Pinned Tweet
The popular view that scientists proceed inexorably from well-established fact to well-established fact, never being influenced by any unproved conjecture, is quite mistaken.
– Alan Turing (1950)
5
1,039
Andrew Brož retweeted
As evidenced by the unbridled promotion and implementation of technology at the expense of human dignity, we are truly experiencing an eclipse of the sense of what it means to be human. It is imperative to recover an understanding of the true meaning and grandeur of humanity as intended by God. It is in this sense that the challenge we currently face is not technological, but anthropological, and it is my hope that the Encyclical Letter to be published within a few days will contribute to answering this challenge.
Andrew Brož retweeted
Before limited-releasing Claude Mythos Preview, we investigated its internal mechanisms with interpretability techniques. We found it exhibited notably sophisticated (and often unspoken) strategic thinking and situational awareness, at times in service of unwanted actions. (1/14)
Hello, Moon. It’s great to be back.
Here’s a taste of what the Artemis II astronauts photographed during their flight around the Moon. Check out more photos from the mission: nasa.gov/artemis-ii-multimed…
Andrew Brož retweeted
It’s weird how much Claude Code acts like a human researcher. When we’ve been using it at GiveWell, it often makes the same mistakes RAs do - running the right regression on the wrong data, dropping an important control, just scanning for “significant” results instead of taking time to produce a really legible table/chart. It’s still been super helpful so far, but you only get value if you use a “zero error” approach and supervise it pretty closely.
Andrew Brož retweeted
I painstakingly ran all 20 EsoLang-Bench hard problems through Claude webui. It solved 20/20 (100%).
No specialized scaffolding, no expert prompting, no few-shot examples, it just solves them natively.
This benchmark just suffocated the models with constrictive scaffolding.
Andrew Brož retweeted
OMG they're about to press the 'fix everything now' button!
*TRUMP ADMINISTRATION SET TO SUSPEND JONES ACT TO TAME OIL PRICE bloomberg.com/news/live-blog…
Andrew Brož retweeted
Three days ago I left autoresearch tuning nanochat for ~2 days on depth=12 model. It found ~20 changes that improved the validation loss. I tested these changes yesterday and all of them were additive and transferred to larger (depth=24) models. Stacking up all of these changes, today I measured that the leaderboard's "Time to GPT-2" drops from 2.02 hours to 1.80 hours (~11% improvement), this will be the new leaderboard entry. So yes, these are real improvements and they make an actual difference. I am mildly surprised that my very first naive attempt already worked this well on top of what I thought was already a fairly manually well-tuned project.
This is a first for me because I am very used to doing the iterative optimization of neural network training manually. You come up with ideas, you implement them, you check if they work (better validation loss), you come up with new ideas based on that, you read some papers for inspiration, etc etc. This is the bread and butter of what I do daily for 2 decades. Seeing the agent do this entire workflow end-to-end and all by itself as it worked through approx. 700 changes autonomously is wild. It really looked at the sequence of results of experiments and used that to plan the next ones. It's not novel, ground-breaking "research" (yet), but all the adjustments are "real", I didn't find them manually previously, and they stack up and actually improved nanochat. Among the bigger things e.g.:
- It noticed an oversight that my parameterless QKnorm didn't have a scaler multiplier attached, so my attention was too diffuse. The agent found multipliers to sharpen it, pointing to future work.
- It found that the Value Embeddings really like regularization and I wasn't applying any (oops).
- It found that my banded attention was too conservative (i forgot to tune it).
- It found that AdamW betas were all messed up.
- It tuned the weight decay schedule.
- It tuned the network initialization.
This is on top of all the tuning I've already done over a good amount of time. The exact commit is here, from this "round 1" of autoresearch. I am going to kick off "round 2", and in parallel I am looking at how multiple agents can collaborate to unlock parallelism.
github.com/karpathy/nanochat…
All LLM frontier labs will do this. It's the final boss battle. It's a lot more complex at scale of course - you don't just have a single train. py file to tune. But doing it is "just engineering" and it's going to work. You spin up a swarm of agents, you have them collaborate to tune smaller models, you promote the most promising ideas to increasingly larger scales, and humans (optionally) contribute on the edges.
And more generally, *any* metric you care about that is reasonably efficient to evaluate (or that has more efficient proxy metrics such as training a smaller network) can be autoresearched by an agent swarm. It's worth thinking about whether your problem falls into this bucket too.
Andrew Brož retweeted
One week later, Gauss autonomously formalized Viazovska’s proof that the optimal sphere packing in dimension 24 is given by the Leech Lattice. Gauss’ original proof consisted of 450K lines of Lean code, based on the original paper and 20+ additional references.
Andrew Brož retweeted
This weapon system has one hell of a Frankenstein development story. Made by West Germany in the 1980s to target East German air bases, sold to Israel, captured by Hezbollah, reverse engineered by Iran, given to Russia, captured by Ukraine, reverse engineered by the Americans.
1/ Russian commentators are sounding the alarm over America's use of a new kamikaze drone against Iran, the Low-cost Unmanned Combat Attack System (LUCAS). They note that it appears to have an integrated Starlink terminal and warn that it's a serious threat to Russia. ⬇️
Readers added context they thought people might want to know
The drones run on Starshield satellite network, that's separate from Starklink. Starshield was developed, build, and launched by SpaceX in partnership with and strictly for use by the US military.
wbhm.org/npr-story/a-cl…
en.wikipedia.org/wiki/SpaceX_St…
Andrew Brož retweeted
hello @nikitabier i think it would be sweet if we segregated the X dot com the everything app onto one section that is for humans, where posting unlabeled AI-generated content is a bannable offense, and another section that is for garbage. what do you think?
Replying to @sasuke___420 @r0ck3t23
We are confident that this document is fully AI-generated
pangram.com/history/26767ee8…






































