So basically more ways of trying to make people buy things, do things, think things than before? I feel like our whole world more and more circulates around manipulation and the absence of truth and discourse.
Then again, I do think LLMs are an incredible technological achievement. The issue is not so much what they do or that they exist, but how they are utilized. Right now, they are utilized to further the class divide between rich and poor.
Who are we to trust in the future? Not big companies, not the state, not LLMs. Time to organize around groups and collectives that we know we can trust and that we know have our wellbeing in mind.
> The issue is not so much what they do or that they exist, but how they are utilized
This is exactly how we got here though. Technology is not passive. It changes incentives, procedures, ideas and shapes the world. If we don't structurally limit what and how it's used, then we are not in control, no matter what are choices personally are.
A major problem is that if we structurally limit what technologies do, we are still not in control. Now whoever we empowered to control and limit the technology is in control. Who keeps them accountable?
You’ll probably get one of three outcomes: regulatory capture by monopolies, self dealing by bureaucrats to enrich themselves or gain power, or regulatory capture by self absorbed ideologues who halt all progress or force it down some ideologically approved path.
In none of those scenarios is anything aligned with the best interest of the people.
The disincentives to nuclear war are glaringly obvious enough that even politicians (and their masters) get it.
AI isn't like that. One problem is that it's rather generally misunderstood at this point. "AI" is not "intelligence". It's intelligence-adjacent, and something like LLMs is part of our psyche...the subconscious facility that allows us to form sentences without really thinking about it.
At any rate, I have to agree with most of the points the blog author brings up.
The majority of human history has been written by the ruling class of the day. Transparency only seems to follow in the wake of their inevitable fall, usually at great cost in retrospective research via the oft thankless unraveling of threads of truth from their more copious fictions. Much like the machines we construct in our likeness, we too seem to get stuck in endless regressive cycles.
Folks in the "now" have always had a tendency to cling to their fictions as if they were truth for whatever reason; like nationalist exceptionalism, racial superiority, or religions rooted in "othering", etc. Humans seem to have an innate desire to fool themselves and trust in things they should not. Perhaps it's simply a sort of existential coping mechanism of living in a cold, unforgiving reality. We seek the comfort of lies.
Organizing around groups of trust, tends to lead to factionalism and conflicts. Knowing and trusting are sadly very different things in our species.
> Time to organize around groups and collectives that we know we can trust
I’ve had the same thoughts, but if you look deeper, it all circles back to what we already had: (open, transparent) public institutions, society, and government by the people. The foundation wasn't the problem; the environment was.
Along the way, social media noise, engagement-optimisation and Kardashian-style "entertainment news" infecting real news made an attention economy where, no matter how scandalous you are, attention can be minted into dollars. That is what polluted our infosphere and lead to the lack of trust.
Now, nobody trusts these previously mentioned public entities any more - sometimes due to state-actor or ad-tech disinformation, and sometimes for good reason like when the poisoned public allowed these 80s-style telemarketer-style political weirdos and their cronies to take over public administration.
> Right now, they are utilized to further the class divide between rich and poor.
Ironically this was the main reason LLMs were introduced in the first place, not to benefit the poor, but to widen the gap between the rich and the poor.
Our society, pre internet, built systems to manage trust. The conditions that allowed those systems to exist (the speed of transmission of data, the ratio of content generation to verification, the ability to shape consensus), have changed.
You are ringing the clarion call for community and cooperation, and it will not work. Not because people don’t want community or the better things, but because incentives make the world go round.
The choice between making some money at the cost of polluting the information commons is no choice at all. That degradation of the commons means no one can escape. No community you form, no group you build, dodges the fallout when someone decides to set fire to shared infrastructure.
We are moving into the dark forest era of the information economy. As models improve, inference costs drop, and capacity increases, the primary organism creating content online will be the bot.
Instead of building communities of people, build collections based on rules of engagement. Participants - be it bots or humans - must follow proscribed rules of conflict and debate.
That way it doesn’t matter if you are talking to a machine or a person. All that matters is that the rules were followed.
