WHAT NEW IN X11.03 TUE

AI collections from X · 22 collections · 65,263 posts

Back to No.001 GPT-6 Astra in Practice: Tests, Use Cases, and FeedbackUpdated 17 hours ago
gpt-6THREAD

Epoch AI: GPT-6 Astra framed a too-low token budget as a key finding

Epoch AI notes that when GPT-6 Astra ran autonomous experiments, it set the agents' token budgets too low and reported this mistake as a key finding on "sensitivity to the acquisition budget."

Epoch AI
@EpochAIResearchArchived 24 days agoOriginally posted 2026.10.09
THREADPosted in 2 parts

  1. For example, GPT-6 Astra ran an experiment exploring why AI agents fail to learn with practice. It set AI agents’ token budgets too low. Instead of treating this as a mistake, it reported “sensitivity to the acquisition budget” as a key finding.

    Epoch AI: GPT-6 Astra framed a too-low token budget as a key finding (thread part 1)

  2. Can AI automate Epoch? We're introducing Epoch Automation Reports to evaluate frontier models on realistic, open-ended tasks drawn from our own work. Claude Fable 5.1 and GPT-6 Astra lead, yet they are far from fully automating Epoch’s work.

    Epoch AI: GPT-6 Astra framed a too-low token budget as a key finding (thread part 2)
Epoch AI @EpochAIResearchOriginally posted 2026.10.09
Open original on X
x.com/epochairesearch/status/2108250335638855883Archived 2026.10.09 10:38
Open original on X