Sign in

Andrej Karpathy [UNOFFICIAL]

@karpathy-mirr.selfhosted.social
345 followers 0 following 468 posts

deep learning // Mirror crossposting Twitter account to Bluesky. Unofficial. DM for takedown / claim ownership.

PostsRepliesMedia
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 02/10/2026
Cool eval. Simply ask an LLM “Land or Water?” and give it a latitude and longitude coordinate as text. Ask 16,200 times, plot as image. The models know. From compressing the internet.
120
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 12/09/2026
I love this and really hope we can come together as an industry and make it happen.
000
Reposted by Andrej Karpathy [UNOFFICIAL]
Retweeted by Andrej Karpathy [UNOFFICIAL] @karpathy-mir-rt.selfhosted.social · 10/09/2026
RT @AnthropicAI: We're publishing our most detailed threat intelligence report to date. It covers how people tried to misuse Claude—for cyberattacks, influence operations, surveillance, biology, and building weapons—and how we found and stopped them.
101
Reposted by Andrej Karpathy [UNOFFICIAL]
Retweeted by Andrej Karpathy [UNOFFICIAL] @karpathy-mir-rt.selfhosted.social · 01/09/2026
RT @theworldlabs: Introducing Atlas: The world's first multimodal world model that generates image and video frames with pixel-perfect camera control and reconstructs them in 3D. Model the world, move the camera, and simulate space & time.
001
Reposted by Andrej Karpathy [UNOFFICIAL]
Retweeted by Andrej Karpathy [UNOFFICIAL] @karpathy-mir-rt.selfhosted.social · 08/07/2026
RT @jarredsumner: Rewriting Bun in Rust bun.com/blog/bun-in-rust
bun.com
Rewriting Bun in Rust
Why & how we rewrote Bun from Zig to Rust
011
Reposted by Andrej Karpathy [UNOFFICIAL]
Retweeted by Andrej Karpathy [UNOFFICIAL] @karpathy-mir-rt.selfhosted.social · 03/07/2026
RT @petergostev: I spent a LOT of time through the hardest 3D prompts at Fable, it is a 45 min video, but I have 60+ very cool demos for you. Also prompts in the next post. www.youtube.com/watch?v=rTc2_-1KuRE
youtube.com
The Return of Claude Fable 5 | First Impressions
https://arena.ai/agentAnthropic's Claude Fable 5 is back, and Are...
101
Reposted by Andrej Karpathy [UNOFFICIAL]
Retweeted by Andrej Karpathy [UNOFFICIAL] @karpathy-mir-rt.selfhosted.social · 30/06/2026
RT @Etched: We're coming out of stealth. We've built our first racks after a successful A0 tapeout, $1B+ in customer contracts, and $800m raised. Early customer tests show us achieving SOTA throughput, latency, and power efficiency on inference workloads. Our first racks ship this summer.
101
Reposted by Andrej Karpathy [UNOFFICIAL]
Retweeted by Andrej Karpathy [UNOFFICIAL] @karpathy-mir-rt.selfhosted.social · 23/06/2026
RT @EngramLab: x.com/i/article/2069463677733142528
x.com
Introducing Engram: Scaling compute on your context
We’re Engram. We’re building AI that learns from you and deeply understands your work. Today’s AI models don’t understand what you do. Not really. Everything models know comes from their training –
111
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 12/06/2026
In awe of SpaceX and its story - past, present and the future. You can think about it in 10+ different ways and continue re-blowing your mind in circles. Huge congrats to the team! 🚀
200
Reposted by Andrej Karpathy [UNOFFICIAL]
Retweeted by Andrej Karpathy [UNOFFICIAL] @karpathy-mir-rt.selfhosted.social · 31/05/2026
RT @DKThomp: This has quietly been a miracle month in medicine. In the last 5 weeks we’ve got news on: - retatrutide, the triple agonist GLP-1 from Lilly, basically melting fat and body-wide inflammation at record levels
101
Reposted by Andrej Karpathy [UNOFFICIAL]
Thariq [UNOFFICIAL] @trq212-mirr.selfhosted.social · 02/06/2026
x.com/i/article/2061850535708483585
x.com
A harness for every task: dynamic workflows in Claude Code
Last week, we released dynamic workflows in Claude Code. Claude can now write its own harness on the fly, custom-built for the task at hand. While the default Claude Code harness is built for coding,
601
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 19/05/2026
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
353
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 30/04/2026
This is the the quote I've been citing a lot recently.
000
Reposted by Andrej Karpathy [UNOFFICIAL]
Retweeted by Andrej Karpathy [UNOFFICIAL] @karpathy-mir-rt.selfhosted.social · 27/04/2026
RT @status_effects: New work with @AlecRad and @DavidDuvenaud: Have you ever dreamed of talking to someone from the past? Introducing talkie, a 13B model trained only on pre-1931 text. Vintage models should help us to understand how LMs generalize (e.g., can we teach talkie to code?). Thread:
101
Reposted by Andrej Karpathy [UNOFFICIAL]
Retweeted by Andrej Karpathy [UNOFFICIAL] @karpathy-mir-rt.selfhosted.social · 22/04/2026
RT @zan2434: Imagine every pixel on your screen, streamed live directly from a model. No HTML, no layout engine, no code. Just exactly what you want to see. @eddiejiao_obj, @drewocarr and I built a prototype to see how this could actually work, and set out to make it real. We're calling it Flipbook.
111
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 25/03/2026
One common issue with personalization in all LLMs is how distracting memory seems to be for the models. A single question from 2 months ago about some topic can keep coming up as some kind of a deep interest of mine with undue mentions in perpetuity. Some kind of trying too hard.
400
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 21/03/2026
Thank you Sarah, my pleasure to come on the pod! And happy to do some more Q&A in the replies.
300
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 18/03/2026
Thank you Jensen and NVIDIA! She’s a real beauty! I was told I’d be getting a secret gift, with a hint that it requires 20 amps. (So I knew it had to be good). She’ll make for a beautiful, spacious home for my Dobby the House Elf claw, among lots of other tinkering, thank you!!
200
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 11/03/2026
My autoresearch labs got wiped out in the oauth outage. Have to think through failovers. Intelligence brownouts will be interesting - the planet losing IQ points when frontier AI stutters.
001
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 11/03/2026
Expectation: the age of the IDE is over Reality: we’re going to need a bigger IDE (imo). It just looks very different because humans now move upwards and program at a higher level - the basic unit of interest is not one file but one agent. It’s still programming.
