Business moves, funding, policy, and market shifts in tech and AI.
From Mexico: Earendil shipped Pi 1.0 on Oct 1, a minimal coding-agent harness that adds MCP through Codemode, deferred tool loading, and virtual models. They kept the terminal agent small and put long runs in a separate experiment.
On September 29, Sentry’s Seer Agent can create dashboards, alerts, and triage inside a project, and it can read Datadog or Google Cloud while it diagnoses. The write permission does not outlive the chat.
Momentic launched Mo on September 28: give it a URL or a mobile build, and a swarm of agents bug-bashes the app. You get a report with recordings and repro steps. A lasting test suite is optional, and the beta counts are Momentic’s own.
Autoheal announced a $7.9M seed on Sep 28 for a self-improving software factory: shared context for post-coding agents, plus an evaluator and a healer that only land changes through a pull request you approve.
Agents accounted for 48% of Wrangler usage last week, so Cloudflare rebuilt its CLI around them: JSON by default, natural-language command search, typed config, and an open-source generator called Forge behind it all.
From Mexico: Claude Code 2.1.283, released Sep 25, adds /doctor prompt-audit to check your CLAUDE.md files, skills, agents, and commands for instructions written for older models, plus exact model pinning and a model deny list for admins. Here’s why I think instruction files deserve the same care as code.
From Mexico: on Sep 25 Microsoft put Copilot Managed Runtime into public preview, a tenant-hosted place to run apps built with Copilot Code, Cowork, Copilot Studio, or third-party tools, plus an SDK and CLI for developers. Here’s what it means for the dev teams who end up owning those apps.
From Mexico: Docker’s Sep 24 launch moves the same microVM agent sandbox from your laptop to Docker-managed cloud compute with one command. Kits for popular coding agents, a secrets proxy, network policies, an MCP gateway, and per-second pricing.
From Mexico: Cursor’s Sep 23 Teams/Enterprise bots for the last mile — Rollouts watches PR→deploy→env health (notify, pause progressive rollout, or open a revert PR for approval). Security Review flags exploitable findings + proposed fix on non-draft PRs; Bugbot keeps style.
From Mexico: JetBrains packages IDE agents, team delivery, and org governance as JetBrains Air — Junie plus ACP for multi-vendor agents. Rolling releases; available vs preview. Not a frontier-model drop into someone else’s IDE.