22 April 2025

ChatGPT o3 Deep Dive: 3 Wild New Use Cases

2Frameworks

7Insights

Frameworks in this episode

Strategy6 steps

Framework-to-Business Application Workflow

Turn an abstract business framework into a practical decision map

Read →

Marketing6 steps

Multi-Model Creative Campaign Pipeline

Pair a reasoning model with an image model to build coherent campaign assets

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Hot Take· 1

Hot Take04:30

Why the AI Era May Reward Generalists

The host argues that AI may shift advantage from narrow specialists toward people who combine knowledge across fields. He connects this prediction to Sam Altman's comparison between the AI revolution and the Renaissance rather than the specialization-heavy Industrial Revolution.

The Renaissance rewarded broad creative capability
The Industrial Revolution intensified specialization
AI may handle more bounded specialist work
Humans may gain leverage by connecting disciplines

“One of the things I believe, is the AI era, the generalists are going to win.”
04:30

“I think the AI revolution is going to look a lot more like the Renaissance than the Industrial Revolution.”
Sam Altman · 04:30

#generalists#specialization#ai era#renaissance

Explainer· 1

Explainer05:00

Kind vs. Wicked Problems: Where Specialization and AI Fit

David Epstein's distinction separates stable, rule-bound environments from complex situations whose rules and feedback can change. The host hypothesizes that AI excels at kind problems while humans retain an advantage in wicked ones, although collaboration can outperform either alone.

Kind problems have limited scope and clear rules
Wicked problems are complex, gray, and changeable
Specialization performs well in kind environments
Human-machine teams can exceed either participant alone

“specialization is really good for kind problems, things that have a very limited scope and clear sets of rules and outcomes.”
05:00

“Humans are really good at Wicked problems.”
05:30

#kind problems#wicked problems#range#human-ai collaboration

Story· 2

Story01:30

o3 Crops and Zooms Instead of Giving Up on an Image

Dan Shipper tested o3 with a distant photograph containing a notebook near a keyboard. Rather than simply saying the text was unreadable, the model repeatedly cropped and zoomed the image until it inferred the song title was Moon River.

The source text was too distant to read directly
o3 used iterative cropping and zooming
The model adapted its process when the initial view was insufficient

“I've never seen any AI model being able to say, Hey, I can't see this. So let me crop and zoom right.”
02:00

#o3#computer vision#reasoning

Story11:30

Using o3 to Name the Post-Inbound Era of Marketing

The host supplied roughly ten pages of thinking about marketing's future and asked o3 to propose names for the next era. The model clustered the material around themes such as resonance, taste, immediacy, and personalization, then compared its shortlist with the popular phrase “vibe marketing.”

The model synthesized a long document of raw ideas
It grouped names by underlying strategic themes
It produced a heat-check map and shortlist
It critiqued “vibe marketing” as catchy but potentially vague

“vibe marketing is catchy, but it carries a soft drink commercial fuzziness that can feel ephemeral or even vague.”
13:00

“we were able to work through all of this in a few minutes, versus like what probably would have taken days or weeks with a…”
13:30

#marketing strategy#naming#thought partner#o3

Tool· 2

Tool02:30

Turn a Week of Meeting Transcripts Into Leadership Feedback

Dan Shipper combined transcripts from a week of meetings and asked o3 for a comprehensive leadership review. The model identified subtle conflict-avoidance behavior that conventional management examples might overlook.

AI note takers can supply a longitudinal record of meetings
A week of transcripts provides patterns beyond one conversation
The analysis surfaced subtle conflict avoidance
The output can function as personalized leadership coaching

“What he did is he took all of his transcripts from a week of meetings and put them into 03 and basically asked it to…”
02:30

“he really avoided conflict in meetings in really subtle ways.”
03:00

#leadership#meeting transcripts#coaching#o3

Tool03:30

Build a Daily Mini-Course With o3 and ChatGPT Reminders

The episode describes using o3 to create a compact design course and ChatGPT's reminders feature to deliver one lesson each day. The combination links model-generated curriculum with a practical learning cadence.

o3 can structure a mini-course on a chosen topic
ChatGPT reminders can deliver daily lessons
Scheduled delivery turns a static curriculum into a learning habit

“he used oh three to create a design course. And then he uses remind, the reminders tool in chat, GPT, to give him a lesson…”
03:30

#learning#mini-course#reminders#o3

Takeaway· 1

Takeaway19:00

The Two Jobs o3 Does Best

The host concludes that o3 is especially effective as a prompt writer and as a high-skilled thought partner. Its speed and conversational framing make complex reasoning feel smoother and more human than many competing reasoning models.

o3 writes strong prompts for other models
It works well on uncertain and forward-looking questions
Its responses feel comparatively conversational
Speed improves the experience of iterative strategic work

“I think oh three is quickly going to become your, one of your favorite prompt writers and high skilled peer thought partners.”
19:00

“I find it to be much more, much smoother, much more humanistic than a lot of the other reasoning models today.”
19:00

#o3#prompt writing#thought partner#reasoning models