Teaching LLMs to think ahead; Learning about ML since 2017
Joined September 2017
- Tweets 185
- Following 289
- Followers 200
- Likes 11,765
7 Photos and videos
Apr 11
I’ve being poking around with agent harness and fully autonomous companies and I think I’ve seen end game
3
69
Mar 5
CLI tools is quickly becoming the standard for agents. If your company don’t have a cli tool, you will be left behind
3
63
Feb 28
This is a open lie. OpenAI now helps the mass surveillance of civilians with AI, it helps to erode the core values of privacy.
Feb 28
Tonight, we reached an agreement with the Department of War to deploy our models in their classified network.
In all of our interactions, the DoW displayed a deep respect for safety and a desire to partner to achieve the best possible outcome.
AI safety and wide distribution of benefits are the core of our mission. Two of our most important safety principles are prohibitions on domestic mass surveillance and human responsibility for the use of force, including for autonomous weapon systems. The DoW agrees with these principles, reflects them in law and policy, and we put them into our agreement.
We also will build technical safeguards to ensure our models behave as they should, which the DoW also wanted. We will deploy FDEs to help with our models and to ensure their safety, we will deploy on cloud networks only.
We are asking the DoW to offer these same terms to all AI companies, which in our opinion we think everyone should be willing to accept. We have expressed our strong desire to see things de-escalate away from legal and governmental actions and towards reasonable agreements.
We remain committed to serve all of humanity as best we can. The world is a complicated, messy, and sometimes dangerous place.
2
53
Feb 21
SpaceX already has Pentagon contracts.
Anthropic just refused mass surveillance and autonomous weapons.
The moment Anthropic gets labeled a “supply chain risk,” xAI steps right in.
1
27
Feb 12
Its an unpopular opinion but @openclaw is amazing and unreliable. It's amazing when it works, but it often don't
2
100
Feb 11
As of Feb 11 I no longer take part in the efforts of pushing @xai forward. I've learned a lot throughout this journey and I'm glad for everyone I've met along the way. I ask you please to respect my privacy.
1
1
42
Jan 23
Day 23 of 2026:
I haven’t written a single line of code.
Edited 3 lines.
It’s over.
Time to reinvent myself.
1
20
Jan 15
Anyone wants to build J.A.R.V.I.S? @Marvel
Jan 15
Introducing json-render
AI-generated UI. Deterministic output.
1. Define your component catalog
2. AI steams JSON
3. Render interactive UI
Let users prompt dashboards, widgets and apps - safely constrained to components and actions you define
27
Fsbonetto retweeted
10 Dec 2025
Quick new post: Auto-grading decade-old Hacker News discussions with hindsight
I took all the 930 frontpage Hacker News article discussion of December 2015 and asked the GPT 5.1 Thinking API to do an in-hindsight analysis to identify the most/least prescient comments. This took ~3 hours to vibe code and ~1 hour and $60 to run. The idea was sparked by the HN article yesterday where Gemini 3 was asked to hallucinate the HN front page one decade forward.
More generally:
1. in-hindsight analysis has always fascinated me as a way to train your forward prediction model so reading the results is really interesting and
2. it's worth contemplating what it looks like when LLM megaminds of the future can do this kind of work a lot cheaper, faster and better. Every single bit of information you contribute to the internet can (and probably will be) scrutinized in great detail if it is "free". Hence also my earlier tweet from a while back - "be good, future LLMs are watching".
Congrats to the top 10 accounts pcwalton, tptacek, paulmd, cstross, greglindahl, moxie, hannob, 0xcde4c3db, Manishearth, and johncolanduoni - GPT 5.1 Thinking found your comments to be the most insightful and prescient of all comments of HN in December of 2015.
Links:
- A lot more detail in my blog post karpathy.bearblog.dev/auto-g…
- GitHub repo of the project if you'd like to play github.com/karpathy/hn-time-…
- The actual results pages for your reading pleasure karpathy.ai/hncapsule/
237
568
5,384
609,147
9 Dec 2025
Since this post I've been querying @antigravity with "How would a google senior engineer implement XYZ" and its output have been much better
7 Dec 2025
Don't think of LLMs as entities but as simulators. For example, when exploring a topic, don't ask:
"What do you think about xyz"?
There is no "you". Next time try:
"What would be a good group of people to explore xyz? What would they say?"
The LLM can channel/simulate many perspectives but it hasn't "thought about" xyz for a while and over time and formed its own opinions in the way we're used to. If you force it via the use of "you", it will give you something by adopting a personality embedding vector implied by the statistics of its finetuning data and then simulate that. It's fine to do, but there is a lot less mystique to it than I find people naively attribute to "asking an AI".
52
16 Nov 2025
Gotta a glimpse of what Gemini 3.0 pro can do. It can boost your post engagement over 95% on SWE verified
237
28 Aug 2025
Actively doing nothing is better than doom scroll TikTok. Boredom is incredibly important to the brain.
1
60
8 Aug 2025
I’ve tried to use GPT-5 to refactor my app to use GPT-5 - It couldn’t. I had to use Claude Opus 4.1 to do it. The joke wrote itself.
Open AI is officially behind.
157
7 Aug 2025
We got gpt-5 before gta-6
113
Fsbonetto retweeted
1 Jul 2025
Jim Simons (the most successful hedge fund investor ever) on the importance of collaboration and aesthetics in building something great
16
264
2,240
64,797