About as solid as those of the Ultraviolet Catastrophe, I've been known to say.

It's called the Turing barrier, and is not specific to deep learning.

Today I'm doing my best to explain counterfactual quantum computation or how to compute with a computer that doesn't compute. No, not a joke! Bear with me while I plot out the beam splitters and mirrors... youtube.com/watch?v=dQIfSr…

wow guys, i was skeptical at first but it really seems like AI is the future

LeCun throws in the towel with current approaches.

This is the way. Fund radical science now. youtube.com/watch?v=JImCFs…

All unit tests passing.

The bait and switch behind AI risk prediction tools open.substack.com/pub/aisnakeoil…

If this Turing AI had actually achieved understanding, as Penrose claims is impossible, it would have said, "Penrose needs no recommendation".

Every talk on AI should have machine generated subtitles, just to keep things in perspective.

Frontiers in Psychology Artificial Intelligence Is Stupid and Causal Reasoning Will Not Fix It frontiersin.org/articles/10.33…

“The human mind is not, like ChatGPT and its ilk, a lumbering statistical engine for pattern matching,” write Noam Chomsky, Ian Roberts and Jeffrey Watumull. nyti.ms/41SD1Lu

A microgram is getting very close to the 'Planck mass' needed to test Penrose's alternative to quantum theory. phys.ethz.ch/news-and-event…

Transformers slapped down by the Turing Barrier.

The Hype Cycle has grown together with tech valuations: No longer a peculiarity of enthusiastic nerds, it is now the very engine of global finance. One that never self-corrects on past failure. Interesting PALLADIUM Magazine article. palladiummag.com/2023/08/03/you…

New paper by Google provides evidence that transformers (GPT, etc) cannot generalize beyond their training data

Quantum Tunnelling is a fascinating process in which a particle goes through a barrier faster than the speed of light. Or maybe it doesn't. A clever, new experiment could finally end the debate once. youtube.com/watch?v=2kTxX-…

Newsletter: There is no AI Revolution. Consumer adoption of generative AI outside of ChatGPT is barely 100 million MAUs, every single company in generative AI is deeply unprofitable, Microsoft is pulling capex, and OpenAI spent $9bn in 2024 to lose $5bn. wheresyoured.at/wheres-the-mon…