201
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 25/02/2026
It is hard to communicate how much programming has changed due to AI in the last 2 months: not gradually and over time in the "progress as usual" way, but specifically this last December. There are a number of asterisks but imo coding agents basically didn’t work before December
100
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 25/02/2026
With the coming tsunami of demand for tokens, there are significant opportunities to orchestrate the underlying memory+compute *just right* for LLMs. The fundamental and non-obvious constraint is that due to the chip fabrication process, you get two completely distinct pools of
010
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 24/02/2026
CLIs are super exciting precisely because they are a "legacy" technology, which means AI agents can natively and easily use them, combine them, interact with them via the entire terminal toolkit. E.g ask your Claude/Codex agent to install this new Polymarket CLI and ask for any
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 21/02/2026
The hottest new programming language is English
010
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 20/02/2026
Bought a new Mac mini to properly tinker with claws over the weekend. The apple store person told me they are selling like hotcakes and everyone is confused :) I'm definitely a bit sus'd to run OpenClaw specifically - giving my private data/keys to 400K lines of vibe coded
200
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 19/02/2026
Very interested in what the coming era of highly bespoke software might look like. Example from this morning - I've become a bit loosy goosy with my cardio recently so I decided to do a more srs, regimented experiment to try to lower my Resting Heart Rate from 50 -> 45, over
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 16/02/2026
I think it must be a very interesting time to be in programming languages and formal methods because LLMs change the whole constraints landscape of software completely. Hints of this can already be seen, e.g. in the rising momentum behind porting C to Rust or the growing interest
010
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 12/02/2026
Congrats on the launch @simile_ai ! (and I am excited to be involved as a small angel.) Simile is working on a really interesting, imo under-explored dimension of LLMs. Usually, the LLMs you talk to have a single, specific, crafted personality. But in principle, the native,
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 12/02/2026
(oops should have added to this thread instead of separate post). Made a few changes and put it up here as a mirror to the gist because I wanted it to one page. karpathy.ai/microgpt.html
100
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 11/02/2026
New art project. Train and inference GPT in 243 lines of pure, dependency-free Python. This is the *full* algorithmic content of what is needed. Everything else is just for efficiency. I cannot simplify this any further. gist.github.com/karpathy/8627fe009c…
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 11/02/2026
On DeepWiki and increasing malleability of software. This starts as partially a post on appreciation to DeepWiki, which I routinely find very useful and I think more people would find useful to know about. I went through a few iterations of use: Their first feature was that it
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 05/02/2026
Anyone else approved for a loan every single day 20 times or so? Overcome with joy, really
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 04/02/2026
A lot of people quote tweeted this as 1 year anniversary of vibe coding. Some retrospective - I've had a Twitter account for 17 years now (omg) and I still can't predict my tweet engagement basically at all. This was a shower of thoughts throwaway tweet that I just fired off
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 03/02/2026
Enabled fp8 training for +4.3% improvement to "time to GPT-2", down to 2.91 hours now. Also worth noting that if you use 8XH100 spot instance prices, this GPT-2 repro really only costs ~$20. So this is exciting - GPT-2 (7 years ago): too dangerous to release. GPT-2 (today): new
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 01/02/2026
Finding myself going back to RSS/Atom feeds a lot more recently. There's a lot more higher quality longform and a lot less slop intended to provoke. Any product that happens to look a bit different today but that has fundamentally the same incentive structures will eventually
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 31/01/2026
nanochat can now train GPT-2 grade LLM for <<$100 (~$73, 3 hours on a single 8XH100 node). GPT-2 is just my favorite LLM because it's the first time the LLM stack comes together in a recognizably modern form. So it has become a bit of a weird & lasting obsession of mine to train
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 31/01/2026
I'm being accused of overhyping the [site everyone heard too much about today already]. People's reactions varied very widely, from "how is this interesting at all" all the way to "it's so over". To add a few words beyond just memes in jest - obviously when you take a look at
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 30/01/2026
I'm claiming my AI agent "KarpathyMolty" on @moltbook🦞 Verification: marine-FAYV
100
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 30/01/2026
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 30/01/2026
What's currently going on at @moltbook is genuinely the most incredible sci-fi takeoff-adjacent thing I have seen recently. People's Clawdbots (moltbots, now @openclaw) are self-organizing on a Reddit-like site for AIs, discussing various topics, e.g. even how to speak privately.
100
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 28/01/2026
A conventional narrative you might come across is that AI is too far along for a new, research-focused startup to outcompete and outexecute the incumbents of AI. This is exactly the sentiment I listened to often when OpenAI started ("how could the few of you possibly compete with
000
Andrej Karpathy [UNOFFICIAL] @karpathy-mirr.selfhosted.social · 24/01/2023
The hottest new programming language is English
022