llm 0.32

Simon Willison, the developer behind the widely used llm command-line utility, pushed out version 0.32 this week. If you've followed his work before, you know the drill: llm is the Swiss Army knife a lot of engineers reach for when they want to hit OpenAI, Anthropic, local models, or dozens of other providers from a single terminal command, without wiring up a new SDK every time a new model drops.

Willison didn't bury this release under a wall of changelog notes in the post itself. Instead, he used the moment to plug something a bit different: a paid monthly briefing, priced at $10, that boils down the month's most consequential LLM news into one curated email. His pitch is refreshingly blunt — pay him to send you less, not more. In a space where every AI lab ships updates weekly and inboxes fill up with noise, that's a pointed bet that curation, not volume, is what busy developers actually want.

It's worth sitting with why that pitch even makes sense right now. Willison has spent years as one of the more trusted narrators of the LLM tooling world, partly because he actually ships software like llm rather than just commentating on it. A 0.32 release is a small, unglamorous milestone on its own. But stacking it against a subscription offer signals where he thinks his real value lies: not in being first with news, but in being the filter that tells you which of the fifty things that happened this month actually matter.

That combination — open-source tool builder plus paid synthesis product — is becoming a familiar shape in this corner of tech. The tools stay free and iterative, version after version, while the attention and judgment around them get monetized separately. llm 0.32 is a data point in that ongoing story, even if the release notes themselves are modest this time around.