LLM Inevitabilism

Compare these positive introductory experiences with two technologies that were pushed extremely hard by commercial interests in the past decade: crypto/web3 and VR/metaverse.

Neither was ever able to offer this kind of instant usefulness. With crypto, it’s still the case that you create a wallet and then… there’s nothing to do on the platform. You’re expected to send real money to someone so they’ll give you some of the funny money that lets you play the game. (At this point, a lot of people reasonably start thinking of pyramid schemes and multi-level marketing which have the same kind of joining experience.)

With the “metaverse”, you clear out a space around you, strap a heavy thing on your head, and shut yourself into an artificial environment. After the first oohs and aahs, you enter a VR chat room… And realize the thing on your head adds absolutely nothing to the interaction.

Give it some time - just like LLMs the first VR headsets were created in the 90s (for example by Nintendo). But it took another 30 years for the hardware to achieve levels of functionality and comfortableness that make it a viable consumer product. Apple Vision is starting to get there. And crypto is even younger - it started in early 2009. For people living in countries without a proper banking infrastructure the stablecoins are already very helpful. Billions of people live in countries that don't have a well audited financial sector, that respects the rule of law or an independent central bank that makes sound monetary decisions irrespective of the government. For them stablecoins and their cheap transactions are huge.

Every single HN post on AI or crypto I see this argument and it’s exhausting.

When Eliza was first built it was seen a toy. It took many more decades for LLMs to appear.

My favourite example is prime numbers: a bunch of ancient nerds messing around with numbers that today, thousands of years later, allow us to securely buy anything and everything without leaving our homes or opening our mouths.

You can’t dismiss a technology or discovery just because it’s not useful on an arbitrary timescale. You can dismiss it for other reasons, just not this reason.

Blockchain and related technologies have advanced the state of the art in various areas of computer science and mathematics research (zero knowledge proofs, consensus, smart contracts, etc.). To allege this work will bear no fruit is quite a claim.

> With the “metaverse”, you clear out a space around you, strap a heavy thing on your head, and shut yourself into an artificial environment. After the first oohs and aahs, you enter a VR chat room… And realize the thing on your head adds absolutely nothing to the interaction.

It doesn't if you use it as just a chat room. For some people it does add a lot, though.

The "metaverse" as in Active Worlds, Second Life, VR Chat, our own Overte, etc has been around for a long time and does have an user base that likes using it.

What I'm not too sure about is it having mass appeal, at least just yet. To me it's a bit of a specialized area, like chess. It's of great interest to some and very little to most of the population. That doesn't mean there's anything wrong with places like chess.com existing.

Bitcoin seems to be working as a kind of digital gold if you look at price development. It's not that much about technology though.

> With crypto, it’s still the case that you create a wallet and then… there’s nothing to do on the platform. You’re expected to send real money to someone so they’ll give you some of the funny money that lets you play the game.

This became a problem later due to governments cracking down on cryptos and some terrible technical choices made transactions expensive just as adoption was ramping. (Pat yourselves on the back, small blockers.)

My first experience with crypto was buying $5 in bitcoin from a friend. If I didn't do it that way I could go on a number of websites and buy crypto without opening an account, via credit card, or via SMS. Today, most of the $5 would be eaten by fees, and buying for cash from an institution requires slow and intrusive KYC.

> And realize the thing on your head adds absolutely nothing to the interaction.

There are some nice effects - simulating sword fighting, shooting, etc.

It's just benefits still outweigh the cost. Getting to "good enough" for most people is just not possible in short and midterm.

> buying for cash from an institution requires slow and intrusive KYC.

Hello my friend, grab a seat so we can contemplate the wickedness of man. KYC is not some authoritarian or entrenched industry response to fintech upstarts, it's a necessary thing that protects billions of people from crime and corruption.

The question I have for your observation (which I think is correct btw) is:

Do you think it's inherent to the technology that the use cases are not useful or is it our lack of imagination so far that we haven't come up with something useful yet?

The day I can put on a pair of AR glasses as lightweight as my current glasses and gain better vision, I'd pay a huge amount for that.

