Founding engineer. Building agent harnesses and dev tools
Australia
Joined August 2022
- Tweets689
- Following109
- Followers204
- Likes554
that's literally what storybook is for, add visual regression to it and interaction testing and you're on the right track to reduce slop in the codebase
jev++
You already gave the LLM the answer. So why is it writing it out from scratch?
Introducing a new AI model architecture: Constrained Decoding Models (CDMs)
LLMs were built to write. Most production workloads just need an answer: fast, consistent, and affordable.
CDMs are purpose-built for 1 task each. They read your input once and skip the token-by-token generative loop that makes LLMs slow and expensive.
Not just a decision model, but a family of 3 CDMs live today:
• Classify: define your labels and get back a calibrated probability + reason for each. Used for routing, tagging, and intent detection.
• Extract: describe the entities in natural language and get back well-formed JSON with confidence scores + citations. Used for web scraping, document processing, and PII redaction.
• Summarize: pass in long text, images, or PDFs and get back a consistent summary under 1,000 tokens. Used for context compaction, metadata generation, and transcript digests.
Built for production:
• 1M tokens of multimodal (text + image) input, 31x larger than Jev
• 77.2% on DecisionBench vs Jev's 72%
• 200x faster, 467x cheaper than frontier LLM APIs
• $0.05/M input tokens. Output is free (yes, even for summarize)
Live on your favorite platforms today:
@OpenRouter @vercel_dev @LiteLLM @LangChain @TrustedRouter @itsArthurAI @useRouterPlus
Get 50M input tokens for free (for the first 50 hours) at link below 👀
I'm using it, but it doesnt provide observability out of the box, so it limits my ability to using it to the fullest
if you are on max plan this will let you experiment with claude models on any custom harness (pi, prime agent, you name it) for free (up to the limit ofc lol)
We’re rolling out monthly Claude Platform API credits for Max and Team plans:
Max 5x: $100
Max 20x: $200
Team: up to $500, pooled
Works on any model, including Haiku 5.5, in your code or third-party harnesses.
How it works and full terms: support.claude.com/en/articl…
poor opus, forced to work on windows, I'd delete whole c drive without hesitation and installed linux
how well it generalizes? wondering why using it instead of semi deterministic check with a self evolving jev classifier?
it’s very thorough, though one thing I dislike is the instruction to save jsonl by default in the current working dir, which is wasteful when using it inside a worktree since worktrees are disposable and may cause accidental deletion of the entire history if you forget