Kind: captions
Language: en
I watched someone spend 20 minutes
reexplaining their entire project to
chat GPT because they started a new
chat. All that context, all those
preferences just gone.
And here's the thing, I was doing the
exact same thing until 2 weeks ago. I've
been using GPT 5.2 daily since it
dropped and I found something that
genuinely shocked me. We're all leaving
about 80% of this tool on the table.
The gap between what you think chatgpt
can do and what it actually does now is
borderline ridiculous. Welcome back to
bitbiased.ai
where we do the research so you don't
have to join our community of AI
enthusiasts with our free weekly
newsletter. Click the link in the
description below to subscribe. You will
get the key AI news tools and learning
resources to stay ahead. So in this
video I'm showing you 30 ways to
actually use what you're already paying
for.
These aren't theory. I've tested every
single one of these in real work over
the past two weeks. You're going to
learn how to make chat GPT remember
everything you've ever told it. How to
feed it 200page documents without it
choking. And how to get answers that
don't sound like they came from a robot.
First thing, chat GPT can now pull up
conversations from 6 months ago, like
the exact chat. Let me show you how that
works. The memory revolution.
Here's something wild. Chat GPT can now
search through every conversation you've
ever had with it. We're not talking
about some vague context feature. GPT
5.2 literally remembers.
Ask it something like, "What did I ask
you about 6 months ago regarding that
essay topic?" And it'll pull up the
exact conversation.
No more scrolling through dozens of
chats trying to find that perfect recipe
or that brilliant idea you had at 2 a.m.
But here's where it gets even better.
If you're on plus or pro, there's the
sources button that appears in
responses.
Click it and chat GPT shows you direct
links to the past conversations it's
referencing. It's like having a personal
archavist who never forgets where you
filed things. You discussed a recipe you
loved 3 weeks ago. Click sources. Jump
straight to that chat. Done.
Personalization that actually works.
Now, I know what you're thinking.
Great. Another settings menu I'll never
use.
But wait, go into your chat GPT settings
and look for custom instructions and the
personality presets. This is where
things get interesting. You can choose
between friendly, professional, candid,
and a few others. Pick professional and
suddenly every response comes back in
formal business style. Switch to
friendly and it's like chatting with a
colleague over coffee. And that's just
the beginning.
Under personalization, there are these
sliders for conciseness, warmth, emoji
usage, the works.
Want bullet list answers that you can
scan in 5 seconds?
Slide it to concise mode. Prefer a
warmer, more conversational vibe with
the occasional emoji? Slide the other
way. You're literally designing how chat
GPT talks to you. The thinking time
breakthrough. Here's something GPT 5.2 2
introduced that changes everything about
how you get answers. Thinking time.
You've got options ranging from light to
heavy. And this isn't just marketing
speak. When you switch to extended or
heavy mode, the model actually takes
longer to process your question. It's
not just stalling. It's genuinely
reasoning through the problem step by
step. Try this. Ask a complex research
question in light mode. Then ask the
same thing in heavy mode.
The difference is night and day. Heavy
mode gives you structured, well-th
thoughtout answers because it's
literally spending more time thinking.
For quick stuff, stick with light. For
anything that matters, go heavy and
watch the quality jump.
Web browsing and real-time data. If
you're on plus or have a plan with
tools, turn on browsing.
Yes, chat GPT can now pull information
from the web. This means you're not
stuck with knowledge from whenever the
model was last trained.
Want the latest release notes for GPT
5.2? Just say using browsing mode, find
the latest chat GPT 5.2 release notes.
It'll actually go fetch that information
for you. And for those of you working
with data, the advanced data analysis
feature, also called the code
interpreter, is a gamecher.
Upload a CSV of your sales data and ask
it to compute monthly totals and create
a bar chart. ChatGpt writes the code,
runs it, and shows you the results. No
spreadsheet software needed.
Handling massive documents.
GPT 5.2's context window is enormous.
We're talking about uploading 50, 100,
even 200page documents and getting
meaningful analysis.
that research paper you've been dreading
reading. Upload the PDF and ask for a
200word summary of the key findings.
That's it. You just saved yourself 3
hours. Better yet, upload multiple files
at once and have Chat GPT compare them.
Two product manuals, three contracts,
whatever. Ask, "What are the main
differences between these documents?"
And let the AI do the heavy lifting.
This alone makes the subscription worth
it if you deal with documents regularly.
Structured output and organization.
Here's a hack that'll save you so much
time. Ask for structured output. Want a
budget spreadsheet? Say, "Make a budget
spreadsheet for a trip to Tokyo with
columns for item, cost, and notes.
Chat GPT will generate the whole thing
in table format. Need a presentation
outline?
Outline a five- slide presentation on
climate change solutions gives you
instant slide content you can copy
straight into PowerPoint and the image
analysis has gotten significantly
better. Upload a chart, a graph,
anything visual and GPT 5.2 can
interpret it. Ask what trends do you see
in this sales line chart over time and
you'll get actual insights, not just
generic descriptions.