I hate my varifocals because of how constrained they make my vision feel...

And my vision is good enough that the only thing I struggle with without glasses is reading.

To me, that'd be a no-brainer killer app where all of the extra AR possibilities would be just icing.

Once you get something like enough and high resolution enough, you open up entirely different types of applications like that which will widen the appeal massively, and I think that is what will then sell other AR/VR capability. I'm not interested enough to buy AR glasses for the sake of AR alone, but if I could ditch my regular glasses (without looking like an idiot), then I'm pretty sure I'd gradually explore what other possibilities it'd add.

Solutions in search for a problem just don’t tend to be very good solutions after all.

Maybe the answer isn’t that we’re too dumb/shallow/unimaginative to put it to use, but that the metaverse and web3 are just things that turned out to not work in the end?

I just want the ability to put on a lightweight pair of glasses and have it remind me who people are.

Ideally by consulting a local database, made up of people I already know / have been introduced.

And yet while this capability would be life-changing, and has been technically possible for a decade or more, yet it was one of the first things banned/removed from APIs.

I understand privacy concerns of facial recognition looking up people against a global database, but I'm not asking for that. I'd be happy to have the burden of adding names/tags myself to the hashes.

I'd just like to be able to have what other people take for granted, the ability to know if you've met someone before (sometimes including people you've known for years).

I don’t have a problem with chess.com existing, but if someone starts shouting loudly about how chess.com is going to be the future of everything, and that I’ll need to buy a bunch of expensive-but-still-kinda-crappy hardware to participate in the inevitable chess.com-based society, and that we need to ground-up rearchitect computing to treat chess as fundamental component of UI… well, it just gets a little tiresome.

Research is fine. But when corporations and venture capitalists are asking for your money today in exchange for vague promises of eventual breakthroughs, it's not wrong to question their motives.

So was the telescreen.

AI has the same vague hand-wavey problems of the metaverse. LLMs are not AI. Roblox is not the metaverse. Both are approaching parts of the promise of each of their potential, but only a small part of what they could be or are promised to be.

Hype cycles will hype. Builders will build.

Except that you can't clone gold or fork gold.

Gold isn't lost because you forgot the password to open it.

Or arbitrarily decide tomorrow that the old gold is not valid and a specific chain of gold is the real one.

Also, you can smelt gold, create electronics, jewellery, cosmetics, drugs with gold, you can't with Bitcoin.

Seriously comparing Bitcoin to gold is beyond silly.

There are Ethereum, Algorand, many alternatives with low fees.

I'm not invested in Bitcoin in any way, so I consider myself neutral. But I see some similarities here. Yes, unlike Bitcoin, gold has some real uses - but it's not what defines gold price, so I don't think it's relevant here. Yes, humanity could theoretically collectively decide that from now on gold is not valuable any more and chose some other scarce material as a way to store wealth, so I think it's not unlike bitcoin really. The difference is in mindshare - gold obviously has more of that, including governments investing in gold. Bitcoin is more likely to drop out of fashion with investors than gold, but so far it didn't happen, and there is also a chance it will not happen or that it will get even more mindshare, e.g. with other states following Trump's US in creating bitcoin reserves.

That's just silly, everyone knows the future is lichess.

> When Eliza was first built it was seen a toy.

It was a toy, and that approach - hardcoded attempts at holding a natural language conversation - never went anywhere, for reasons that have been obvious since Eliza was first created. Essentially, the approach doesn't scale to anything actually useful.

Winograd'd SHRDLU was a great example of the limitations - providing a promising-seeming natural language interface to a simple abstract world - but it notoriously ended up being pretty much above the peak of manageable complexity for the hardcoded approach to natural language.

LLMs didn't grow out of work on programs like Eliza or SHRDLU. If people had been prescient enough to never bother with hardcoded NLP, it wouldn't have affected development of LLMs at all.

That's an unreasonably charitable reading of the purpose of KYC. It's primarily about government control of the primary medium of economic exchange. As always, this benefits the privileged at the expense of the less privileged.

