Yeah that makes sense. I took the optimizations of my hamming distance library to a bit of an insane level. I wrote this about the first round https://emschwartz.me/unnecessary-optimization-in-rust-hammi... The second round brought the best result down to 0.8 nanoseconds per comparison on x86 with SIMD (for a batch of 1000).
Separately, I’m a huge fan of your writing about search! It’s been very helpful while I’ve been building https://scour.ing.
I remember similar observations for an earlier use case in computer vision, loop closure and place recognition for visual SLAM algorithms. In this case the goal was to find a needle (or needles) in a big haystack of visual descriptors (in some sense, proto-embeddings for small image patches or in some cases, whole images). Several approaches used hierarchical data structures for the NN search such as k-mean trees. But linear search - especially with binary descriptors, also became popular as a fast and simple alternative.
Definitely put your finger on what I’ve been feeling.
> What AI did was give every engineer a small team of tireless, fast, occasionally-wrong direct reports. And with the team came the manager’s problem. The discomfort engineers are feeling right now isn’t an AI problem. It’s a delegation problem, and delegation is the oldest unsolved problem in our discipline.
What fascinates me is that the idea is not really a novel finding, plenty of people are making the same observation. What's strange is that the conversation always stops there... nobody seems to ask the harder question: management is a skill that took most of us years to learn badly before we got decent at it. "Welcome to our world" is not giving any help.
Couldn’t agree more. And for what it’s worth, its aptly-named author Merlin Sheldrake is an absolutely charming human too. As quirky iconoclastic woodland enthusiasts go.
I suspect some folks here might also appreciate his early-career (2011) musical collaboration with brother Cosmo Sheldrake and friends, as the Gentle Mystics:
Cosmo is his brother, wow! (have deeply adored his solo work since discovering it last year)
Aaah, it all connects - a web within a web indeed...
I have got some weekend reading and listening to get to now - thank you all kindly.
As a contribution wanted to mention Paul Stamets and his works - it's all somehow about fungi (and bees sometimes!), and all deeply fascinating
As someone who's read a lot of pop sci on this topic as well as published research I went into this book expecting the tropes I've become very familiar with. Merlin Sheldrake actually brought some refreshing and unique takes and I walked away actually learning some new things. It's become my goto introductory book to get people interested in the world of fungi
I’ve been using MixedBread, which is a pretty old model at this point. Recently, I tried comparing it to some newer models and was disappointed that the results weren’t dramatically and uniformly better.
You probably can’t go wrong if you pick a recent one that scores decently well on benchmarks and is at the right price point (or memory requirement) for whatever you’re trying to do.
I have feedback. My main piece of feedback is that it should be more obvious at signup that you can select interests from a list (because I prefer doing that). My other feedback is that the categories based on technology from a country (like “Swedish filesystems”) should probably be removed because there is probably noone who cares what country the filesystem came from.
Thanks for that! I added the first one to the feedback board here https://feedback.scour.ing/142 and I'm tweaking the recommendations to avoid topics like Swedish filesystems.
