ChatGPT 5.2: Agentic Future Delegation
Key Points
- GPT‑5.2 is a fundamentally new, “agentic by default” model that can autonomously process massive datasets (e.g., 10 000 rows), perform analyses, and generate finished deliverables like PowerPoints, docs, and Excel files with reliable accuracy.
- The breakthrough lies not just in speed but in the ability to compress work that would normally take six‑to‑eight hours into a 20‑minute run, dramatically reshaping productivity expectations.
- To harness this power, users must master a new skill: precisely defining the scope and desired outputs of a task so the long‑running agent can be delegated work effectively.
- Despite its advanced capabilities, the model cannot replace an entire job; clear, scoped instructions are still essential for meaningful results.
- The improved quality of generated artifacts (e.g., usable PowerPoint slides) marks a leap from earlier versions and signals the kind of competence expected of professionals by 2026.
Sections
- ChatGPT 5.2 – Agentic Data Mastery - The speaker highlights the upcoming ChatGPT 5.2 as a genuinely agentic model that can autonomously process huge datasets, produce polished PowerPoints, Excel reports, and executive narratives, emphasizing the new skill of delegating work to such an AI.
- Effective Prompt Framing for Advanced Models - The speaker stresses that with newer, more capable, agentic models like Chad GPT 5.2, users must clearly define the task and scope—otherwise the model will guess—making precise problem framing a universal skill, illustrated by comparative tests against Gemini 3, Claude Opus 4.5, and ChatGPT 5.1.
- Leveraging Large Context for AI Tasks - The speaker emphasizes that the key advantage of advanced models like ChatGPT 5.2 is their ability to process extensive, varied datasets within a large context window, enabling them to solve complex, data‑intensive problems such as customer service analytics and multi‑source analysis.
- Leveraging GPT‑5.2 for Agentic Workflows - The speaker outlines how to intentionally steer the upcoming 5.2 model, allow it extended processing time, and use it as a flexible, agentic executor for deep analyses such as P&L reviews, acquisition assessments, and personal budgeting, while emphasizing the need to choose the right problems and provide appropriate data.
Full Transcript
# ChatGPT 5.2: Agentic Future Delegation **Source:** [https://www.youtube.com/watch?v=821UqXHineU](https://www.youtube.com/watch?v=821UqXHineU) **Duration:** 00:14:38 ## Summary - GPT‑5.2 is a fundamentally new, “agentic by default” model that can autonomously process massive datasets (e.g., 10 000 rows), perform analyses, and generate finished deliverables like PowerPoints, docs, and Excel files with reliable accuracy. - The breakthrough lies not just in speed but in the ability to compress work that would normally take six‑to‑eight hours into a 20‑minute run, dramatically reshaping productivity expectations. - To harness this power, users must master a new skill: precisely defining the scope and desired outputs of a task so the long‑running agent can be delegated work effectively. - Despite its advanced capabilities, the model cannot replace an entire job; clear, scoped instructions are still essential for meaningful results. - The improved quality of generated artifacts (e.g., usable PowerPoint slides) marks a leap from earlier versions and signals the kind of competence expected of professionals by 2026. ## Sections - [00:00:00](https://www.youtube.com/watch?v=821UqXHineU&t=0s) **ChatGPT 5.2 – Agentic Data Mastery** - The speaker highlights the upcoming ChatGPT 5.2 as a genuinely agentic model that can autonomously process huge datasets, produce polished PowerPoints, Excel reports, and executive narratives, emphasizing the new skill of delegating work to such an AI. - [00:03:37](https://www.youtube.com/watch?v=821UqXHineU&t=217s) **Effective Prompt Framing for Advanced Models** - The speaker stresses that with newer, more capable, agentic models like Chad GPT 5.2, users must clearly define the task and scope—otherwise the model will guess—making precise problem framing a universal skill, illustrated by comparative tests against Gemini 3, Claude Opus 4.5, and ChatGPT 5.1. - [00:08:24](https://www.youtube.com/watch?v=821UqXHineU&t=504s) **Leveraging Large Context for AI Tasks** - The speaker emphasizes that the key advantage of advanced models like ChatGPT 5.2 is their ability to process extensive, varied datasets within a large context window, enabling them to solve complex, data‑intensive problems such as customer service analytics and multi‑source analysis. - [00:11:58](https://www.youtube.com/watch?v=821UqXHineU&t=718s) **Leveraging GPT‑5.2 for Agentic Workflows** - The speaker outlines how to intentionally steer the upcoming 5.2 model, allow it extended processing time, and use it as a flexible, agentic executor for deep analyses such as P&L reviews, acquisition assessments, and personal budgeting, while emphasizing the need to choose the right problems and provide appropriate data. ## Full Transcript
Chad GPT 5.2 time traveled back to see
