lukeplato

6 ms·

1.
On Proof and Progress in Mathematics (1994) [pdf](math.toronto.edu)
2 points by lukeplato 5d ago|0 comments
2.
by lukeplato 5d ago
Many comments seem to be misinterpreting this, but all this seems to be suggesting is that the AI companies should not be treating mathematical milestones as benchmark targets but need to release the cumulative findings in a human interpret
3.
The Mantle(johncarlosbaez.wordpress.com)
3 points by lukeplato 25d ago|0 comments
4.
by lukeplato 1mo ago
related paper on the long-horizon AI research system used: https://arxiv.org/pdf/2608.11195
5.
by lukeplato 2mo ago
they're probably referring to CASP13/14
6.
From Quantum Relative Entropy to the Semiclassical Einstein Equations(journals.aps.org)
2 points by lukeplato 2mo ago|0 comments
7.
Naturally Occurring Quasicrystals(johncarlosbaez.wordpress.com)
134 points by lukeplato 4mo ago|11 comments
8.
The Pentagon threatens Anthropic(astralcodexten.com)
191 points by lukeplato 7mo ago|126 comments
9.
Unidentified Aerial Phenomena Near Nuclear Facilities – Enigma Labs(enigmalabs.io)
1 points by lukeplato 8mo ago|0 comments
10.
The universal weight subspace hypothesis(arxiv.org)
358 points by lukeplato 9mo ago|132 comments
11.
Active open mindedness and mindfulness – Vervaeke(youtube.com)
2 points by lukeplato 10mo ago|0 comments
12.
Token Embeddings Violate the Manifold Hypothesis(arxiv.org)
3 points by lukeplato 10mo ago|1 comments
13.
by lukeplato 1y ago
I think an unrestricted model with sufficient inference time compute could create unprecedented catastrophes. Not sure why you would be opposed to whistleblower protections
14.
The Information-Computational Universe (ICU) Theory(zenodo.org)
3 points by lukeplato 1y ago|0 comments
15.
The Claude Bliss Attractor(astralcodexten.com)
71 points by lukeplato 1y ago|87 comments
16.
Andromeda's asymmetric satellite system challenges cold dark matter cosmology(nature.com)
3 points by lukeplato 1y ago|0 comments
17.
Nuclear Explosions for Large Scale Carbon Sequestration(arxiv.org)
6 points by lukeplato 2y ago|1 comments
18.
by lukeplato 2y ago
hopefully nuclear power plants
19.
by lukeplato 2y ago
Michael Edward Johnson has an interesting theory called vasocomputation with the core hypothesis: > vasomuscular tension stabilizes local neural patterns. A sustained thought is a pattern of vascular clenching that reduces dynamic range
20.
by lukeplato 2y ago
I don't see why a mixture of experts couldn't be distilled into a single model and unified latent space
21.
by lukeplato 2y ago
reading this thread was a good reminder that being intelligent but closed off to alternative hypotheses is the same thing as being ignorant
22.
Explore new styles of music with Suno V4(suno.com)
4 points by lukeplato 2y ago|0 comments
23.
Artificial Kuramoto Oscillatory Neurons(arxiv.org)
4 points by lukeplato 2y ago|0 comments
24.
ESM3 Preprint [pdf](evolutionaryscale-public.s3.us-east-2.amazonaws.com)
3 points by lukeplato 2y ago|0 comments
25.
by lukeplato 2y ago
I remember reading that for k-d trees to be able to split on k dimensions the dataset needs to be > 2^k, which becomes unwieldy pretty quickly
26.
by lukeplato 2y ago
This could make for an interesting UI for exploring clusters in data. I only wish K-d trees could handle higher dimensions
27.
by lukeplato 2y ago
interatomic potential is a promising scale and there's been some interesting recent developments using equivariant graph neural networks: https://www.nature.com/articles/s41467-022-29939-5 https://arxiv
28.
by lukeplato 2y ago
When is a good time to buy AAPL on the dip?
29.
Interpolation in Statistical Learning, Or, Memorizing the Training Data(bactra.org)
3 points by lukeplato 3y ago|0 comments
30.
by lukeplato 3y ago
you should use a throwaway account without your name so as to not impact your company

More ›