reshlo

> it’s just a standard s-expression syntax

I.e. they are correct that it is arcane. What percentage of programmers today do you think have ever seen code written in any Lisp dialect, let alone understand it?

You’re still failing to understand that a model being able to output a prediction of something is not the same thing as it “knowing” that thing. The Newton-Raphson method doesn’t “know” what the root of a function is, it just outputs an approximation of it.

> It’s extremely silly then don’t you think to make such bold declarations on what doesn’t have it?

I don’t find it particularly bold to respond to your assertion that a piece of mathematics is sentient life by stating that you haven’t proven that it is, and that in the absence of that proof, the most rational position is to continue to believe that it is not, as we have done for millennia. The burden of proof is on you.

> if it does anything a conscious being does

You haven’t shown that it can do anything that only conscious beings can do.

Being able to generate a passable approximation of text that might follow some prompt doesn’t mean that it understands the prompt, or its answer. As an obvious example, if you give LLMs maths problems, they change their answers if you change the names of the people in the question. They’re not actually doing maths.

> Notice anything? It’s not just that the performance on MathGLM steadily declines as the problems gets bigger, with the discrepancy between it and a calculator steadily increasing, it’s that the LLM based system is generalizing by similarity, doing better on cases that are in or near the training set, never, ever getting to a complete, abstract, reliable representation of what multiplication is.[0]

[0] https://garymarcus.substack.com/p/math-is-hard-if-you-are-an...

The important part is that the algorithm doesn’t either.

You are the one claiming that LLMs are conscious, so it falls to you to prove it.

I argued that LLMs do not have the capacity to have ideas or to know things, and you tried to prove me wrong by providing examples of papers that show, for example, that LLMs have internal states that can be used to predict the likelihood that what they will output will be facts. But that doesn’t disprove what I said, because that’s not what it means to have ideas or know things. By definition, only conscious beings can do those things.

If we run the Newton-Raphson algorithm on some input and it hasn’t run long enough to come up with a good enough answer yet, then we are uncertain about the answer. It is not the case that the algorithm is uncertain about the answer. It would make no sense to make any claims about the algorithm’s level of certainty, because an algorithm does not have the capacity to be certain.

You’re missing my point. Take one of the articles described in that comment, titled “The Internal State of an LLM Knows When It's Lying”. It states “In this paper, we provide evidence that the LLM's internal state can be used to reveal the truthfulness of statements.” Both of these are untrue, for a number of reasons.

- An LLM knowing when it is lying is not the same thing as its internal state being able to “reveal the truthfulness of statements”. The LLM does not know when it is lying, because LLMs do not know things.

- It is incapable of lying, because lying requires possessing intent to lie. Stating untrue things is not the same as lying.

- As the paper states shortly afterwards, what it actually shows is “given a set of test sentences, of which half are true and half false, our trained classifier achieves an average of 71% to 83% accuracy”. That’s not the same thing as it being able to “reveal the truthfulness of statements”.

No intellectually honest person would claim that this finding means an LLM “knows when it is lying”.

It may be terminology that some people use in that way, but it’s becoming increasingly common for people describing LLMs to use such terminology to mean that the LLM literally has the capacity for understanding.

Personally, until recently I can only recall people saying things along the lines of “applying the model indicates that we can state this fact about the data with this much confidence”, never “the model has this much confidence” in some truth statement, especially one independent of its training data.

The point is that neither the model nor the sampler algorithm can possibly have “confidence” in its behaviour or the system’s collective behaviour.

If I put a weight on one side of a die, and I roll it, the die is not more confident that it will land on that side than it would be otherwise, because dice do not have the ability to be confident. Asserting otherwise shows a fundamental misunderstanding of what a die is.

The same is true for LLMs.

A statistical model which is instructed to output the token that is most likely to come next doesn’t have “confidence” in its choice based on the distribution of possible tokens. We might, but it cannot. A statistical model cannot be confident or unsure. It has no mind.

It also has no concept of what it means for the choice of token to be an “error” or not, or what a “correct” answer would be.

The total length of the relevant sections of the Southern Cross Cable is 12,135km, as it goes via Hawaii.

The main reason I made my original comment was to point out that the real numbers are more than double what the other commenter called “devastating” latency.

https://en.wikipedia.org/wiki/Southern_Cross_Cable

I did misunderstand it, because looking at it again now, they spent the entire post talking about how difficult it is to make the cables, except for the very last sentence where they mention die area one time, and it’s still not clear that they’re talking about die area for something that’s inside the computer rather than a chip that goes in the cable.

