Pocock: agents that build state machines ship sturdier code
Matt Pocock says he can feel the robustness jump every time an agent reaches for a state machine, and XState's David K agrees agents are now extremely good at writing them. Meta also put an 816B open-weight model into the coding race.
Receipts
Pocock says agent-built state machines make frontend code noticeably sturdier
Matt Pocock says every time he sees an agent build a state machine or even a reducer, he can feel a large bump in robustness, and expects future updates to be trivial and testable. He recommends it for frontend work, noting the XState core team member in him is showing.
they make complex logic explicit, impossible states impossible, and agents are extremely good at making them now
How they actually work
dax now keeps one long-running session per project, delegating to subagents
dax says the opencode mobile app pushed him into keeping a bunch of long-running sessions, one per project, and telling each one to do everything via subagents so he can work on several things at once.
Model and agent watch
Meta ships Muse Spark Open, 816B params and built for agentic coding
Meta released Muse Spark Open, calling it the most capable US open-weight model to date: 816B total parameters with 42B active, a 1M token context window, and natively multimodal. Alexandr Wang says it is built for long-horizon tasks, hard software engineering and multi-step reasoning. Nathan Lambert says adoption will go gangbusters if real performance matches the published scores.
First open-weigh model in the US with near frontier agentic capabilities