AI - AI Dynamics

→ View original post on X — @patio11

AI Accountability: Legal Compliance with Government Demands

By

@patio11

–

06 November 2023 8h00

Every penny and/or yen that the people of the United States and Japan have, through their elected representatives, deemed fit to demand of me; not more than that.

Open-source LLMs surge: Can AI R&D regulation work?

By

–

06 November 2023 6h21

High-performance open-source LLMs popping up around the world.
Does anyone believe regulating AI R&D can be effective?
(Assuming it's useful. Which it is not).

→ View original post on X — @kaifulee

China’s Yi-34B Model Challenges US LLM Ecosystem Leadership

By

@kaifulee

–

06 November 2023 6h08

I must say China is now behind the U.S. in the LLM ecosystem momentum. The underlying vision for me to start @01AI_Yi is to make better AI accessible to more people. We are glad our first moderate-size Yi-34B performs competitively at a global level. More to come soon after

FSD Falls Short of Human Learning Abilities Despite Data Investment

By

–

06 November 2023 6h08

FSD has required enormous amounts of data, years of engineering to take care of corner cases, and the reliability is nowhere near that of a human (though it's clearly useful. I use it).
Clearly, we are very far from the learning abilities of a 17 year old.

How Do Humans Learn World Models?

By

–

06 November 2023 6h03

Exactly.
That's my whole point.
How do humans learn world models?

Joint Embedding Architectures and Non-Contrastive Learning Methods

By

–

06 November 2023 5h30

Well, technically, my first paper on Joint Embedding Architectures (AKA Siamese nets) is from NIPS 1994.
but that paper and many subsequent works use sample-contrastive learning and no predictor.
The JEPA idea trained non sample-contrastive losses (Barlow Twins, VICReg, MCR2) or

→ View original post on X — @jeremyphoward

Inquiry about tokenizer capabilities and features

By

@jeremyphoward

–

06 November 2023 5h29

Thanks for all this great info Stella. I'm not familiar with that tokenizer — what's special about it?

LLMs Retrieval vs Reasoning: Understanding Fundamental Limitations

By

–

06 November 2023 5h20

Don't confuse the approximate retrieval abilities of LLMs for actual reasoning abilities.

Censorship and Knowledge Exchange Control in Authoritarian Regimes

By

–

06 November 2023 5h14

In 1983, when the free world was starting to play with personal computers, the Ceaucescu regime in Romania required a license to own a typewriter.
Obscurantism isn't just preventing people from accessing knowledge.
It's also preventing people from exchanging knowledge.