loyalty comes from investing in people before it was obvious you should

the benefits of new technology are always disperse among billions and the drawbacks are loud vocal and concentrated among displaced workers or specific systems tipped over the edge

if reincarnation is true a life becomes ergodic and you can take bigger better risks

choosing a hill to die on is so much more honorable than looping in the strange attractor between the hills

adhd is a mid game build

getting riled up thinking about Meta wasting all those beautiful H100s

guy who is cursed to narrate perfectly but never participate

“nature has never produced a wheel” heads btfo’d

crimson sky

jensen loves julius

Transformers may not seem so dangerous now, but with current progress I expect that within 10 years they'll be able to seamlessly disguise themselves as cars

one observation from SF so far: Tenderloin etc got nothing on NYC. much less harassment, much less active schizo, seems mostly just down and out + drugged out. i feel far less acute perception of risk, despite quiter streets (which in NYC is a very high risk signal).

- language gives us some level of CEV for free
- crossed the superintelligent threshold on certain types of tasks without requiring theoretical alignment guarantees to make safe
- any type of self improvement takeoff will involve humans in the loop at first

reasons to be optimistic about alignment:
- even “emergent” capabilities arise continuously/gradually
- current generation rlhf generalizes far better than anyone had guessed
- iterative deployment is ever more iterative as labs inch progress publicly

FTC please look into mkbhd.

this is such a hot take that I have to respect it