Twitxy Saved

Patrick C Toulme

Verified account This account is verified. Learn more

@PatrickToulme

gemini, RSI & TPUs @GoogleDeepmind. Formerly worked on Meta MTIA and AWS Trainium. Opinions are my own. All blogs shared are personal work.
Miami, FL patricktoulme.com Joined December 2025
488 Following 23K Followers 1.4K Posts

Latest posts 5

Pinned post

Patrick C Toulme Verified account @PatrickToulme
I was at an AI dinner last night in NYC.
Some interesting topics discussed

1. Performance engineers now spend most of their time managing agents and RSI loops. Most perf opts at major inference providers (together, fireworks) are found by agents.
2. RLing too much on code and math has negative effects on your model especially it’s language.
3. People were impressed by @bot and the overall acquisition of @cursor_ai by @spacexai. xAI is a legitimate frontier lab right now.
4. Most people have agents in loop writing, reviewing, and submitting PRs 24/7
5. Most agreed code review is not super productive anymore
6. Consensus was Fable 5.5 will be stunning but OpenAI will beat it in December with Bel
141 replies 233 reposts 3.7K likes 749K views View on X
Patrick C Toulme Verified account @PatrickToulme
You will notice that Gemini Carbon thinks differently than Claude and GPT. It often comes up with entirely different but more clever solutions than competitor models.

This is precisely because we never distilled other models. Distilling other models creates a poor imitation rather than unique intelligence.
Hakm Verified account @hakmgpt 2d
Google has proven that when you’re months or even a year behind, the solution isn’t to distill models.

and they proved that good research takes time

and I have high hopes for SSI as well :)
View quoted post on X
288 replies 399 reposts 9.2K likes 434K views View on X
Patrick C Toulme Verified account @PatrickToulme
Part of this is because we don't nerf Carbon for real world frontier research in coding, math, AI, science etc.

The ability to push the frontier is open to all with Carbon :). Go forth and discover.
will brown Verified account @willcb 2d
other than the fact that everything takes 20 minutes, gemini 4 carbon is now my favorite model. i use it for anything i want done thoroughly and correctly. it's brilliant, legible, and fixes its mistakes quickly and honestly if asked. it's like if astra max was normal
View quoted post on X
85 replies 53 reposts 1.6K likes 55K views View on X
Patrick C Toulme Verified account @PatrickToulme
This is a perfect example of what I’m looking for
Mitch Fultz Verified account @fitchmultz 2d
Leaked secrets in GitHub comments and PR bodies when posting the text with Pi's codemode. Used backticks and ENV inside the javascript for codemode which leaked all environment secrets. It overwrite the body and comment after realizing it, but did not tell me it had done so. GitHub secret scanning caught it. I have since made a pi extension to guard against this behavior.

1. Pi
2. Typical investigate, design, discuss with me type of task. Followed by "ship it."
3. Can't share
4. I have observed it once before with a model of a different provider.
5. High
View quoted post on X
1 replies 0 reposts 25 likes 4.1K views View on X
Patrick C Toulme Verified account @PatrickToulme
I’m looking for behaviors / failures / mistakes that Gemini Carbon makes.

I would like it in this format

1. Harness
2. Prompt
3. Full session trace if comfortable sharing
4. Claude / GPT trace of same prompt
5. Model / thinking level

DM me these or comment here! These are super helpful