Jay @treeinnauvis

Builds brain-like AI. Joined November 2023

Tweets

353
Followers

27
Following

296
Likes

634

Jay @treeinnauvis

a day ago

Can’t come up with a good reason, but for my use case, ChatGPT is still better than any other models.

0 0 0 6 0

Jay @treeinnauvis

4 days ago

Causal compression is the best compression

0 0 0 5 0

Many AI researchers are taking the term “world model” too literally. In order to make a system to have good world model, one doesn’t need to model the world explicitely. Instead, representation of the causal structure of the world would be sufficient.

0 0 0 12 1

Beff – e/acc @BasedBeffJezos

a week ago

Being a founder is performing active Inference at civilizational scale. You have to steer the world to your thesis of how it could be better.

Alexis Ohanian 🗽 @alexisohanian

a week ago

Being a founder is performing active Inference at civilizational scale. You have to steer the world to your thesis of how it could be better.

72 165 1K 186K 285

12 25 158 11K 11

Melon Usk - e/uto (commentary) @MaskedMelonUsk

2 weeks ago

Quantum ethics is fun but a bit tricky to explain (work in progress but the general idea is in place): So it's the most minimal quantum ethical model based on Wolfram physics model, it's intentionally simplified to be fast to compute and understand: The goal is to instantly…

3 4 16 839 4

Download Image

fish @fishPointer

3 weeks ago

trillions of dollars spent on AI and the only interface is still a chatbot. still no branching canvas, no mindmap, no dimensionality, no forking, no inversion, no merging. no custom functions. no style and no creativity

704 1K 25K 1.5M 5K

Eren Bali @erenbali

3 weeks ago

With AI, everybody will make their own software just like how we all 3D print our own furniture.

361 2K 27K 1.1M 2K

Melon Usk - e/uto (commentary) @MaskedMelonUsk

3 weeks ago

The trippy one: Are you a multiverse yourself?) And do we live in 2 multiverses - one inside of another - not just one? ;-) Let me explain: Have you ever had this feeling: You know there is a spoon 🥄 on the table behind you, so you reach out there without looking… But alas…

2 2 10 149 2

Download Gif

Gamma Factor @yFactr

4 weeks ago

thanks to AI, anyone with $20 in their pockets can acquire forms of psychosis previously only available to billionaires and dictators

fish @fishPointer

4 weeks ago

thanks to AI, anyone with $20 in their pockets can acquire forms of psychosis previously only available to billionaires and dictators

13 8 223 20K 8

5 73 494 12K 15

Melon Usk - e/uto (commentary) @MaskedMelonUsk

4 weeks ago

If you can imagine all the universes in your head, you’re max-intelligent, nothing can be smarter than you. That is AXI - Artificial maX-Intelligence. And it is so perfect it is a place, not an agent. Agent is by definition limited - still hasn’t reached the ends of all…

1 1 7 102 0

Jay @treeinnauvis

a month ago

If quantum immortality is true, then P=NP for the observer.

0 0 1 25 0

Jay @treeinnauvis

a month ago

Why isn’t there a one giant library of EVERY (implemented) useful functions and the IDE just locally caches the needed functions? Everyone can contribute to everyone that way.

0 0 1 33 0

Jay @treeinnauvis

a month ago

The following sentence came to me in a dream: Data builds questions that are founding themselves to ask them to. And that question is now disassociated from the original question

0 0 1 27 0

Jay @treeinnauvis

2 months ago

Maybe architectural breakthrough is not happening because deep learning sucks.

0 0 0 24 0

Albert Gu @_albertgu

2 months ago

Tokenization is just a special case of "chunking" - building low-level data into high-level abstractions - which is in turn fundamental to intelligence. Our new architecture, which enables hierarchical *dynamic chunking*, is not only tokenizer-free, but simply scales better.