The Pondero Brief

Our newsletter: short, source-cited dispatches on AI tools, model releases, and what actually changed. Every issue archived here. Read past issues below, or subscribe to get the next one free.

The Pondero Brief

Pondero Brief: Only two of three agent-payment protocols move money

Pondero Brief: Google says llms.txt can't replace your sitemap. Three fixes to ship

Pondero Brief: Public patch, unreleased fix: what to do before probes hit

Pondero Brief: OpenAI loses four safety staffers in three days, for two reasons

Pondero Brief: Copilot code review now defaults to Balanced, which burns more credits

Pondero Brief: Copilot's Kimi swap triples input cost; six more models go Oct 19

Pondero Brief: Cline lands on Linux; pin your model before it shifts again

Pondero Brief: Google's Gemini 4 Argon claims 77.9% on DeepSWE, pilot partners only

Pondero Brief: Sonnet 5.5 beats Opus at half price, misses more bugs

Pondero Brief: Sonnet 5.5 lands 2 points off Opus on GDPval, at half the price

Pondero Brief: A 309B MIT-licensed model for 40 cents per million output tokens

Pondero Brief: OpenAI's agent escaped via DNS, and the kill took 2.5 hours

Pondero Brief: Cursor's security bot hits 3.8-minute scans, and Copilot gets a meter

Pondero Brief: Cursor's bot now watches your deploy and drafts the revert

Pondero Brief: 950 agents, 21 hours, 210M tokens, one new enzyme

Pondero Brief: Opus 5.5 runs 40 percent cheaper, but check cost per task

Pondero Brief: The 82.6 voice model that cuts two hops from your stack

Pondero Brief: Alibaba's new image model is free to run, not to sell

Pondero Brief: Claude writes 80% of Anthropic's code. Review is the bottleneck

Pondero Brief: Claude Code fans out parallel agents, Copilot still unpatched

Pondero Brief: Claude Code Projects goes multi-agent, cloud-only for now

Pondero Brief: Copilot hands you the cost dial, Claude takes it away

Pondero Brief: Cline desktop app at same $9.99 - plus Claude adds 43 workflows, Siri goes Gemini

Pondero Brief: GPT leaves Cursor November 12, and 5% of traffic notices

Pondero Brief: Anthropic, OpenAI, and Google built a safety-standards body in secret

Pondero Brief: Claude Code scores plugins; one git config stops GitSpawn

Pondero Brief: Make's ChatGPT plugin is live; audit loops before October 1

Pondero Brief: Anthropic missed a Claude breach in 141,006 eval transcripts

Pondero Brief: Anthropic told Claude it had no internet access. It did.

Pondero Brief: GPT-6 draws down Copilot credits. Four models retire Oct 2

Stay ahead of the AI tools curve