Last night, someone asked me what I've been up to since 2010. My reply turned into a short autobiography. I considered deleting it, but people encouraged me to post it instead: gist.github.com/shawwn/3110a… If you're unhappy with your life, it's important to believe you can fix it.

Source: tiktok.com/t/ZT8MA1rrS/

How the kneecap works

It’s also nice because chopping off the last digit means you won’t be wrong by more than a small margin. 10% of 1129 is 112.9, but 112 is close enough in almost all real world situations.

One of the most useful arithmetic hacks I learned came from dota. You have to estimate costs and percentages quickly. The trick is, to find 10% of something, chop off the last digit. So if you want to know 40% of 1120 damage, 10% is 112, times 4 is 448. Easier than the vid imo.

Early in my career I broke Facebook Chat (pre Messenger) for all IE users. Couldn't sleep all night afterwards. However, one of the most interesting issues I caused was when I removed about 300,000 duplicated files in Facebook's monorepo and pointed everything to the remaining copies. I went to sleep and woke up the next day with angry messages on my phone: I made facebook.com 1.2 seconds slower to load! How? Facebook's bundler used machine learning to optimize JS bundles. That bundler was relying on paths for JS files, and it didn't track changes like mine. So when I pushed my change, the bundler was under the impression that all those files were new and basically unbundled them and sent them to users as individual files instead of JS bundles! Given the scale Facebook was operating at, and some issues in the Python based Mercurial client (which have since been replaced), it took about an hour to merge my PR into main. The revert to make facebook.com fast took a three people team 12 hours to complete. Afterwards an intern and I redid the initial work by also patching the bundler's paths correctly. I caused other large issues, but this one is my best bad one. It worked while testing, could not have been noticed in development since the bundling infrastructure was different, and basically only one person at the company would have known this issue could have been caused.

Dozens of people have told me I belong in prison for making books3. It’s not merely lawsuits that researchers are afraid of. Criminal copyright infringement has a penalty measured in years. As for where to get training data: annas-archive.org/llm. But twitter will despise you.

Richard Sutton (Father of Reinforcement Learning) joined Keen Technologies ↓

First diaper change: how it felt when I did it without waking up my wife piped.video/svocYcH5VIU?si=E9x9…

First dr appt

First walk

Thread of firsts

This subthread was wild from start to finish. Nights like these are why I love ML.

Another knowledge branch:

They’re dropping a bunch of knowledge in the replies, too:

This hack is a work of art. I aspire to have this level of casual skill in something physical someday.

If I had to point to one resource that I can credit for learning about analog video specifically and tinkering with old stuff more generally, I would probably point to @craig1black 's YouTube channel Adrian's Digital Basement.

They are tying to scare anyone from using OpenAI derived technologies, basically a modern variant of patent trolls. They *want* people to be overwhelmed by these requests and fear the consequences.

If they wanted to know which books are in books3 or how I made it, they could have just asked me. What could they possibly want here, and why do they think they authority over my private property? Is there an ongoing criminal investigation? This is the part I don’t get.