hn_20591249

Seems like sama may have put a big hole in that argument when he tweeted "her", now it is very easy to say that they knowingly cloned ScarJo's likeness. When will tech leaders learn to stop tweeting.

A startup founder lying to get juicy government contracts? Say it ain't so!

Yeah I would very much likely to know what caching library has a failure mode of returning content for the wrong keys, that seems pretty bad if not a highly suspect explanation

Driving 10 hours both ways to an AWS event to look at some slide decks you can already find online and eat some mid-tier catering doesn't seem like a great use of a day to me.

I asked for an image of The Muppets as the counter-terrorist team from Counter Strike and that got blocked, so "terrorism" is definitely getting picked up understandably.

I find it very surprising that this is not the top post at the moment. My partner went to work at her bar at 5PM ET, and their Square services which processes reservations, transactions, payroll and even staff scheduling is at a stand-still. It seems like this is the case for a huge amount of bars, restaurants, and whatever other line-of-business Square seems to have infiltrated itself into to cause one massive single point of failure.

I feel for the engineers at this point in time, but their status page being updated at a 45 minute cadence with very little in terms of resolution is.. not so good.

There has been some shuffling of seats but from what I am hearing FAIR is the best setup as far as staffing and funding that they have been in quite some time. Mark is pivoting hard to stay competitive in AI and is providing the resourcing to do so, the results speak for themselves.

> you should have a cloud-native setup ready to go

why is storing the file as a FAISS/LanceDB on-disk vector store not "cloud native"? I am running this setup in production across dozens of nodes, we migrated all of our infrastructure off Pinecone towards this solution and have seen 10x drop in latency, and the cost improvements have been dramatic (from paid, to totally free).

I have a bit of an axe to grind in the vector DB space, it feels like the industry has gaslit developers over the last year or so into thinking SAAS is necessary for vector retrieval, when low latency on-disk KNN across vectors is a solved problem.

Should a one person developer shop be bringing their own stack to a client? Seems like a problem once you inevitably roll off the project.

Maybe I'm grumpy today but I am shocked at how many responses you are getting where people think this is a novel idea. Has the engineering mindset really shifted into a default of "buy" even when build could take less than a week?

15 minutes of install, install, install beats getting into the vicious SAAS vendor cycle of pay, pay, pay with heavy lock-in.

I've hosted embedding models on AWS Lambda (fair that this is a vendor, but 1 vs. 3), if you try an LLM with 1B+ parameters you will struggle, but if the difference between a light-weight BERT-like transformer and an LLM is only a few % of loss, why bother getting your credit card out?

Edit: another thought, skip lambda entirely and run the embedding job on the server as a background process, and use an on-disk vector store (lancedb)

Am I being unreasonable to find it bizarre that the tutorial begins with subscribing to 3 different SAAS vendors?

Especially seeing these days you can run a vector store on-disk if you have less than 10 million records, pull any free embedding model straight from HuggingFace and run on consumer hardware (your laptop).