Follow me for thoughts on the ∩ of people, AI, and business.
CEO of Proof+Geist, co-founder of NormalNext and PauseOnError.
a mournful, defiant celebration of Americans that is in sharp contrast to the story of the USA that is being fed to us today. This is the counter-narrative we need in this moment.
a mournful, defiant celebration of Americans that is in sharp contrast to the story of the USA that is being fed to us today. This is the counter-narrative we need in this moment.
let count = 0;
let name = "Bob";
const unchanging = 42; // const is fine because it never changes
// React equivalent
const [count, setCount] = useState(0);
const [name, setName] = useState("Bob");
const unchanging = 42; // still just a regular const
let count = 0;
let name = "Bob";
const unchanging = 42; // const is fine because it never changes
// React equivalent
const [count, setCount] = useState(0);
const [name, setName] = useState("Bob");
const unchanging = 42; // still just a regular const
A surprising result from Databricks when measuring embeddings and rerankers on internal evals.
1- Reranking few docs improves recall (expected).
2- Reranking many docs degrades quality (!).
3- Reranking too many documents is quite often worse than using embedding model alone (!!).
It's called "School Marketing The Right Way" and it should be published in the New Year.
I'm excited to share the cover of the book with you. 😀
www.thecontextpodcast.com/98
www.thecontextpodcast.com/98