> Look at the supported lengths and cables. … relatively generic cables. Your USB-C or HDMI connections cannot go nearly as far and require significantly more tightly controlled cables and shielding. … They used a long list of tricks to squeeze every possible dB of SNR out of those cables.

You explained why 10G Ethernet cables are expensive, but why should it be so expensive to put a 10G-capable port on the computer compared to the other ports?

> Anyone who lives in Asia or Australia should relate to this. 100ms RTT latency can be devastating.

When I used to (try to) play online games in NZ a few years ago, RTT to US West servers sometimes exceeded 200ms.

You just used the letter f 13 times to make that file, though.

You might be able to get by in Haskell if you avoid case expressions.

Many websites are using passkeys not as a primary factor, but as the second factor, or as both factors. That implies that they are meant to serve as some combination of “something you are” and “something you have”. The fact that you logged in with one by using biometrics proves both that you are you and that you have your phone. They’re certainly not “something you know” because they are designed specifically so that you are not allowed to know them.

Allowing both “something you are” and “something you have” to be reset simultaneously via proof only of “something you know” (the password to your email account) means that once that reset happens, you’ve gone from two or three factors to one factor.

Allowing passkeys to be reset by email is not compatible with using them as anything other than the primary factor. If you’re using them as both factors, you’d get equivalent security if you implemented sign-in via only magic links. If you’re using them as the second factor along with a password, but you allow them to be reset via email, you actually only have one factor.

Yes.

https://github.com/microsoft/winget-pkgs/tree/master/manifes...

> choco and scoop

And winget.

> In Windows you… google and download random .exes?

Windows has an official package manager now too.

I meant that hiding your boredom from everyone on the call is more professional than revealing you are bored, regardless of whether your camera is on or not.

That’s obviously better than being seen as unprofessional.

Isn’t the whole point of a passkey that it’s meant to use a chain of trust to prove that you’re you via biometrics or a physical factor? I’ve read that they’re intended to remove the need for 2-factor authentication because they are both factors, which implies you shouldn’t be allowed to reset them.

Resetting 2-factor authentication by proving access to only one factor (email) defeats the purpose of requiring 2 factors. If they can be reset via email, they might as well not exist at all. Even if we assume that nobody other than the user has legitimate access to the emails sent to the user (which is often untrue), emails can be trivially intercepted by a third party.

Not to mention that if I’ve lost access to the device where I am signed in to my email account, I won’t be able to access my email account to reset my passkeys anyway, because access to my email account would also require a passkey that I no longer have.

FOMO stands for Fear Of Missing Out.

That’s the main subject of the linked post by Doctorow.

> I mean, isn't the idea from them that you have 2 or more of them?

So now I need to buy an extra phone from a different manufacturer than the one I already own, or sign up for another paid service? I’m starting to see what their motive might be now.

Is it even a requirement of the passkey standard to allow the user to create more than one passkey for your website?

Passkeys are terrifying, and I don’t understand why the companies pushing them are doing so. What are their motives? What do they gain from catastrophically increasing the risk that users completely lose access to our ability to conduct our lives?

If someone steals my phone today, I can still access most of my accounts, and can regain access quickly to the others.

Now let’s assume passkeys are ubiquitous and used to log in to every website.

If they can’t be exported, then your entire digital existence is at the mercy of whatever device or technology platform you use to store your passkeys. If you lose access to the platform, you also permanently lose access to every single account you’ve ever signed up for. For me, that would include losing access to my retirement savings, tax records, and the ability to communicate with many of my friends, to give a few examples.

My computer also doesn’t have Bluetooth, which means I can no longer log in to any websites on it even when I do have access to my passkeys.

I’ll answer your question with a question: can you memorise your passkey for a given website?

Not to mention the money spent on funding food stamps for Amazon’s underpaid employees.

Why inflict that thinking on environments that aren’t consulting projects if you don’t have to? That kind of thinking is a big contributor to the lack of trust in consultants to do good work that is in the client’s best interests rather than the consultants’. We don’t need employers to start seeing actual employees in the same way too.

> What key concept underlying how we now think about QM doesn’t go back to this list of people?

Off the top of my head: quarks, and therefore the existence of the colour charge quantum number; and the Higgs field.

All of the people in the list were also building on prior research by the likes of James Clerk Maxwell and Ludwig Boltzmann. Einstein himself said “I stand on the shoulders of Maxwell.”

There are other obvious candidates for inclusion like Henri Poincaré, Hendrik Lorentz, Satyendra Nath Bose…