I've published detailed notes and an annotated transcript to accompany the video of the keynote I gave at @WeAreDevs World Congress North America in San Jose on Friday - here's my rundown of everything that's happened with LLMs and agents in 2026 so far …↗

For 3 years, OpenRouter has focused on the ultimate benchmark for LLMs: empirical usage. Today, we added industry benchmarks as well, to help give you the full picture of a model's capabilities. More to come!↗

Because of LLMs, the only type of people I’d hire now are what I’d call “super-unicorns” - Developer - Designer (or great design taste) - Product-focused (vs feature-focused) - Marketer - Agent expert - High emotional intelligence - 5+ years experience creating web app products…↗

others: better run in a sandbox llms are not safe me: literally hooks up claude to whatsapp xD you can limit actions to specific phone numbers tho. Supports both twilio and reverse engineered whatsapp sync (using a second phone is way easier than Meta WA Business verification) …↗

*gets up on soap box* With the announcement of this new "code mode" from Anthropic and Cloudflare, I've gotta rant about LLMs, MCP, and tool-calling for a second Let's all remember where this started LLMs were bad at writing JSON So OpenAI asked us to write good JSON schemas…↗

talked to another great team today who ripped out their ai framework the story is the same every time -- most of the value of an abstraction is abstracting across LLMs the rest eventually weighs you down↗

@thesamparr Alright, I didn't think that would blow up like that haha. Heaps of people have been asking in my DMs about other mega prompts I use, how I manage them, and how to create shortcuts so you can access them across different LLMs. Here is a quick video on it.↗

Human feedback for open source LLMs needs to be crowd-sourced, Wikipedia style. It is the only way for LLMs to become the repository of all human knowledge and cultures. Who wants to build the platform for this?↗

I think this is mostly right. - LLMs created a whole new layer of abstraction and profession. - I've so far called this role "Prompt Engineer" but agree it is misleading. It's not just prompting alone, there's a lot of glue code/infra around it. Maybe "AI Engineer" is ~usable, th…↗