Fetching from the wire…
Top 5 · 2026-05-07 · source-backed
In a Latent Space podcast episode, Shopify CTO Mikhail Parakhin disclosed the most detailed enterprise AI adoption numbers I've seen from a public company.
The headline stats: 90-100% of Shopify employees use AI tools daily. The company provides an unlimited Claude Opus 4.6 token budget. Search throughput jumped from 800 to 4,200 QPS at the same quality level. And the target that made me do a double-take: Shopify and major customer Mercado Libre are targeting 90% autonomous coding by Q3 2026. That's three months from now.
Three internal systems anchor the strategy. Tangle handles content-based caching for data processing, creating cross-team network effects where one team's cached computation benefits another. Tangent optimizes experiments. SimGym simulates customer behavior for testing. These aren't chatbot wrappers. They're infrastructure systems that treat AI as a core compute primitive.
This is the strongest counter-narrative to the vibe coding skepticism from Story #1. Where Willison expresses honest doubt about unreviewed code, Shopify is betting its entire engineering org on AI-first development and building purpose-built infrastructure to manage the risk. The question is whether "90% autonomous" means 90% of code written by agents (plausible) or 90% of code shipped without human review (terrifying). Parakhin didn't clarify.
The unlimited token budget detail matters. Most companies I talk to gate AI tool access through approval processes, cost centers, or per-team budgets. Shopify said "no limits" and is watching what happens. At a $200B+ market cap, they can afford the experiment. For smaller companies, the signal is that token budgets are becoming a hiring and retention lever. If your competitor gives engineers unlimited Opus access and you're rationing Haiku, you're going to lose people.
Combined with OpenAI's Symphony announcement and Anthropic's rate limit increases, a pattern is forming. The tooling for AI-augmented development is maturing fast enough that the constraint is shifting from "can we build with AI?" to "have we reorganized our processes to assume AI is building?" Shopify reorganized. Most companies haven't.
Each link below shares sources, entities, or timing with this story.
Mikhail Parakhin doesn't do half-measures. In a Latent Space deep-dive interview, Shopify's CTO (ex-Microsoft, ex-Bing) revealed that 100% of Shopify's workforce now uses AI daily, and the company actively discourages anyone from using a model less capable than Opus 4.6. Not r...
Shopify built an internal coding agent called River. It generates over half the company's code. And it won't talk to you in private. That last part is the interesting bit. River operates exclusively in public Slack channels, refusing DMs entirely. Every prompt, every response,...
This one's been building for days and it crystallized this week. Per The Register, the incident behind the US export-control block on Anthropic's Fable 5 and Mythos 5 wasn't a jailbreak or a guardrail bypass. It was a plain three-word prompt, "fix this code," run against CVE-l...
GLM-5.1 from Zhipu AI scored 58.4% on SWE-bench Pro. GPT-5.4 scored 57.7%. Claude Opus 4.6 scored 57.3%. That's the first time an open-weight model has ever topped a major coding benchmark against the best proprietary models. The specs matter. GLM-5.1 is a 754B-parameter mixtu...
The coding agent wars just entered a new phase. Cursor isn't just an IDE anymore. It's a model company. Cursor released Composer 2.5 on May 18 with a custom agentic coding model trained using 25x more synthetic tasks than Composer 2 and a novel "targeted textual feedback" appr...
MAI-Code-1-Flash, a 5B-parameter coding model, is in GitHub Copilot and VS Code, and Microsoft says it beats Claude Haiku 4.5 across core coding benchmarks, +16 points on SWE-Bench Pro at 51.2% versus 35.2%, using up to 60% fewer tokens. MAI-Thinking-1, a 35B-active MoE with a...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.