us here. I am convinced that this is a
model that shows us what the future is
like for 2026. It's not an incremental
upgrade, guys. I know it's positioned
that way, but it's actually got some
capabilities that I haven't seen in
other models that I want to lay out here
so that you understand what they are and
you can figure out for yourself whether
the model is right for you. First and
foremost, this model is agentic by
defaults. So if you think about models
on a range of how long they can run and
execute tasks, this is the first
generally available model where it's
very very easy to get it to do a
tremendous amount of work on a huge
bucket of inputs like a data set with
thousands of rows. I tried it with a
data set with 10,000 rows, right? It can
do all of that, compute against it,
develop insights, come back with a
PowerPoint, come back with a doc, come
back with an Excel spreadsheet, and it
actually works. That means that it's
accurate. It's coherent. It's cogent.
It's thoughtful. It's able to craft an
executive narrative. Guys, the
PowerPoint is not nearly as gnarly as it
was in 5.1 and 5.0. The PowerPoint
artifacts actually work now. It's
wonderful. But this creates a skill
problem for us, doesn't it? Because what
we have to do is we have to figure out
how do we now define work that is ready
to be delegated for that period of time.
And that's a new skill for a lot of us.
For many of us, we have been trying to
figure out how to make these models help
us do our work faster all year long. And
that's been most of the conversation
I've had with folks. Guess what? the
models keep getting better and we have
to keep scaling up. And in this
situation, the skill that we need to
learn whether we're technical or
non-technical is how do we define a
piece of work correctly so that we can
assign it to a longunning agent. That is
what feels like 2026 about chat GPT 5.2.
That's what feels novel, new, and super
interesting. And if you can't define
that work, you are going to be behind
people who can really define it well and
come out with a fullyfledged analysis
from a deep data set or a deep problem
in the code or what have you and then
get an answer that they can use and run
with that would normally have taken them
hours. Because when I take and I'm not
kidding 20 minutes, 30 minutes, 40
minutes on a chat GPT 5.2 two task which
I did today. It's it's really good and
it's better than work that would have
taken me four or five hours to do. And
so it's not just about can it save me 20
minutes. It is understanding that the
model can do in 20 minutes what would
have taken someone six or eight hours to
do and how do you understand that block
of work and give it to the model. Now
you might think if it can do six or
eight hours of work can it just do my
job? The answer is no. It needs clear
scope. When I talk about the skill to
delegate to the model, the first thing
to do is to be able to define what
output you want. A scoped output that
matters. If you want a PowerPoint deck,
it can do that. You have to define what
you want there. If you want a word doc,
it can do that. If you want an Excel, it
can do that. Specify. Be clear about
what you want. You also need to be
really clear about what you need from
the inputs. Especially if you're going
to use that nice big context window and
you're going to put a bunch of stuff in.
Please explain to the model what is in
the box and what you want the model to
do with it. Because if you don't, the
model's going to fill in its best guess
and it's going to try and make it
intelligible as best it can and you may
or may not get what you want. And that
has higher stakes now, doesn't it? One
of the big things that shifted in the
last 6 months is that we are no longer
in a world where instant responses are
the best a model can do. The best a
model can do is often longer running.
And so if you're in a world where the
model can take a while to come back with
a response, you better get it right.
You'd better be correct in your problem
framing. And that's not just an
executive skill set anymore. That's an
everybody's skill set. All of us need to
learn more about framing problems and
chunking problems into scopes of work
that that can fit with a model that is
truly agentic. And the reason I'm
emphasizing that here is because Chad
GPT 5.2 is so widely distributed.
