759
at a level that's far beyond the vast majority

00:15:11.759 --> 00:15:14.200
of users. These features aren't gimmicks. They

00:15:14.200 --> 00:15:15.879
fundamentally change your frame of reference.

00:15:16.080 --> 00:15:18.500
You have to stop just asking questions and instead

00:15:18.500 --> 00:15:21.200
start building consistent systems. The key to

00:15:21.200 --> 00:15:23.539
moving it to that top tier of power users is

00:15:23.539 --> 00:15:26.100
intentional setup, especially using projects

00:15:26.100 --> 00:15:29.220
and system prompts to explicitly define the AI's

00:15:29.220 --> 00:15:32.019
role and behavior before it ever generates a

00:15:32.019 --> 00:15:34.340
single line of content. That is the crucial shift.

00:15:35.030 --> 00:15:37.370
intentionality so my challenge to you is to start

00:15:37.370 --> 00:15:40.330
small today create one simple project right now

00:15:40.330 --> 00:15:42.669
maybe it's just a dedicated folder for learning

00:15:42.669 --> 00:15:45.269
a new skill and maybe use that pro technique

00:15:45.269 --> 00:15:47.669
we talked about of having the ai rewrite its

00:15:47.669 --> 00:15:50.269
own instructions to maximize its efficiency from

00:15:50.269 --> 00:15:52.850
the very start the real discovery here is mapping

00:15:52.850 --> 00:15:54.929
these new capabilities to your own unique high

00:15:54.929 --> 00:15:57.649
value workflows what repetitive task could you

00:15:57.649 --> 00:15:59.789
turn into a mini automation today that's where

00:15:59.789 --> 00:16:02.370
you're going to find true lasting leverage out

00:16:02.370 --> 00:16:02.990
to your own music