svnt

The more common english term is “held back” instead of gated.

And nvidia probably doesn’t want to say “probabilistic responses including hallucinations preclude reliable ai analytics”

As it is the sentence communicates very little. Things are unreliable for a lot of different reasons. Hardware? Software? Users? Intentional and required use of RNGs? Claude presumes the reader has all the same weights activated.

Aging and cancer have some shared metabolic pathways, but your framing is an extreme oversimplification.

Consider the accumulation of mutation that would occur in a cell not subject to aging.

You could just as easily find support for the counterposition: aging, in a sense, prevents cancer. [0]

0: Judith Campisi, "Senescent cells, tumor suppression, and organismal aging: good citizens, bad neighbors," Cell 120:513–522 (2005). https://pubmed.ncbi.nlm.nih.gov/15734683/

Edit: you edited your comment to talk about mitochondria transplants, but the science I am aware of actually points in the opposite direction: https://pubmed.ncbi.nlm.nih.gov/25565207/

Do you have supporting evidence?

How many parameters are you planning to use for your doc editor? Email prose is optimal on a different model. Web parser? You might only be able to do one at a time with switching, which will work, but could be tiresome.

Don’t eat Jimmy!

I’m sorry that has been your experience. The language, like anything available and useful, can be used by anyone, and for any purpose.

It is sad that it has been abused by previous generations and in corporate uses to the point where we can no longer even receive emotional support without suspicion of intent.

I hope you are able to find a community where people who say that, feel it, and mean it, and partner with you to resolve the underlying issue.

A good experienced designer who has done esp32 before will have all these modules or similar enough in a library and can just plug them together.

Possible in a day depending on risk tolerance, almost always better imo to sleep on it. There would be no iterations.

But they will not likely do it happily in that timeframe. And you would practically never get them to quote it at one day, especially not without an inflated rate.

Yeah the framing seems inverted. We have a faster system 2 with llms, and system 1 is what decision models are hoping to be, but so far they just seem to be system 2 models without language output.

The terms don't seem to translate well, possibly because they were the wrong framing to begin with. Possibly also because the billions of years of progressively complex evolutionary constraint that brought about animal cognition does not seem likely to be reproduced by running backprop on human language artifacts.

I understand your point, but the context is that repeatedly, experiments have shown that humans inject too much pattern to be good rngs. It turns out that is visible with training after the fact.

Yes, it is possible that any two 50-sequence patterns are actual rng. But some of them are much more likely, and we can tell.

The issue the anecdote is reflecting is that across the data of 50 naive made-up coin tosses there are too many alternating sequences and probably zero HHHHHH.

It sounds an awful lot like the halting problem.

But they are the ones releasing an unverifiable (no model release) and massive and unchecked body of mathematics into the public while making exaggerated claims about its capabilities to replace human work. And they are the ones doing it in advance of a fractional sale of the company to the public.

Fable will already do this, just very subtly.

Gemini has been like that. One early model I used I asked it about another company’s model. It insisted it didn’t exist. I told it when it was released and to search. It did, and found results, but told me it wasn’t really a public model and I couldn’t use it. I screenshotted the pricing page. It told me I got that from somewhere else, it couldn’t be correct, because it couldn’t access that page.

It is one thing to avoid sycophancy, but Gemini often seems to go well beyond that.

Your statements here reflect a fairly typical reductive understanding of not only mystical states, but entire other areas of empirical science that you apparently don’t understand or can’t put in words.

How, precisely, are you linking information theory to this?

I didn’t claim a side for you, you created that. From my perspective it is one system. Right now you are behaving as a person running an incentivized and specialized negentropic process. You are unknowingly chasing theological propositions and claiming victory every time the perspective can be durably shifted.

To move it back into a more familiar frame: theology loads in the low res view and you claim victory while most of what you did was disaggregate pixel values, but collectively a lot of people like you can form new more accurate images in certain parts of the low res view.

While you move through doing that, though, adjacent areas open up where you have no purchase, religion steps in again, and the process repeats.

There is a core that doesn’t disaggregate. That is where the pixels are coming from. That is what you cannot see when all you want is higher resolution.

Beginning at the end of your comment and working backward, what do you think of Gödel Escher Bach or even Logicomix? It isn’t really a matter of debate that you cannot use logic as a universal construction toolkit for a rational reality, it has already been disproven. Gödel vs Bertrand Russell and Alfred Whitehead etc.

How do you propose causation emerges exactly? You are a kind of relationist?

You seem to be conflating “doesn’t have reasons” with “is random” because your basis seems be something like a relational rationalist perspective, which I don’t know of as a thing that exists. I don’t believe I have encountered it before.

I can see that some aspects of your perspective make sense, but I don’t see how the aggregate holds together.

edit: I forgot to answer “how so”: your argument against free will seems to be that any choice is in response to a cause, and causes infinitely regress, therefore free will cannot exist.

Both move the choice (or the cause), out of you or me and into the origin of existence. It doesn’t cease to exist, really, it just moves. But your argument, used to argue for the lack of existence of choice, can be used to argue for the lack of existence of cause, because every cause is caused. The term becomes meaningless then, one static continuation.

