Mullet loving woodchuck skinning model merge of AI, search, and redneck. I refuse to believe vector retrieval is the final form of how systems get context.
It’s been wild to see how quickly I kick my Claude code habit with a little incentive. Free Ox-alpha was all it took to jump to open code. I wonder how widespread this is. I had that experience with codex a few months ago but somehow ended up back in Claude land
You don't always need a frontier model.
A recent benchmark found that Gemma 4 31B matches Sonnet 5 on answer quality at ~40x lower cost.
With high cost-efficiency and low latency, Gemma unlocks high-volume use cases that are uneconomical with larger models.
I love web search and search engines. I have worked at a few and spent nearly a decade trying to make them more LLM-ready. @chaschwartzz and I decided I wanted to see how Frontier models perform on hard queries, comparing Web to Alphasense. What we found blew my mind.
I had my most apparent perfect example of the benefits of in person work at the office today. I heard someone yapping about a tool they built to make their and some of their teams life better. It was a direct compliment to a tool my team was building. We made PB&Js immediately.
