Viewing profile — lukemerrick

I am not super familiar with C and CUDA, so I read solely for the README and enjoyed it supremely. The blend of cheerful walking through instructive examples and your philosophical…

Meta is a very large organization, and I'm willing to believe that a good chunk of Meta FAIR (the lab releasing all of this stuff) truly do care about innovations for advancing AI …

Looks like there is both an ARM and x86 version according to the docs. Probably need two different binaries, but you still get cross-OS for each architecture.

Just spent two hours on this. Would happily spend the time again (and probably will soon!). Awesome work.

Just skimmed so far and didn't see any reference to the Simplified Transformer block of https://arxiv.org/abs/2311.01906 (and it seems they also left out grouped query attention, t…

Related to the idea of "no one trains on data they own, they shouldn't own the resulting model": since big public datasets like The Pile have CC-SA items in them, is anyone conside…

As you guessed, the history tracking is one of the killer features. Imagine it being super easy to edit the history of a REPL session (delete, reorder, merge, and edit contents of …

Thank you for the tip! Getting to implement a different language from Lox feels like a nice way to cut down on the tedium.

Wow, the whole concept of a peephole optimizer is a bit mind blowing to me. I'm appreciating all the reasons to power through to writing a bytecode VM as the next step. I'm not sur…

Thank you for the suggestions! I was actually just searching about MLIR today after reading some Julia language community discussions on the new Mojo language that uses MLIR.

I actually played with Unityper.jl and SumTypes.jl, but my conclusion was that if I was going to depart from dispatch on Julia types in my code, I might as well just stick to an un…

In general, if you keep the source code positions of every nontrivial token, and you keep the raw source code, then yeah you can print out those pretty specific point error message…

Disclaimer: I wrote this blog post. If this were an "Ask HN" post, though, the question would be "What next after reading Crafting Interpreters?" I have only done the tree-walk int…

I'm late to the party, but I want to say thank you for sharing this. It's inspiring to look at how much you've built and (hopefully) enjoyed the process of building! I'm loving eve…

Disclaimer: I'm the author. I was torn whether to share my own post, but I figured the HN crowd might include a few others who will also really geek out about this topic and apprec…

Built using Pyodide [1] (web-assembly-compiled Python), stalkmyself.com lets you convert your GMail inbox into a CSV file summary without relying on any backend service -- everythi…

HN is special, and you all here in this comment tree are the best of the best -- love the fact that everyone is having fun with this in such a civil way and historically knowledgea…

Ah, this is a nuanced point I totally left off the README. Each raw file is ~5GB, but the raw files are a dump of network traffic from the firehose feed that tracks not just trades…

Backstory: I wanted to play with intraday stock data but couldn't find a free dataset anywhere. IEXCloud [1] offers API access to 1-minute granularity intraday historical price dat…

This is an incredibly salient point. Others have also pointed out that there are numerous applications in which black box models appear to offer significantly greater accuracy than…