Join 📚 Josh Beckman's Highlights
A batch of the best highlights from what Josh's read, .
Yes, it would be great if everyone did fully reproducible science and made all their data available and *actually responded* when you asked them questions and got a pony. But in the current, *actual* world, *most* papers are missing important details. The problem of having to scan your eyeballs past a few extra digits is a silly non-issue compared to the problem of meaningless results everywhere. So please stop spending your energy actively trying to convince people to delete one of the very few error correction methods that we *actually have* and that *actually sort of works*.
Please Show Lots of Digits
DYNOMIGHT
The value of a product is the number of problems it can solve divided by the amount of complexity the user needs to keep in their head to use it. Consider an iPhone vs a standard TV remove: an iPhone touchscreen can be used for countless different functions, but there's very little to remember about how it works (tap, drag, swipe, pinch). With a TV remote you have to remember what every button does; the more things you can use the remote for, the more buttons it has. We want to create iPhones, not TV remotes.
My Heroku Values · GitHub
gist.github.com
But roughly speaking (and today), you're not asking some magical AI. You're asking a human data labeler. Whose average essence was lossily distilled into statistical token tumblers that are LLMs. This can still be super useful ofc ourse. Post triggered by someone suggesting we ask an AI how to run the government etc. TLDR you're not asking an AI, you're asking some mashup spirit of its average data labeler.
People Have Too Inflated Sense of What It Means To...
Andrej Karpathy
...catch up on these, and many more highlights
Readwise