s3 does vectors now vector databases are officially dead

haha, 691x now due to incidental changes! (this doesn't include the 5-10x low-hanging fruit - that's still in the back pocket) 🚀🚀🚀

Interpret this if you will: last night i dreamed that i went for a doctor's appointment and it turned out the doc was Sam Altman. He asked me a bunch of weird questions and looked at me the whole time through a magnifying glass. I was pretty spooked.

be me: start work early because brain buzzing with ideas; immediately spill coffee on new M4 macbook pro; kernel panic 😱

I like this for so many reasons

Very interesting design of a range filter that learns the dataset’s distribution by sampling keys and storing them in a cache-efficient trie. Uses y-fast tries instead of a traditional trie to achieve an 𝑂(log2 L) number of cache misses for predecessor and successor queries

The deluge of AI slop posted on Medium is getting ridiculous tbh

In the industry for 27 years, and only called myself a software engineer for the first 2 of those because that was the job title and I didn’t know better.

Ugh, I hate how the word "utilize" gets utilized* when "use" would be more correct and make the sentence easier to comprehend. * As in "put to use beyond its original purpose", the actual narrower meaning of utilize ;)

For the over-engineers who need to hear this, Dropbox got to several hundred million users before needing to shard the User db... and then went to two shards.