Sam 🌱

I like to learn about things that keep the world moving.
mechatronics student 🔧

[Sam 🌱 @samsyntaxerror]

to E, or not to E, that is the question

[Sam 🌱 @samsyntaxerror]

I love how we can put things in space and everyone just pretends it's normal

[Sam 🌱 @samsyntaxerror]

@a__tomala the amazing digital gradients

[Sam 🌱 @samsyntaxerror]

this dude just seriously came over to my desk and said:
"so what's the difference between donald trump and me?"

[Sam 🌱 @samsyntaxerror]

there's no way out of population collapse at this point in time

the only way to achieve escape velocity is through technology

[Sam 🌱 @samsyntaxerror]

worst than not being able to sleep is BEING able to sleep and see people arguing about fucking mermaid charts

[Sam 🌱 @samsyntaxerror]

atp it's very safe to say that self-recursive improvement can happen with current models

i'd bet that even if we freeze all sota models right now, with the right harness & compute we could probably find a new architecture that overthrows the transformer
https://t.co/sLjWIQ1SoN

[Lisan al Gaib @scaling01]

> be me
> build "AGI" benchmark
> actually version 3 already
> we don't talk about 1 and 2
> (they saturated in a year)
> invent new scoring method
> if human scores above AI, use squared efficiency
> example: human took 10 steps to solve level
> AI took 100 steps to solve a level
> then AI's score is (10/100)
> correction, add square penalty to be extra cheeky
> (10/100)^2
> AI scores plummet
> smug_face.jpg

> observe that AI is scoring higher than human on some levels
> "damn"
> "CLAMP THE SCORES!!!"
> new metric now takes minimum of 1 and AI's score
> AI literally NOT ALLOWED to score higher than human
> meanwhile human get squared rewards
> doubly_smug_face.jpg

> AI thinks at 200 tokens/s
> human thinks at 5 tokens/s
> cries about it
> "UNFAIR ADVANTAGE"
> add restriction that models can't use more than 5X the steps that a human used

> DeepThink, Pro and xhigh models exist that think longer to score better
> excluded them

> DECLARE AI HIT A WALL

[Maxime Chevalier @Love2Code]

Somehow our nerve impulses travel at mere meters per second, which is an obvious disadvantage in hunting or physical combat. Evolution never figured out how to build better wires, but our brains surely run on quantum photonic memristors guys...

[Sam 🌱 @samsyntaxerror]

been thinking about using pulsars for deep space navigation and it was already a thing lol

demo'd by NASA some time ago

[Sam 🌱 @samsyntaxerror]

the good thing about the Epstein files is that you can clearly see who's epistemic model is based on what everyone is saying rn

[Sam 🌱 @samsyntaxerror]

posted some slop
https://t.co/RJkWfGSAqQ

[Sam 🌱 @samsyntaxerror]

https://t.co/Ql15qHmcHy

[Sam 🌱 @samsyntaxerror]

https://t.co/Ql15qHmcHy

[Sam 🌱 @samsyntaxerror]

cognitive dissonance hypothesis

[Sam 🌱 @samsyntaxerror]

sha256 commit: 96ff89e5630466a3dc0ae0a965ecea22a28ea19d180cc13eff98b1e60f634f42

[Sam 🌱 @samsyntaxerror]

I dreamed that I was chasing a penguin inside a mall what does this mean

[Sam 🌱 @samsyntaxerror]

so redundant it hurts but I want to publish this old draft anyways
https://t.co/r8Z0bmbUvf

[Sam 🌱 @samsyntaxerror]

I remember this game that asked you if you would either feel bad or disgusted if you stepped on a snail barefoot

it called you a liar if you felt bad, but I do truly feel bad when I step on a snail barefoot

makes me wonder the proportion of oversocialized people that do lie

[Sam 🌱 @samsyntaxerror]

@groas_ai @scottastevenson why

[Sam 🌱 @samsyntaxerror]

wtf is happening on threads lmao

[Sam 🌱 @samsyntaxerror]

@cryptidcynic fair enough, i think it's important to publish results and to suggest possible causes. but that's not what the abstract is selling

[Sam 🌱 @samsyntaxerror]

you can just do linear regression on a point cloud with 0.001 correlation, draw conclusions from it and get published

this is the state of science rn lol
https://t.co/TpAdSftB8n

[Crémieux @cremieuxrecueil]

Smarter ADHD sufferers tend to get diagnosed with ADHD at later ages.

Consistent with the idea that smarter people are able to compensate for ADHD-related deficits, being smart seems to be 'protective' for noticing that a person has the condition.