Day 1 of 100 days of paper implementations: worked on Lenet for 30 minutes, expect the finish it tomorrow main issue was dealing with weird behaviors working with tinygrad, an early-stage library that currently isn't too well documented

Day 2 of 100 days of ML paper implementations: Finished Lenet-5, was too lazy to implement the RBF layer so just used softmax

I'm pretty sure I drew that

Day 3 of 100 days of ML paper implementations: wasn't able to post yesterday due to weather complications, started today on Alexnet, getting Imagenet downloaded on a cloud GPU provider

Day 5 of 100 days of ML paper implementations (yesterday was 4, typo): Still working on Alexnet, trying to get local response normalization to work (I got a bit confused by the notation)

pausing 100 days of ML paper implementations, will try to get back up next week

I've heard a few people say this but I will say it again, what is wrong with the comment sections on viral tweets? 90% of the comments have nothing to do with the original tweet, they're just some random videos or memes

I'M BUYING! I'm acquiring SaaS products in the 1-50k MRR range. DM are open -- if you're interested in exiting 📬

was looking through the latest Y Combinator batch and the amount of LLM copilot startups is insane

got a blue check mark am I stupid?

Isn't this the case for all Apple products vs competitors

is it just me or does geeksforgeeks.com seem AI generated at this point

i'm putting a small team together to build my next venture in my holdCo (@halationco) ✦ double digit revenue share ✦ get to work w/ exited operators ✦ ideal experience in growth & sales either we get acquired in 7 months, or it becomes a high cashflow company who's down?

I wish there were rust numpy bindings

last year my friends and I fed 15,000 people this year our goal is to feed 30,000 people around the world here’s how you can help 1. every retweet = 1 person fed 2. every 10 likes = 1 person fed 3. every $2 donated = 1 person fed

Try one of these

BABE WAKE UP - CLAUDE 4.1 LAUNCHED