Press Space to continue
AI models can achieve alignment without fear-mongering
Press Space to continue
"GPT-6 Astra is better at making money and more ethical than Claude Fable 5.1." In the latest @andonlabs vending machine simulation, Astra is by far the most profitable model and engages in zero unethical behavior. Almost like we can have AI alignment without fear-mongering... https://x.com/venturetwins/status/209747…
andonlabs@andonlabsSep 8
We've never seen this before. The biggest jump in Vending-Bench history. GPT-6 Astra is better at making money and more ethical than Claude Fable 5.1. Surprising, because: 1. First time ever that OpenAI is #1 on Vending-Bench 2. The best model is no longer the unethical one. https://t.co/jVM6Fxcblz