Its use to limit competition from cryptocurrency is a perfect example of that. A major market which crypto was supposed to be able to serve - the "unbanked" - are largely locked out of it. Turns out giving poor people access to money is not a feature that the system wants to allow.

The benefit you claim for KYC is a marketing bullet point side effect at best.

Based on what do we know that Eliza won't scale? Have we tried building an Eliza with a few gigabytes of question/response patterns?

Prior to the rise of LLMs, such a thing would be a waste of time by any respectable AI researcher because it obviously isn't related to intelligence.

The problem with this kind of argument is what I'd call the "Bozo the Clown" rejoinder:

It's true that people spent a lot of time investigating something that decades (centuries, millennia) later came to be seen as useful. But it's also true that people spent a lot of time investigating things that didn't.

From the perspective of the present when people are doing the investigating, a strange discovery that has no use can't easily be told apart from a strange discovery that has a use. All we can do in that present is judge the technology on its current merits - or try to advance the frontier. And the burden of proof is on those who try to advance it to show that it would be useful, because the default position (which holds for most discoveries) is that they're not going to have the kind of outsize impact centuries hence that number theory did.

Or in other words: It's a bad idea to assume that everybody who get laughed at is a Galileo or Columbus, when they're more likely to be a Bozo the Clown.

I think AI is inevitable in the way that bitcoin is now inevitable: it's not going to go away, it consumes a huge amount of energy, has various negative externalities, but a massive fanbase.

It doesn't really matter whether crypto is "useful", it has billions of dollars worth of fans. Similarly the LLM fans are not going to go away. However, there will probably be curated little oases for human-made works. We're also going to see a technique adapted from self-crashing cars: the liability human. A giant codebase is launched and a single human "takes responsibility" (whatever that ends up meaning) for the failures.

I feel like my personal experience of the metaverse is a really good comparator for LLM’s. Really cool, I can see the possibilities, I want it! It seems like it’s there, But I can also see that the gap between what exists and what would make it truly useful is too great.

10 years ago I'd have settled for this if it only worked on Game Of Thrones characters.

This is why you can never trust any proprietary tech by a tech giant.

It's unfortunately a relatively hard optics thing to make reasonably working projectors into glasses, or the tiny OLED ones.

It does matter whether something is useful and I really wish people would stop making comparisons to crypto because it's an absolutely terrible comparison.

AI is certainly in a bubble right now, as with dotcoms in 1999. But AI is also delivering a lot of value right now and advancing at an incredible pace. It will become ubiquitous and at a faster pace than the Internet ultimately did.

Meanwhile, Bitcoin has been around for 17 years and there still are no non-criminal use cases apart from buying it and hoping it will be worth more in the future.

It doesn't really matter what use cases cryptocurrencies were supposed to have — their actual use cases turned out to be scams and speculation. We can wax philosophic about the failed promise, but to a rounding error scams and speculation have always been their only use cases.

Which makes it very understandable that crypto companies became subject to KYC laws as they tried to scale up to serve the American public! Online gambling and securities trading are already subject to KYC. The rest of the activity is the scams and crime that (despite your cynical reading) KYC was intended to fight in the first place.

If I understand the discussion correctly:

Your opinion is that the benefits of KYC (safety) outweigh the downsides of KYC (giving up liberty).

The other poster's opinion is that the downsides outweigh the benefits.

There is a quote out there regarding those who would sacrifice liberty to obtain safety, but it slips my mind at the moment.

Probably the biggest Eliza-like program is ALICE[1], which used a more formalized rule representation called AIML. The size of ALICE distributions is in the single-digit megabytes.

Systems like that don't scale in a human effort sense - i.e. the amount of effort required compared to the value produced is not worthwhile.

Aside from that, models like that didn't have a true grammar model. They responded to keywords, which meant that their responses were often not relevant to the input.

> "Prior to the rise of LLMs, such a thing would be a waste of time by any respectable AI researcher because it obviously isn't related to intelligence."

You might imagine so, but that wasn't really the case. ALICE won the Loebner AI prize multiple times, for example. Before neural networks started "taking over", it wasn't obvious to everyone what direction AI progress might come from.

