Intelligence is similarity + compositionality. Transformers' similarity is dot product of embeddings and their compositionality is attention, both still very primitive.
@pmddomingos
-
LLMs Face the Same Problem as NLP: Unclear Use Cases
By
–
The problem with LLMs is what it's always been with NLP: if people can't tell what they're good for and not, they'll wind up not using them for much.
-
LLMs Improving Faster Than Academic Publishing Quality
By
–
Considering the speed at which LLMs are getting better and Nature/Science worse, it won't be long before they cross.
-
Algorithms Outpace Moore’s Law in Computing Speed
By
–
Better algorithms have sped up computing by more than Moore's law.
-
Hardware exponential growth versus linear software and human progress
By
–
Hardware progresses exponentially, software progresses linearly, and humans don't.
-
Geometric Deep Learning and Symmetry-Based Learning Future
By
–
Geometric deep learning is a type of symmetry-based learning. Lots more will happen in this space.
-
Human-produced Internet content will approach zero
By
–
The fraction of Internet content produced by humans will inevitably dwindle to near zero.
-
Earth as Evolution’s Data Center with Biological Processing Units
By
–
The Earth is evolution's data center, and it has trillions of BPUs (biological processing units).
-
Tesla’s Superior Software-Intensive Car Manufacturing Capabilities
By
–
Tesla knows how to make software-intensive cars better than anyone in the world, for one.
-
Criticism of Elon Musk Clouded by Bias and Lack of Objectivity
By
–
I’m not an Elon fan: I’ve repeatedly criticized Tesla, his views on AI, etc. The problem is you’re so biased you can’t even see straight.