Min Choi
@minchoi
Building with AI. Sharing what's wild, what's practical, and what's next.
1.3K Following 388.1K Followers
This is literally my new workflow now:
Realtime Research → Grok Bot
Planning & Orchestration→ Grok Bot
Day-to-day Coding/Debug → Grok Build + Grok 4.6
Write & Run Tests → Grok Build + Grok 4.6
Complex Coding/Debug → GPT-6 Astra
Frontend → Fable 5.1
Bookmark this.
Show more
Kevin Rose just shared his Grok Bot.
It takes years of your saved IG video, runs Transcribe 2.0 + Vision on them, and builds a local Karpathy-style markdown wiki you can search and ask.
Ok Muse is actually useful.
People are using Muse to get their money back, save on bills, get efunds.
10 wild examples:
Ok GPT-6 Astra is insane.
People can't stop building.
10 wild examples.
Ok GPT-6 Astra is insane.
People can't stop building.
10 wild examples.
There is no "best" AI model.
Martian just showed what happens when you stop pretending there is: 46% fewer errors than the single best LLM, across 16 benchmarks, by routing each task to the right model instead of locking into one.
The part that got me: quoted token pricing is a bad predictor of real cost. Some models charge more and think less. Some charge less and burn tokens.
Backed by an oral at ICLR 2026.
If you're hardcoding one model into everything, this is the dashboard to sit with:
Show more
We got 46% fewer errors than the single best LLM across the 16 most used benchmarks (TerminalBench, LiveCodeBench, etc).
Here's how that's possible and what each model can achieve when used optimally (every benchmarks misses the majority of model capabilities) 👇
Interactive Site:
Academic Paper:
Show more
sir, they just mogged Fable 5.1 with GPT-6 Astra
This is GPT-6 Astra.
Anything you can do on a computer, Astra can do for you. Fast.
Ok Fable 5.1 is insane.
People can't stop building games + entire worlds with it.
10 wild examples:
sir, they just mogged Fable 5.1 with GPT-6 Astra
This is GPT-6 Astra.
Anything you can do on a computer, Astra can do for you. Fast.
There is no "best" AI model.
Martian just showed what happens when you stop pretending there is: 46% fewer errors than the single best LLM, across 16 benchmarks, by routing each task to the right model instead of locking into one.
The part that got me: quoted token pricing is a bad predictor of real cost. Some models charge more and think less. Some charge less and burn tokens.
Backed by an oral at ICLR 2026.
If you're hardcoding one model into everything, this is the dashboard to sit with:
Show more
We got 46% fewer errors than the single best LLM across the 16 most used benchmarks (TerminalBench, LiveCodeBench, etc).
Here's how that's possible and what each model can achieve when used optimally (every benchmarks misses the majority of model capabilities) 👇
Interactive Site:
Academic Paper:
Show more
Time to touch some grass folks...
Ok Fable 5.1 is insane.
People can't stop building games + entire worlds with it.
10 wild examples:
Grok Bot is way more powerful if you set it up right.
The @bot team shared 10 pro tips to make your Grok Bots work more like an actual AI team.
Bookmark this:
SAT prep just got a personal AI tutor.
I built AcuPrep to help students study smarter with AI, not just grind through another question bank.
It learns your weak spots, tells you what to work on next, and trains you for questions you've never seen.
Waitlist open. Invite-only.
Show more
How it started: How it’s going:
This Grok Bot prompt is literally $300/hr chief of staff
Grok Bot is way more powerful if you set it up right.
The @bot team shared 10 pro tips to make your Grok Bots work more like an actual AI team.
Bookmark this:
Grok Bot is way more powerful if you set it up right.
The @bot team shared 10 pro tips to make your Grok Bots work more like an actual AI team.
Bookmark this:
SAT prep just got a personal AI tutor.
I built AcuPrep to help students study smarter with AI, not just grind through another question bank.
It learns your weak spots, tells you what to work on next, and trains you for questions you've never seen.
Waitlist open. Invite-only.
Show more
How it started: How it’s going:
