📡 Radar — 1 Sep 2026

The daily catch for 1 September — all days →

📡 1 Sep 2026

🛠️ AI Development

News
Google Antigravity adds a /boost deep-reasoning mode for coding agents
A dedicated reasoning mode aimed at multi-file debugging, refactoring and verification. It's Google's own launch post, but a notable capability shift for anyone running coding agents.
via @antigravity ↗
Article
Vercel's DESIGN.md: getting coding agents to build on-brand pages
A committed DESIGN.md standard plus an eval loop so agents follow a design system, with a claimed 57% drop in layout failures. A reusable technique, though it comes from Vercel's own team.

Article
Between Algorithm and Intuition: keeping designer agency in AI-assisted UX research
New research on preserving human judgment when AI helps make sense of qualitative UX data — a check against over-delegating the interpretive core of design research to models.
Article
NN/g's AI glossary for design and product work
Plain-language definitions of ~80 AI terms, from tokens and embeddings to automation bias and prompt injection. A reference rather than news, but an authoritative one to keep bookmarked.

Post
One person plus agents ships a 114k-line Azure platform in 7 weeks
A first-hand account of replacing a Scrum team with agentic AI on an enterprise build, covering architecture, testing and ops. One case study, not a controlled result — but a concrete data point.
via @patricksavalle ↗
Post
AI-native orgs and the case for machine-readable coordination
Argues that AI changes how work gets routed, and asks what that does to the PM role, Scrum artifacts and cross-team coordination. Analysis rather than evidence, but a sharp framing of the role.
via @bibryam ↗