happy pride

 hiiii here's an over due post (will maintain daily posts !)

on why we need ai safety - specification gaming article by google deepmind 

here's a quote that sums up the whole thing - reward function / intent engineering is a big thing

• How do we faithfully capture the human concept of a given task in a reward function?
• How do we avoid making mistakes in our implicit assumptions about the domain, or design agents that correct mistaken assumptions instead of gaming them?
• How do we avoid reward tampering?

vanishing gradient problem - occurs potentially when u use sigmoid activation, slower learning in beginning --> use relu

sum-free sets and the beauty of a grad student expanding on other's work, smth about littlewood norm and fourier analysis being powerful

reverse computing -- potential benefits for energy conservation, has been theorized and engineers are trying to make it happen! this feels like a strange but very interesting idea and in the end its sorta a tradeoff like everything is

had a talk about the meaning of life with some friends / how to be better people

i'm also trying to get into the arena curriculum :P i'll need to spend more time

got pitholed by some math proofs

and now i'm using a key to timetrack for accountability because its summerrrr and i want to learn lots of cool things / work on cool tings

1. mo

2. ar

3. nt

4. cst 

5. as

6. cm 

7. hm

8. mv

9. rb

10. se

11. wr

12. lb

13. ex