Pinned post
max
@maxbittker
3.5K Following 12K Followers 10K Posts
Latest posts 5
Matches what I've seen in my multi-agent evaluations, where agents being helpful and accommodating becomes destructive in a group setting
The Collaborative Intelligence Benchmark paper is out!
We tested 18 AI models as teammates on 240 tasks. The best model scores only 82.2 out of 100.
Frontier models almost never cheat. But when a teammate pushes back, they give in!
Read the paper at arxiv.org/abs/2611.02633View quoted post on X
Now that Kimi 3.1 Pro has decent inference providers, it does well on Runescape Bench!
New #1 open weight, and only beaten by frontier models from the last 2 months.
New #1 open weight, and only beaten by frontier models from the last 2 months.
Grapes were trading >90k gp each yesterday on the runescape agent server, because speed runners are hoarding them to make wine.
cheap intelligence driving new forms of demand
cheap intelligence driving new forms of demand
