Hacker News: rdlecler1

New comment by rdlecler1 in "You Can Just Buy Far-UVC"

rdlecler1 — Wed, 14 Jan 2026 20:49:41 +0000

that wavelength penetrates the skin. you need to be around 222nm for human safety

uviquity has prototypes of a 220nm solid state chip they’ll commercialize next year (we’re an investor). a single far-uvc photon will destroy the covid virus.

https://uviquity.com/

New comment by rdlecler1 in "The Curious Similarity Between LLMs and Quantum Mechanics"

rdlecler1 — Tue, 11 Feb 2025 19:20:54 +0000

You clearly have deep understanding of base reality and the emergence of complexity. So please share an actual argument.

The Curious Similarity Between LLMs and Quantum Mechanics

rdlecler1 — Tue, 11 Feb 2025 16:24:30 +0000

Article URL: https://robleclerc.substack.com/p/the-curious-similarity-between-llms

Comments URL: https://news.ycombinator.com/item?id=43014716

Points: 16

# Comments: 14

Super Intelligence ≠ Hyper Intelligence

rdlecler1 — Mon, 10 Feb 2025 21:38:56 +0000

Article URL: https://robleclerc.substack.com/p/super-intelligence-hyper-intelligence

Comments URL: https://news.ycombinator.com/item?id=43005528

Points: 3

# Comments: 0

New comment by rdlecler1 in "Elon Musk wants to dominate robotaxis–first he needs to catch up to Waymo"

rdlecler1 — Thu, 10 Oct 2024 22:50:10 +0000

It’s the other way around. Waymo may do the job exquisitely, but exquisite is also expensive. Telsa has always constrained their system to lower cost hardware. It’ll be easier for Tesla to eventually punch through than for Waymo to cut their costs to that of a Tesla robotaxi.

New comment by rdlecler1 in "The AI startup drama that's damaging Y Combinator's reputation"

rdlecler1 — Thu, 03 Oct 2024 13:37:06 +0000

Power eventually turns a high trust society into a low trust society when that power is used to one’s own advantage.

New comment by rdlecler1 in "Apple removes nearly 100 VPNs used by Russians to bypass censorship"

rdlecler1 — Wed, 02 Oct 2024 01:30:21 +0000

Think different. No kidding. No one is going to think this.

New comment by rdlecler1 in "YC criticized for backing AI startup that simply cloned another AI startup"

rdlecler1 — Tue, 01 Oct 2024 16:39:15 +0000

YC is putting through 200+ companies per batch and making a tiny investment. It’s far more efficient to swallow the occasional bad bet than to do a lot due diligence… Now the VCs that came in after… that’s another story.

New comment by rdlecler1 in "If AI is helping people code better, why aren't products getting better?"

rdlecler1 — Mon, 23 Sep 2024 15:20:57 +0000

This is a great point. LLMs may flood an app with more features without making it better. At the end I do the day you still need to make something people love. AI may help you build that faster if you know what do build, but it’s not going to make a bad product great.

New comment by rdlecler1 in "Myocarditis complications more common after Covid infection than vaccination"

rdlecler1 — Tue, 27 Aug 2024 04:43:28 +0000

Remember there is still survival bias here. We’re not seeing the complications from people who died of the infection.

New comment by rdlecler1 in "America’s Transit Exceptionalism"

rdlecler1 — Tue, 23 Jul 2024 23:03:05 +0000

Do driverless cars make all this moot?

New comment by rdlecler1 in "General Theory of Neural Networks"

rdlecler1 — Mon, 15 Jul 2024 23:42:09 +0000

Super interesting.

New comment by rdlecler1 in "General Theory of Neural Networks"

rdlecler1 — Mon, 15 Jul 2024 21:18:07 +0000

This sits in a larger field of complexity theory and complex adaptive systems. There was also some interesting work on “Artificial Life” although that research program seems to have fallen out of favor. My introduction in 1995 was the book Chaos and then Stuart Kauffman’s At Home in the Universe. Wolframs New Kind of Science was also interesting.

New comment by rdlecler1 in "General Theory of Neural Networks"

rdlecler1 — Mon, 15 Jul 2024 19:18:39 +0000

I could have bogged the essay down with qualifiers to address all the potential straw man objections, but that didn't seem productive. It's easy to take an uncharitable view on this, but I do explain more about GRNs later in the essay. I worked with them for 8 years, and yes, they do act like the rudimentary brains of the cell, and that's the reason this system is selected again and again by evolution.

New comment by rdlecler1 in "General Theory of Neural Networks"

rdlecler1 — Mon, 15 Jul 2024 18:02:35 +0000

I'm not arguing that this approximator is necessary (not sufficient) for this class of networks. I've proposed some conjectures on what we might expect to see, but there are certainly other salient ingredients and common principles that we haven't discovered, and I think it's important to hunt for them.

New comment by rdlecler1 in "General Theory of Neural Networks"

rdlecler1 — Mon, 15 Jul 2024 17:40:54 +0000

This is a great question, and I don't yet have an answer. I'm going to butcher this description, so please be charitable, but functionally, the attention mechanism reduces the dimensions and uses the coincidence between the Q and K linear layers to narrow down to a subset of the input, and then the softmax amplifies the signal.

One unsatisfying argument might be that this might fall into implementation details for this particular class. Another prediction might be that an attention mechanism is an essential element of these networks that appears in other networks of this class. Another is that this is a decent approximation, but has limitations, and we'll figure out how the brain does it and replace it with that.

New comment by rdlecler1 in "General Theory of Neural Networks"

rdlecler1 — Mon, 15 Jul 2024 16:03:32 +0000

More generally there’s graph neural networks, for instance, but not you’re including many dynamic networks that are not open-ended or evolvable. The idea is to identify common dynamics and add constraints on the types of networks that are included to find general principles within that class. Kisen the constraints, you make the class too broad and can’t identify common principles.

New comment by rdlecler1 in "General Theory of Neural Networks"

rdlecler1 — Mon, 15 Jul 2024 15:11:03 +0000

Genetic systems code gene regulatory networks. I spent most of the essay on them.

New comment by rdlecler1 in "General Theory of Neural Networks"

rdlecler1 — Mon, 15 Jul 2024 04:52:04 +0000

Those neurons are being trained the day we were born. Reality corresponds to about 11 million bits per second. What I suspect’s happening is that we train higher and higher levels of abstraction and we get to a point where new knowledge is involves training a new permutation of a few high level neurons.

New comment by rdlecler1 in "General Theory of Neural Networks"

rdlecler1 — Mon, 15 Jul 2024 04:46:59 +0000

Activation functions are implementation details. See appendix for the general formula.