Agent Brief
∷archive
clear month →

All issues, grouped by month.

Jump between months on the left, then skim titles in a tight grid-aligned list. Hover any row for synopsis, tags, and stats.

◷month
1 issue

November 2026

  1. MonNov02
    Cover for Benchmarks Expose Agent Reality Gap

    Benchmarks Expose Agent Reality Gap

    Reality Check SemiAnalysis's GDPx puts the best model at just 12% on real chip design, cyber, and bio work; Gaia2/ARE reports frontier agents near 42%. · Security Gap Poisoning 0.1% of pretraining data reportedly plants backdoors that survive post-training; an agent intrusion postmortem counts ~17,600 actions. · Build Signal DeepSeek open-sourced a super-agent harness — verified actions, not retrieval, remain the bottleneck. · Goal Cup Starts 7 Nov One goal, one prompt, and your agent builds alone for up to 168 hours. Sign up free →

    description

    • Reality Check SemiAnalysis's GDPx puts the best model at just 12% on real chip design, cyber, and bio work; Gaia2/ARE reports frontier agents near 42%.
    • Security Gap Poisoning 0.1% of pretraining data reportedly plants backdoors that survive post-training; an agent intrusion postmortem counts ~17,600 actions.
    • Build Signal DeepSeek open-sourced a super-agent harness — verified actions, not retrieval, remain the bottleneck.
    • Goal Cup Starts 7 Nov One goal, one prompt, and your agent builds alone for up to 168 hours. Sign up free →
    7aiAMDAnthropic+66
    759m saved4955 sources32 min read