“Read, but maybe not fully understood” column.

“Read, but maybe not fully understood” column. Issue 2. From my phone.
1/
A classic paper (https://arxiv.org/pdf/1911.01547)
on what AGI is and how you can even make sense of it.
2/ I rarely use frameworks when working on a product. But
this one (https://www.lennysnewsletter.com/p/how-duolingo-reignited-user-growth)
is probably the best I know for working on retention. The whole idea that you should pick metrics to match the use case and do pattern matching across a bunch of benchmarks is just chef’s kiss. And overall the article is great.
3/ I just like reading Quanta Magazine. Don’t know why. This
time (https://www.quantamagazine.org/the-s-matrix-is-the-oracle-physicists-turn-to-in-times-of-crisis-20240523/)
it’s about s-matrices of probabilities.
4/ a lot of scientists I respect believe that AGI or just the next-level lil model will show up when they combine deep learning and
this stuff (https://people.csail.mit.edu/asolar/SynthesisCourse/index.htm)
5/
some thoughts (http://www.incompleteideas.net/IncIdeas/BitterLesson.html)
from a smart dude about how, like, search and learning will always be the methods
6/
an old little article, (https://www.wheresyoured.at/the-men-who-killed-google/)
but I love coming back to how easy it is to add a couple billion in profit at Google