AIUpdateWatch Daily Briefing
Today's most important AI developments across the full landscape, selected for practical importance and checked against source evidence. The complete report is preserved at its permanent dated URL.
Latest verified briefing
Anthropic rolls out Claude Fable 5.1 with slashed prompt-cache rates alongside restricted Mythos 5.1 cyber models; frontier labs abandon saturated public code benchmarks for SWE-bench Pro; and hyperscalers retool infrastructure around rack-level agentic throughput.
Decision-ready intelligence
Facts, interpretation and recommended actions are separated. Quiet days are not padded to a fixed number of items.
Anthropic rolls out Claude Fable 5.1 with 1M context and slashes prompt cache-read pricing by 75%.
SWE-bench Verified hits a 96% ceiling as industry shifts to contamination-resistant SWE-bench Pro.
Hyperscalers shift capex contracts to rack-scale liquid-cooled Rubin NVL72 and GB300 systems.
Project Glasswing and NIST AI 600-2 establish verifiable containment for autonomous agents.
Featured explainer
A plain-English explanation of how prompt caching works at the GPU memory layer, why stateless multi-turn agent loops cost 10x more, and how 75% cache discounts make autonomous software engineering economically viable.
Morning-edition movement
Anthropic expanded Claude Fable 5.1 general availability with 1M context, 128k output, and slashed prompt cache-reads by 75% to $0.25/M tokens.
Original source (opens in a new tab)SWE-bench Verified reached saturation as top frontier models clustered between 94% and 96%, accelerating migration to SWE-bench Pro.
Original source (opens in a new tab)Hyperscalers finalized capex commitments for NVIDIA Rubin NVL72 and Blackwell Ultra GB300 rack-scale systems.
Original source (opens in a new tab)Choose your next step
Verification