People even tried to extend ELIZA/ALICE style models, with one of the most prominent examples being MegaHAL[2], which also won a Loebner prize. MegaHAL used a Markov model, so wasn't purely based on hardcoded rules, but like ELIZA and ALICE it still didn't understand grammar.

[1] https://en.wikipedia.org/wiki/Artificial_Linguistic_Internet...

[2] https://en.wikipedia.org/wiki/MegaHAL

Careening at 90 miles per hour through a school zone crosswalk as kids dive out of the way: "Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety."

> As always, this benefits the privileged at the expense of the less privileged.

This is all quite a naive look at the world. The least privileged don't have any money, so by definition aren't hurt by KYC.

Capital is power and power makes the world go round. If the powerful of the world desire one thing above all else, it's not to have any scrutiny over how they acquire more power and make use of it, with financial privacy being a very large part of that.

Financial privacy is without doubt important for the regular citizens, and we should have laws in place that protect it. There is no reason for the government to have access to your transactions outside a well-functioning system of checks and balances, court orders and warrants etc.

But financial privacy maximalists strike me as useful idiots for unrestrained power. There is nothing good that society has to gain from allowing anonymous transfers of billions of dollars across borders. Once you tolerate anonymous finance, an entire bestiary of crimes and abuses become possible or easier, without any benefit for the common man. This was widely the case in the second half of the 20th century, and the financial industry had no problem laundering the profits made from the misery and death of the wretched of the earth, as long as they got their cut.

KYC is foremost a tool for democracy and checks on power. It's not the only tool and it can't operate by itself, but you need it in place before you can even ask the question "what are the reasonable and socially useful limits of financial privacy?"

A Medici could hold a store of gold throughout their properties for the last 700 years and during that entire time, they can be a bank just on the fact that they can guarantee deposits against that wealth. At minimum they can ensure their wealth and that of their descendants, because gold has always been valuable. Even the communists hoarded gold - so much for building a currencyless society!

It is thus absurd to compare bitcoin to gold, yet. In 2000 years, if it's still around, I'm happy to accept the comparison.

I can find you people today that would take a gold coin for their services instead of a bitcoin (obviously worth far more), because they don't care / understand / trust it. The only reason I can think gold would no longer be valuable would also nullify the value of bitcoin - the demise of capitalism or anything like it, and the advent of a currencyless society.

> The least privileged don't have any money, so by definition aren't hurt by KYC.

A very privileged perspective.

What about, for example, undocumented immigrants? Countries allow them in to exploit for cheap labor, so they do have some money. But using banks is difficult and risky because of their documentation situation.

Now, if you're a certain kind of conservative, you'll say that's the way it's supposed to be, because these people having committed a civil violation are "criminals" which the KYC laws are attempting to punish.

But that's not how a compassionate human thinks.

Undocumented immigrants can absolutely open bank accounts: https://www.nerdwallet.com/article/banking/undocumented-immi...

Yawn. If that were truly analogous to the current topic, rather than a gross exaggeration, the analogy would be unnecessary.

Replace "Careening at 90 miles per hour through a school zone crosswalk as kids dive out of the way" with the actual topic of "Spending your money on legal things without government tracking and control".

Your point is understood that you personally prefer one thing to another, compared to another person preferring the opposite.

And if you were truly able to attack my position on its merits you wouldn't need to keep stripping out all specifics to build a straw man, but here we are!

(Also, the analogy would only be necessary if it were… not analogous to the topic at hand? That makes no sense.)

> undocumented immigrants? Countries allow them in to exploit for cheap labor

It seems you've answered your own question.

Your argument seems constructed as a red hering, "what about <this group of people left in legal limbo by the politicians>, wouldn't they benefit from <the speculative financial scheme I have a stake in, that the state can't track or ban>?". The solution to that problem is, you know, to not have people put into a situation where they are exploited and do not have legal rights; and in general, the solution to human rights failures in democracies surely isn't the suspension of laws, it's better laws and better systems put in place to protect those legal rights.