Multi-perspective thinking. This one's
subtle but powerful. Instead of asking
chat GPT for a single answer, ask for
multiple perspectives. Try this. Explain
blockchain technology three ways. From a
beginner's perspective, an expert's
perspective, and a skeptic's
perspective.
Suddenly, you're not getting just one
viewpoint. You're getting a
comprehensive understanding. Or flip it
around and have ChatGpt ask you
questions first.
Say before suggesting a vacation spot,
ask me five questions about my travel
preferences. Then give one
recommendation.
This turns the whole interaction into a
conversation rather than a oneshot
query. The personalization makes the
final answer way more useful.
Bulletproof planning. Want answers that
actually work in the real world? Request
plans that include failure scenarios.
Ask ChatGpt to help me plan a backyard
party and identify the top things that
could go wrong like rain or power outage
and how to handle them.
You get not just a plan but
contingencies. It's planning that
actually accounts for reality. And when
you need strict constraints followed, be
explicit. Tell chat GPT. Plan a dinner
for $50 max. No oven or stove. Must be a
full meal. Don't suggest buying kitchen
appliances.
When you set hard boundaries like this,
GPT 5.2 is much better at staying within
them than previous versions.
Code analysis and debugging.
For anyone working with code, this is
huge. Paste in a function that's giving
you trouble and ask, "Here's a Python
function. It's returning the wrong
output. What's the bug?" GPT 5.2's code
reasoning has improved dramatically.
It'll walk through the logic, spot the
issue, and explain what's wrong. You can
also use it to understand code you
didn't write. Found some complex
function in a codebase? Paste it in and
ask for an explanation.
The model breaks it down step by step.
This alone makes it worth having open
while you're developing.
Translation and style shifts.
The translation quality has jumped with
GPT 5.2, too. But more interesting is
the stylistic flexibility. You can say,
"Translate this email into Spanish in a
friendly tone or even rephrase this
paragraph as if it were in a Shakespeare
play."
The model understands not just literal
translation, but tonal and stylistic
shifts.
This extends to all kinds of writing.
Need a formal business email turned
casual? Done. Want that blog post
rewritten for a younger audience? No
problem. It's like having a writing
assistant who can shift between any
style you need. Long- form content
creation. Ask GPT 5.2 to draft an
800word article on any topic and you'll
get something genuinely readable. It's
not perfect. You'll want to edit it, but
it's a solid first draft that captures
the main points and maintains a coherent
structure. for blog posts, reports, even
essay drafts. This saves massive amounts
of time. Brainstorming is another area
where this shines.
Give me 10 unique blog post ideas about
healthy cooking. Boom. You've got ideas
ranging from practical to creative,
enough to fuel your content calendar for
weeks. Memory powered personalization.
With memory turned on, here's where
things get really interesting. Over
time, ChatGpt learns your preferences.
It remembers you're vegan, that you
prefer concise answers, that you're
working on a specific project. This
means future conversations start from a
place of understanding rather than from
zero. Try this. After using it for a
while, ask, "Given what you know about
my vegan diet from our past chats,
suggest dinner recipes."
You'll get suggestions that actually
align with your preferences because the
AI has been paying attention.
Personality experiments. Don't be afraid
to experiment with those personality
presets. Switch between efficient,
friendly, nerdy, even cynical and ask
the same question. You'll be surprised
how much the answers can differ.
Sometimes a different perspective
unlocks a better solution.
And here's a powerful technique.
Role-playing.
ask chat GPT to act as a career coach
and advise me on interviewing.
The model shifts into that role and
gives advice from that specific
perspective.
It's like having access to experts in
different fields whenever you need them.
Iterative refinement.
This is where GPT 5.2 really outperforms
older versions. After you get an answer,
just say make it shorter or make it more
detailed or rewrite this in first
person. Each iteration gets you closer
to exactly what you need.
You're not starting from scratch, you're
refining.
For long conversations, ask it to
summarize.
Summarize our chat in five bullet points
so far. This keeps you oriented.
Especially useful in complex
problem-solving sessions where you've
gone down several paths.
Combining hacks for power moves. The
real power comes from combining these
techniques.
Upload a data file, set the thinking
mode to heavy, ask clarifying questions,
and request the output as a bullet list.
You've just created a workflow that
turns raw data into an actionable report
in minutes. Each of these hacks
leverages what makes GPT 5.2 special.
The massive context window, improved
reasoning, better vision capabilities,
and significantly reduced
hallucinations.
The model is about 30% more reliable
than GPT 5.1, which means you can trust
it more for coding help, complex
planning, and detailed analysis.
Though, as always, double check anything
critical.
The key takeaway, stop using chat GPT
like a fancy search engine.
It's a personalized assistant that
remembers context, handles complex
tasks, and adapts to your style. These
30 hacks are your road map to actually
using the tool the way it was built to
be used.
Start with memory and personalization.
Add in the thinking modes and file
uploads. Experiment with the different
features and watch how much more useful
your AI assistant becomes. That's it for
this one. If you found this helpful, let
me know which hack you're most excited
to try in the comments. And if you want
more deep dives on AI tools that
actually work, you know what to do. See
you in the next one.