The Brief
New briefs daily around 7 AM Eastern
Vercel's $1M sandbox challenge finds a kernel zero-day reaching the host
Vercel's sandbox escape challenge drew 1,285 reports and paid $242,000 in bounties. The top find was a Linux kernel networking flaw that let a microVM guest leak host memory, crash the host or run code, with no customer data accessed.
- 01Read
Claude Platform adds OpenTelemetry trace export in public beta
Anthropic now exports OpenTelemetry traces from the Messages API, Managed Agents sessions and the Claude SDKs, and LangSmith accepts them on day one.
- 02Read
Google's first TPUv7 vLLM results claim 1.67x perf per dollar over GB300
SemiAnalysis says Google submitted TPUv7 FP8 vLLM results to its Agentic Inference Benchmark, showing 1.67x better performance per dollar than GB300 NVL72 at FP4.
- 03Read
Subframe benchmark finds 57% of voters prefer AI-built product screens
Subframe's Real Design Benchmark asks people to choose between a real product screen and one built by Opus 5.5 or GPT-6.1 Sol. After thousands of votes, 57% preferred the AI design.
- 04Read
Epoch AI's SpireClimb shows frontier models winning Slay the Spire
Epoch AI's SpireClimb has models play Slay the Spire through the game's graphical interface. Frontier models now win over 90% of easy runs and 20% at the hardest difficulty.
- 05Read
Siemens EDA lifts chip-design agent from 55% to 95% with NeMo Gym
Siemens EDA used NVIDIA NeMo Gym verifiers to find failure modes in a chip-design agent, then rewrote its skills and MCP tool descriptions. Its benchmark score rose from about 55% to 95% as turns fell from 155 to 53.
- 06Read
SWEeper-Bench makes agents find bugs in 200 web apps; top score is 59%
SWEeper-Bench gives coding agents a codebase and an open-ended request, so they must write interactive tests to find bugs in 200 real web apps. The best of 15 agents passes 59%.
A dated brief from the vibe-coding frontier. Today’s Intel.