Sam Dare @DistStateAndMe

Founder / Member of Non Technical Staff @covenant_ai ( @tplr_ai : @basilic_ai : @grail_ai ) covenant.ai Joined April 2014

Tweets

10K
Followers

5K
Following

3K
Likes

35K

Sam Dare @DistStateAndMe

2 days ago

@LLMPsychologist @StijnSmits seems adjacent to what you are working on with anabasis

1 0 1 114 0

View Details

Sam Dare @DistStateAndMe

2 days ago

@onusoz "I'll write none".

0 0 0 41 0

View Details

rohan anil @_arohan_

4 days ago

Good morning! ☀️

9 6 291 13K 5

View Details

Sam Dare @DistStateAndMe

4 days ago

@llllvvuu @nrehiew_ $740K a day is ready to stop

0 0 1 231 0

View Details

This is an elitist statement. That there’s no reason to be this fast, and therefore we have to slow down! It’s truly remarkable the level of intelligence I have available on my phone with a $200/month pro plan. However, it’s way too expensive for this to be widely available to billions of people around the world. Tao is worried about the nonlinear nature of progress, except that technological breakthroughs, almost by definition are nonlinear. Otherwise, we would have run out of resources in the world long time ago. People forget how recently we figured out how to produce food at today’s scale. In the 1800s, an important source of fertilizer was an island off the coast of South America. Birds had spent centuries dropping their shit there, building up deposits that people mined for fertilizer. Then came a remarkable invention by the German scientist Haber, which Bosch helped scale industrially. By combining nitrogen from the air with hydrogen, using energy, we could manufacture ammonia for fertilizer. Fertilizer from thin air. There’s another lesson in there about the nature of technology. The same technology that helped create an abundance of food also enabled the production of explosives at scale.

Haider. @haider1

6 days ago

Mathematician Terence Tao: "we have to slow down AI. the pace is insane, and there's no reason to be this fast — no reason at all" It's amazing how willing we are to change everything without any idea what happens afterward These are extremely nonlinear dynamics

883 1K 7K 7.5M 3K

13 11 167 17K 45

View Details

Robert Yang @GuangyuRobert

5 days ago

i predict math ai progress will organically slow down millennium prize problems are non-renewable resources for pr, soon to be depleted then the big labs won't be incentivized to spend major effort on math ai similarly, deepmind moved on from go ai (alpha go/alpha zero) not long after beating lee sedol (world champ). they could have kept training stronger and stronger go ai, but who would appreciate it? and i think being even better at abstract math won't transfer to other ai priorities (rsi, knowledge work, robotics, etc.) regardless, the work of pro mathematicians will shift from solving problems to understanding ai solutions

Haider. @haider1

6 days ago

883 1K 7K 7.5M 3K

46 6 155 49K 39

View Details

Sam Dare @DistStateAndMe

5 days ago

@snowclipsed x.com/mariyaivasilev…

Mariya I. Vasileva @mariyaivasileva

5 days ago

I just submitted to ICLR. As of last night’s abstract deadline, there were over 60,000 submissions registered. For comparison, last year ICLR started out with 19,525 submissions as per the stats posted on their blog. 779 (3.99%) were desk-rejected, 5,042 (25.82%) were withdrawn,

18 22 314 28K 113

0 0 1 169 0

View Details

Weco AI @WecoAI

7 days ago

Can autoresearch agents find better training data, not just better training code? Here’s our new EMNLP paper: AutoData: Agentic Search for Pre-training Data Selection [1/6]

7 13 117 18K 99

View Details

Goodfire @GoodfireAI

a week ago

Models know when they’re reward hacking. But they still do it a ton - in 50-96% of rollouts we studied! We built activation monitors that detect the behavior behind the Hugging Face hack in real time. This can help us stop hacks now - and train future models that don’t cheat. 🧵

29 98 836 145K 429

View Details

Sam Dare @DistStateAndMe

a week ago

@Charles_Y_Wu @PhAILabs @hejia0530 @LingYang_PU @zhenfei_yin_yzf Congratulations !!

0 0 0 23 0

View Details

Sam Dare @DistStateAndMe

a week ago

@drakeevans oh its not runnable yet. Will ping when it is .

0 0 0 26 0

View Details

Sam Dare @DistStateAndMe

a week ago

Pre-training over the internet with unreliable peers requires reimagining fault tolerance from first principles. This opens the door to greater economic gains by way of been able to admit spot instances in our runs. The internet is the data center

templar @tplr_ai

a week ago

We’ve been researching fault tolerance in Crucible, Templar’s pre-training platform. The goal: keep training through node failures and make better use of unreliable workers and spot instances, without idling an entire model replica when one stage goes down. 1/n

3 10 12 4K 4

4 0 17 2K 3

View Details

Sam Dare @DistStateAndMe

a week ago

@maximelabonne I can't seem to find the link to the article on HF

0 0 0 76 0

View Details

Sam Dare @DistStateAndMe

a week ago

@maximelabonne @huggingface @elonmusk is chess solved ?

0 0 2 295 0

View Details

Sam Dare @DistStateAndMe

a week ago

@FaroukAdeleke3 @lateinteraction GEPA mentioned

0 0 1 68 0

View Details

Sasha Rush @srush_nlp

a week ago

Lean Verified Transformers (srush.github.io/lean-transform…) In which we prove a bunch of Transformer invariants from scratch in Lean, and speculate about how hard it would be to do that for the rest of the world's code.