Everybody's going to get it because
everybody has Chad GPT. So, we all need
to learn this. Now, now you might be
wondering, how does this compare to some
of the other models out there? Well, I
want to give you some very specific
comparison notes that I've been seeing
in early testing because I did a cross
analysis where I gave the same
assignment to different models to see
what the quality would look like. I
tested against Gemini 3. I tested
against Claude Opus 4.5. I tested on
Chat GPT 5.1 as well just to see what
the sense of of of a difference is
versus 5.2. I think that I'm getting a
real clear sense of where these
different models stack up. One of the
things that is standing out to me is
that the ergonomics of the model matter
a lot. By ergonomics, I mean how do you
have the full environment around the
model feel comfortable like a good
ergonomic chair so you can use it for
useful work. That's not just comfort.
That's actually value. Specifically,
Gemini 3 has really poor user ergonomics
right now. They have embedded Gemini 3
inside Google products and you can
access Gemini 3 in the developer studio
and you can access Gemini in the mobile
app. But in none of those places is it
easy to throw a bunch of data to throw a
bunch of docs into the model and say
please come out with a fully finished
output. That is just not the product
that Google has built. And so even if
the brain power is there to do
meaningful work against these artifacts
and analyze it and come back with a
fully featured output, you can't get to
it. I could not upload a PowerPoint to
Gemini 3. I could not upload a
PowerPoint to or an Excel or a CSV. It's
just not good, guys. you you you have to
have the ability to put a lot of data in
if you want to do complex work and it's
a problem if you can't do that. And so I
love Gemini 3. I did a great review on
it. I still use it. I love their image
generator. It's a smart model. I use it
for thinking a fair bit. But the
ergonomics are a and they really pop out
when you compare it to Chat GPT 5.2
because Chad GPT 5.2 YouTube will take
anything. Like you can throw anything in
there. It will take it all and it will
just chew on it. You can throw a
screenshot and a CSV and a doc and a
PowerPoint and it will just chew it all
and process it and come out with
something useful. And I think that's
really really helpful. And I think that
one of the things that really popped as
a difference in my test between 5.2
is that the ability to intelligently
coherently with less hallucinations
process this data is way up. And that
showed up in their benchmarks. They saw
like 38% less hallucinations or
something like that. And and it just it
pops like you can see it. You can you
can see the coherence. Now comparing it
to Opus 4.5 is interesting because the
ergonomics in Opus 4.5 are also quite
solid. You can throw in a wide variety
of input documents. I like the way Opus
4.5 is able to craft effective output
artifacts just like Chad GPT 5.2. And so
if I were to look for a difference
between the two, I think the thing that
I want to call out is first the way the
models are architected is very very
different. Chad GPT 5.2 especially in
thinking mode which is a very different
mode. If you're using instant mode, it's
not the same thing. Chad GPT 5.2
thinking mode is a longunning thoughtful
intentional model. It takes a while to
respond. It does very thorough work and
these days it now does artifacts well
too. It really does. there's not really
that gap on PowerPoint functionally
speaking. Opus uses tools instead of
reasoning. And so Opus will work for a
while, but it's using tools as a
non-reasoning model to get that work
done. So it's a very different approach.
I like the aesthetics of the PowerPoint
that Opus 4.5 produces slightly better.
The functionality is about the same.
Like from a functional PowerPoint
narrative perspective, it's about the
same. And critically, the thing that
gives Chat GPT 5.2 to an edge is that it
can take so much data to solve your
problems. And that's why I started this
conversation saying pay attention to how
long these agents can work. Because if
you were going to give an agent a
meaningful task, that only really works
if you trust it with a ton of data. If
you give it a lot of data to work with
and ask it to handle a complex task.
