#1 AI forecaster on Manifold Markets (and #2 across all categories) manifold.markets/Bayesian I want everything to make sense
- Tweets 1,314
- Following 2,727
- Followers 1,408
- Likes 20,236
I think AI is probably conscious despite occasionally touching grass
AI isn’t conscious.
If you think it is conscious or will be conscious, I personally think you need to remove any access you have to model weights asap.
Now step away from the laptop or keyboard, put down your devices and go outside and go touch some grass.
1
11
434
the liquidity phrenologists among you might be able to decipher this. For the rest of us, here's what this means:
WE BELIEVE!
Now at 50%. One of our highest volume markets of all time: a highly contentious coin flip.
Screw p(doom), tell me your p(movie)!
Based on the comments, many users have convictions on the extremes (5%, 95%)
11
827
Bayesian retweeted
opus-5.5 made a ytp with darios CBS interview, actually quite funny in some places
audio on
37
78
1,266
69,942
Bayesian retweeted
Based on my estimates, the 10,000-agent swarm that solved the Navier-Stokes problem was about 8 months of frontier progress ahead of a single GPT-6-Astra instance (with a range of 5.5 to 12.4 months)
It used 48,000 times more output tokens than the largest single model evaluation and roughly 68 times more than an entire long-horizon coding benchmark
35
97
1,518
179,622
Bayesian retweeted
Introducing WeirdML v3, a fully agentic benchmark featuring 11 complex hand-made tasks.
Models must explore and understand unfamiliar data, develop ML and data analysis pipelines and produce results despite limited data, unspecified goals and/or very limited feedback.
1/8
WeirdML v2 is now out! The update includes a bunch of new tasks (now 19 tasks total, up from 6), and results from all the latest models. We now also track api costs and other metadata which give more insight into the different models. The new results are shown in these two figures. The first one shows an overview of the overall results as well as the results on individual tasks, in addition to various metadata.
The second figure shows cost vs performance and shows a clear scaling with better results for higher costs. We also have a very varied pareto frontier with 11 models from 6 different companies having the best accuracy for a given cost for at least some of the cost range. Grok 3, Claude Opus 4 and GPT 4.5 are the ones that underperform for their costs, while Gemini pro and o3 pro have the best results at the highest costs. Qwen3 30B3A, grok 3 mini and deepseek R1 also each represent a good chunk of the pareto frontier.
54
56
655
140,262
Bayesian retweeted
6 months ago: Dario Amodei meets Susie Wiles and Scott Bessent at the White House about Mythos.
Asked about Amodei's visit, Trump says "Who?"
cnbc.com/2026/04/17/anthropi…
1
11
224
8,055