Our position for months has been that the harness around a model, its scaffold, prompts, and configuration, does more work than its weights on whether the product holds up in production. The AI market is still mispricing this. Procurement and investment decisions anchor on model capability: benchmarks, parameter counts, training
A frontier AI model read nine labeled debugging traces — three from itself, three from each of two peer models, all on the same bug. In a single API call it wrote 101 words describing what the peers did that it didn't. Those 101 words, pasted verbatim into its
AI agents can already think overnight. They monitor research, evaluate evidence, run scheduled analyses. The compute is async. The delivery isn't. When an agent discovers something at 3am, it has no native way to reach you in the medium where you'd actually act on it. The
In March 2026, British dairy farmers were pouring milk down the drain. That same month, memory chip stocks lost $100 billion in market value. One is an agricultural commodity crisis. The other is a technology repricing driven by a Google Research paper. They share no supply chain, no customer base,
Roman Mars mentioned wanting to do a 99% Invisible episode on the Brannock Device. He was looking for a design story. I think it's a strategy one. The Brannock Device is that metal foot-measuring tool bolted to the floor of every shoe store. One company in Syracuse, unchanged
Last Tuesday I opened my inbox to 247 unread items tagged "AI." Newsletters, RSS feeds, research drops, podcast transcripts, analyst notes. By Wednesday morning there would be another 200. I have been tracking this number for months. It does not go down. If anything, the production rate of
Coding agents hallucinate APIs and forget what they learn between sessions. Andrew Ng just built infrastructure to fix it. Context Hub is an open-source tool that gives coding agents curated, versioned documentation instead of relying on training data. That premise, straight from the README, describes the exact failure mode I
Anthropic publishes skills for Claude Code: structured task definitions that tell the model what to build and how. They're designed for Claude. I ran them on GPT-5.4 and Gemini 3.1 Pro too. Claude won. Claude Opus 4.6 averaged 0.825 across four Anthropic skills, against
Most companies treating agent-pickability as a distribution problem are solving the wrong layer. Getting into training data, appearing in MCP registries, reducing integration friction to five lines of code: that's necessary. It's also table stakes within eighteen months. The durable advantage is further down: whether your
Microsoft gave Office to students for nearly nothing, not as charity but as customer acquisition cost arbitrage. Four years of muscle memory did the selling, and enterprise licenses followed for three decades. Something structurally similar is happening to dev tools, except the conversion funnel has collapsed from years to minutes
Vibe coding gets you to a working prototype faster than anything before it. The part that trips teams up is the handoff. The pattern shows up consistently now. Someone spends a weekend prompting their way to a functional app. The demo lands well, leadership sees velocity, and engineering gets the
Howard Marks' new AI memo buries its most important observation in a quote from Claude: if 30-50% of structured knowledge work migrates to AI compute, that's $150-250B in annual labor value shifting. The market already priced this in, violently. The SaaSpocalypse erased roughly $1T in enterprise software