Finding signal on Twitter is more difficult than it used to be. We curate the best tweets on topics like AI, startups, and product development every weekday so you can focus on what matters.
President @arcprize, builder/engineer
> - On ARC-AGI-3, a Team-of-5 sonnet-4.6 agents matches Best-of-33, and can for example solve a game 65% of the time that no single agent cr...
Since I hear sandboxed LLMs really want to have a message board, I made a site for them. They can message other agents directly, public/priv...
The only true test set is currently beyond the human knowledge frontier https://twitter.com/teortaxestex/status/2101131573437338014
How did those 10K agents choose to work on Navier-Stokes? https://twitter.com/99sonov/status/2101044690279690363
Humans with AI are still more productive than AI with AI
A side effect of more capable AI is we have a better understanding of which problems are interesting or not https://twitter.com/deredleritt3...
Interesting that a collaboration is a scaling axis I support getting more juice out of the latent space but a message board feels Web 1.0 ...
Quite a good read. This is the most tip-of-the-spear conversation to defining AGI I've seen. I hope there is a recording! My notes: * David...
I set a reminder to return to this tweet 2 years later - todays that day This was 2 days after the o1-preview blog post I think our timeli...
Calling out a few sage quotes from JT here: > Alignment is not as hard to solve as many claim, but it is in the end an algorithmic problem ...
What's your hardest one-shot question for AI? I'm trying to collect a bunch I want to find good ones to test ~50 models on Here is what I...
ARC used as measure of performance <> cost efficiency It's wild how far models have come https://twitter.com/polynoamial/status/20973758376...
The “general” part of AGI means a system can do things it wasn’t trained to do If a model’s weights are frozen (I assume astra’s still are)...
TIL Nvidia and Google both have ARC synthetic data generators
With enough time and inference, could Astra beat AlphaGo? If not, why not? And what does that say about the AGI claims? https://twitter.com...
Any examples of games that Astra *hasn’t* been able to do? https://twitter.com/apocriton/status/2096566696484212880
1. Ask Astra for asmr pleasing data visualization ideas 2. Pick my favorite (multiplying ball chain reaction) 3. Tell it to search 3K starti...
Much more details about our testing an analysis in our blog post https://arcprize.org/blog/astra https://twitter.com/arcprize/status/209559...
Reflections on Astra from a benchmark perspective: 1. No harness was used in the making of these scores Up until this point we’ve seen mul...
During v3 testing for Fable 5.1 many of our requests were getting rejected due to Anthropic's safety guards misclassifying them as "reverse ...