My argument is that the infinite regress ends in you (and me), because of the details of our construction.

I can’t seem to reply to your comment deeper in the chain.

Starbuck collected data and William James turned the data into a compelling book, The Varieties of Religious Experience. Varieties draws on Starbuck's questionnaire data and names him dozens of times. Some of James’ main arguments were:

1) Against "medical materialism": a belief's origins don't settle its value. If Paul's vision is discredited by its physiology, so is the materialist's own conviction.

2) Mystical states: they carry authority for those who have them but don't bind others. Still, they break ordinary rational consciousness's claim to be the only authoritative kind. The sheer numbers of these, and the intact they have on healthy peoples’ lives, make it hard to argue rationally on an empirical basis that there is nothing there.

3) Structurally the conscious self is continuous with a "wider self" through which religious experiences come, which James calls the "more."

4) "The Will to Believe" (1896): Clifford's rule against believing on insufficient evidence is itself a personal preference, choosing to avoid error over gaining truth.

Following on that: your desire to be a reliable producer of material goods outweighs your desire to understand things that require you to take risks in assuming potentially false viewpoints. This is not common sense (whatever that means), it is an epistemological choice that seems common to you because of your educational and professional environment.

Have you read William James/Starbuck, and how do you interpret that?

I’ve read Schopenhauer but this is not complete in my view. He actually advocates for the will as a transcendental, escaping this dilemma.

I have been working on this quite a bit, maybe I can publish it.

One challenge with your approach is that decisions are frequently (if not always) multi-causal in origin, and the balance of influence of causation must be something we can weight in real time.

The bigger basic structural idea is that closures escape causal chains, and freewill is essentially what happens within a (specific type of) causation closure.

A further problem is attempting to enclose and limit reality within logic, which we know to be a finite rule system and subject to specific limitations as such.

Further, your same argument could be used to support the idea that causation doesn’t exist.

It seems pretty clearly structured to be on each person to explore experientially and explain it to themselves.

But the more things a person must close themselves off from, the less open their consideration of possibility, the more I tend to see their worldview as subservient to something else.

In this part of society it tends to be in support of rapid, high reliability problem-solving, and rational explanation thereof.

I responded with an informative post but apparently made the mistake of disclosing it was generated and it was flagged. I reviewed the post and it is correct. If you cannot see it, it says you are wrong/incomplete here and cites three studies I know of:

Beniaguev, Segev & London, "Single cortical neurons as deep artificial neural networks," Neuron, 10 August 2021, Volume 109, Issue 17, pp. 2727–2739.e3 https://doi.org/10.1016/j.neuron.2021.07.002 Open-access preprint: https://www.biorxiv.org/content/10.1101/613141v2

Gidon et al., "Dendritic action potentials and computation in human layer 2/3 cortical neurons," Science, 367(6473), 83–87 (2020) https://doi.org/10.1126/science.aax6239

Randi, Sharma, Dvali & Leifer, "Neural signal propagation atlas of Caenorhabditis elegans," Nature, 623(7986), 2023, pp. 406–414 https://doi.org/10.1038/s41586-023-06683-4 Open-access preprint: https://arxiv.org/abs/2208.04790

Neurons are not individually simple, in any sense. Nor are they the only class of cell involved in cognitive processes — glial cells are involved in inhibition, memory, network modification, etc.

Citation needed. If what you are actually saying is successful empirical science only produces empirical observations about physical processes, you are willingly wearing blinders.

You just moved the magic to “being aware” — now the question is only what has awareness instead of what has consciousness.

It wasn’t the pope, but it is mentioned by I think a rabbi in the article. He relates telling Olah if he thinks models are conscious he should be fighting the South, freeing the slaves.

> Also your comment about us not understanding how a neurone works does vast injustice to the amount of research that has been done in this field. At this point in time we do have a very good understanding of how a neurone works. What we don’t understand is how consciousness arises from it. And that’s the very thing you seem to be arguing that there’s no point researching.

This is the part you keep stating that is wrong. We know a lot about the most obvious electrical way that neurons act. That is nowhere near having a good understanding of how they work.

So you are saying it is deterministic but unknowable in advance.

This does not prevent someone from building a you that runs on a faster substrate and pre running you.

It also gets into sticky territory on responsibility. Are you responsible for your choices?

See also Gödel.

I do not understand how your argument works.

You cannot choose, but you can prechoose? You cannot choose, but your physical identity chose the moment it was deterministically brought into being?

What does choice even mean in this context?

I understand the premise, but I believe success depends on you being able to effectively sort problems for other frontier models better than the frontier.

Micron is the only one to bail, and they are the smallest footprint. No one else has bailed.

Samsung recently shifted some capacity from HBM back to DDR5 modules, freeing up around 80,000 wafers/month, so DDR5 margins are currently good enough to compete with HBM for wafers, or maybe they are anticipating a turn, or both.

If you are training on data labeled by frontier models, how do you expect to exceed the performance of frontier models, other than in the cost dimension by recognizing simpler problems and routing to cheaper models?