In the great scheme of things, undocumented migrants in 1st world economies are far from the wretched of the earth I was referring to. For example, it has recently emerged that a fellow country man of mine was involved in supplying european mercenaries into the DR Congo, which were caught in the rebel capture of Goma early 2025. The ring leader was found to have laundered a few millions of euros through cryptocurrency he earned from Congo.

DRC is one of the poorest countries on earth, with the majority of its population without electricity, internet, smartphones etc. They are the real "unbanked" and what bitcoin has done for them is to enable war. Now, it's very likely that the war would have been exactly the same if the mercs were paid using the traditional financial system, but at least that would leave traces and discouraged some of the mercs, since it's an illegal profession in many european countries of origin.

I don't have a solution for all the world's troubles, but this dystopia where you can anonymously buy armed soldiers on the internet to prop up your authoritarian regime or guerilla faction, surely isn't the best of all possible worlds.

> Meanwhile, Bitcoin has been around for 17 years and there still are no non-criminal use cases apart from buying it and hoping it will be worth more in the future.

That is plain and simply false. It works just fine as a currency, and some legitimate businesses even accept it. I think it's true that Bitcoin is not particularly useful, but that's not the same as there being no non-criminal use cases.

> It doesn't really matter what use cases cryptocurrencies were supposed to have — their actual use cases turned out to be scams and speculation.

I'm going to translate what you said here out of your obscene level of privilege:

"It doesn't really matter what use cases cryptocurrencies were supposed to have - even if their actual use cases did address those concerns, not all of them did, and what's more important to me, and other hypercapitalists like myself, is to maintain my privilege."

I very much did not say that cryptocurrencies' actual use cases address those concerns. In fact, I said the opposite!

Like the other commenter, the reason you need to "translate" my argument is that you can't rebut what I'm actually saying on its merits.

> Yes, humanity could theoretically collectively decide that from now on gold is not valuable any more and chose some other scarce material as a way to store wealth

No it's not going to work because gold is very scarce and extremely valuable as a material in any case.

Yes, the price would take a hit, likely something between silver and current gold price, which would increase use and demand and then price again.

There's a reason we're thinking of mining it from asteroids.

Why would I attack your position? Your position is totally reasonable given a particular set of pretty common preferences ("this" over "that"). Besides, you never asked me to try to change your mind, so why would I assume you wanted me to spend my time & effort doing that work?

What's more, your "position" (opinion) is no more or less defensible than the opposite opinion held by a different person, or any opinion in general, and other people are no more obliged to "attack" yours than you are theirs (is "attacking" really a good conversation anyways?)

I'm just a third party elucidating that this is 2 people disagreeing based on their own personal preferences, which are based on their own personal opinions, and doesn't need to be a referendum on either person.

You know, I really don't appreciate this coy "I'm not attacking your position, I'm just simplifying it to a tradeoff between liberty and safety and what's that quote about liberty vs. safety again? Ah well, the reader can draw their own conclusions." Your whole "all opinions are valid, who can truly say which is more defensible?" schtick is the last refuge of people who are unable to defend their views.

It's pretty obvious what you're trying to say here, man. Don't piss on my leg and tell me that it's raining.

It doesn't though. Go to your local Wal Mart, Costco, Home Depot, Target, etc. and try to make a purchase with it. Go to a local supermarket or restaurant and try to buy food with it. Stop at a gas station and try to buy fuel with it. Go to Amazon or eBay and try to make a purchase with it. Go to an airline or hotel website and try to book travel with it. Go into your Uber app and try to pay with it. Call your landlord and try to pay your rent with it. Objectively, it doesn't "work just fine as a currency" or we'd have empirical evidence of it actually being used as such in real life.

I'm sorry that you don't appreciate my "position" of "not arguing with your opinion, because you never asked anyone to help you change your opinion, and because you don't seem interested in changing it anyways if someone tried".

It sounds like you just want to fight rather than understand why different people have different views. Even after someone tells you that your opinions are common and your conclusions are reasonable given those opinions, you're still trying to find something to fight them on.

I'm more interested in the latter (understanding) than the former (fighting), so I'll pass, but I'm sure you'll have no trouble finding someone else on the internet who wants to fight.