Doug Turnbull
@softwaredoug
Search at Daydream; ex @Reddit, @Shopify & @o19s; Books: Relevant Search & AI Powered Search Find me at 🦋
Ok y'all, this is wild – my next tech talk will be on WhatNot and I'll be auctioning off signed copies of my book, Prompt Engineering for LLMs. ¯\_(ツ)_/¯ True story is that I've got a buddy that works there and this is how I'm delivering a tech talk to their engineering team.…
It would seem a fairly sensible approach in the US - “I don’t want my tax dollars going to kill kids in Gaza”. But what the fuck do I know.
"LLM as a judge" for search is really just learning to rank in disguise. (with the features / ranker being LLM)
by popular demand, now introducing one-dimensional vectors 🎈
A lot of people think mastering your craft is about learning a new topic But it's actually about relearning something you "know" from first principles (and repeating this every now and again, going deeper each time)
Somehow paid @GoogleColab is often worse than the free tier Free tier, I always have had 3 runtimes available at once. Somehow with Pro ($9.99/month) I can have just one runtime at once? Even when running CPU-only. WTF? Disappointing product
You missed this cohort, the next one is: maven.com/softwaredoug/c…
Now available for ACM Members: "AI-Powered Search," by @softwaredoug @treygrainger @binarymax. Learn new techniques such as Semantic search using dense vector embeddings from foundation models, Retrieval augmented generation (RAG), question answering &... share.percipio.com/cd/b5Xynhdy7
I'd be willing to network while showing myself off wearing clothes for coffee... I'd be a latte interaction model
I used to have an ML-Powered Search course that basically was about Learning to Rank and Rerankers. I've been told people enjoyed it If that interests you, please add yourself to the waitlist, and I might do it again... maven.com/softwaredoug/m…