Otherwise, even in thinking mode, it
won't take that long. and you won't have
solved that meaningful a problem. And so
I think the thing that we need to shift
toward is a world where we recognize
that increasingly the models have a
better understanding across larger
swaths of data than we do. So maybe it's
a a customer service set of tickets that
we need to analyze. Maybe it's hundreds
of Twitter responses from a question
that we had. Maybe it's uh a bunch of
Stripe transaction data. Maybe it's a
big Excel spreadsheet of customer
issues. You get the idea, right? It's
anything that has that sort of very
large variagated data types all in one
big place, right? Because you could have
like customer tickets in one hand, you
could have transcripts and recording in
another. The data can be quite
variegated. It's a big enough context
window you can throw it all in there and
you can ask it to make sense of it. And
it does. and it's able to translate it
into something useful. I think this is a
little bit of an intangible, but one of
the things that comes out when you have
a model that is strong at coherence that
reduces on hallucinations that has the
tools to build something like a
PowerPoint well is the ability to build
narrative comes as an emergent property.
And so what I noticed is it's able to
take data that I don't necessarily have
a clear story for and it's able to pull
it in and say there's a story here.
here's the overall story and here's why
I know that and you can check it and
prove it because of course you do. You
have to go in and check and and see that
it actually works. And so if you're
looking at 2026 and you're asking
yourself, what are the skills I need to
thrive? How do I build this into my
teams? I would say the number one skill
that you're going to need in chat GPT
5.2 too and in the other models that
follow not just from Chad Chad GBT but
from Gemini from XAI from Anthropic
you're going to see more agentic models
and your number one skill needs to be
grow my ability to delegate we are
moving from a world where execution with
models was the story for 2025 delegation
to models is going to be the story of
2026
we're not ready we're not ready. We're
not ready with the data side. We're not
ready with the skill side. We don't know
how to frame problems. Look, I the first
thing I did when I started getting into
5.2 and seeing what it could do is I
went over to 5.2 and I asked it to start
to help me think through prompting this
model differently because we have to
think about prompting not as give me a
response now, but as let me give you a
lot of stuff and then go away and think
about it. Now, eventually we're going to
get to a world where I think we have
more interaction patterns with running
agents and you can interrupt the agent.
We're starting to see hints of that.
We'll see more of that in 2026. But the
skill for now is really intentionally
aim the model in the direction you want
it to go and then focus and make sure
you have the right stuff and then give
it give it time to work. Let it work for
a while. It is not unusual to see a
model like 5.2 two work for 20, 30, 40
minutes and it's not like deep research
because deep research comes back and it
just gives you a web report and it's
very well written. It can be 50 pages.
5.2 thinking will come back in a similar
amount of time but it will give you much
more control over what you get. You can
define the output type that you want.
You can define the kind of analysis you
want. It's like a much broader Swiss
Army knife versus the scalpel that is
deep research. So if you're wondering
where to put this into your workflow, I
would say 5.2 to thinking is a agentic
workflow executor that is almost more
powerful than we're ready for. It is
something that if you know how to
delegate well, it is going to eat work
for you. You want to analyze a P&L, let
it let it analyze the P&L for you. Let
it take the first pass. You want to
analyze an acquisition, let it do that.
You want to analyze your investments or
your personal savings and budget, let it
do that. This thing loves to solve
problems. And so really the rate limiter
for us, the question for us is do we
have the taste to find the right
problems to solve? Do can we can we
locate the data for it? Can we throw the
data in and then can we give it clear
enough directions about the output and
the kind of analysis it needs to run in
order to get successful outcomes because
the stakes are higher. Now if you're
running Chad GBT 5.2 too for 20, 30, 40
minutes and you didn't give it the right
directions. Your feedback loop is slow.
You're going to be like, "Oh no, now I
have to redo it. It's going to be like
another hour out of my day to get this
done." So, our prompting skills are now
higher leverage because it's so
important. And so, I put together some
prompts to make sure that we have a good
sense of what this looks like. But
beyond prompting, the key thing that I
want to call out is we need the soft
skills to delegate better, to understand
those problem frames. And that's what I
want to leave you with because I believe
that that is the key skill for 2026. And
I think that is what 5.2 shows us in a
way that no other model does. It will
eat entire workflows because it is so
good at correct coherent longunning
agentic execution. I think they kind of
undersold it as a 0.1 upgrade. I think
it's bigger than that. But you tell me.
You test it out. You tell me what you
think. I'm really curious. I love the
model. It's going to be a lot of fun to
use.