Local models and powerful consumer HW and an informed populace that doesn't hate STEM, but that's not good for the shareholder value so you get expensive everything everywhere all at once instead. And if you dare question the mindset of hating on STEM whilst being addicted to its fruits, that just means you're another one of those maximally SV-aligned sociopaths so why bother? Evolve and let the chips fall where they may because I don't see any other options that play out in the idiocracy craving for strong confidently wrong leadership.
Self inflating nipple shaped balloons that generate their own lift without any helium would be an incredible achievement but that doesn't mean it's useful beyond being novel. Chatbots are ultimately just predictive text on steroids, and only complete fools would base their business, or entire economy around it.
I don't need to conduct 1000 transactions per day. I don't forsee a world in which it will be some sort of fatal inconvenience to need to approve all purchases. I certainly don't plan on ever just handing over my credit card to an LLM, due to its fundamental architectural issues with injection, and I still don't anticipate handing it over to any future AI architecture anytime soon because I struggle to imagine what benefits could possibly be worth the risk of taking down such a basic, cheap barrier.
Agreed. My only real complaint with this article is it frames needing to argue with a machine as though this is a new, freshly annoying thing. I already do this constantly.
Every time I call the Costco pharmacy, I just hit 0 immediately because: Phone. Trees. Suck. They have always sucked, it's just an awful, grindingly slow way to accomplish ANYTHING, and it's so, so much easier to, when I need help, get a person on the line who can figure out what's gone wrong and sort it.
The only people benefiting from cutting that down are the scum class (combo of shareholders and executives) and who's shocked, really. Everything is being ruined nearly at all times to benefit the scum class.
At least phone trees are deterministic and there's still (usually) an option to get to a person for matters that aren't covered by the multiple choice options. Talking to AI is a much worse experience and the hope of the industry is that there won't need to be a human as a fallback anymore because (they believe) the AI is intelligent enough to handle anything.
I'm surprised to find so many people who consider human-based customer support a good experience. I wasted an hour on the phone last month with a series of polite support agents who I'm sure were wonderful people in their personal lives. They kept saying they'd like to try one more thing, making me wait 5 minutes (just short enough that I can't get anything done in the interim!), and then asking for one more pointless permutation of the workflow that did not work because their website was not showing me a button the support scripts said should be there. Talking to an LLM would have let me realize a lot faster that we weren't getting anywhere.
My father just changed his car key battery with the help of ai and he likes that. He also consulted it about about car insurance regulations and he got more out of it than searching the web himself.
For most simple mainstream questions I just ask ai instead of googling shitty results.
Most of the time ai is good enough and often better than the status ante.
People do not care if it is a stupid token prediction machine as long as the job gets done.
> Perhaps we’ll see distributed boycotts where many people deploy personal models to force Burger King’s models to burn through tokens at a fantastic rate.
Given how many people hate AI in general, I'm surprised there hasn't been anything like this happening. They could even get around the irony of using "AI" themselves, I bet low-tech language models like Markov chains could provide sufficient time wasting potential (I'd love to see it done with an old fashioned AIML chatbot). Asymmetric chatbot warfare.
Most people who hate AI have been completely dis-enfranchised by the system. The media won't amplificate their voices, any viable political leader that is seem as threat will be completely and utterly destroyed by the parties and the PAC machine still on the primaries.
It is an incredibly vexing situation to see whatever you're an AI hater or enthusiast.
I, for one, welcome our trillion parameters multiple layers overlords.
If your complaints about AI are largely about the industrial energy use, the poor quality of service, and the displacement of human labor, wasting more CPU time doesn't seem like a viable or useful protest. The lesson Burger King would take away from your DDoS protest isn't that they should provide better customer service, but that they shouldn't provide any customer service. You'd be giving them free cover to blame consumers for making customer service too expensive.
The worst thing is that non-technical people, and actually a lot of technical people without experience in ML, will tend to overstimate the capabilities of those systems, neither the nuances of probabilitic thinking to properly integrate their outputs in a decision.
Remember that the polygraph still exists, now we will be dealing with a massive portion of the decision makers will treat as artificial inteligence not in the technical sense we use, but as real inteligence, maybe even super-inteligence.
The erosion and further diffusion of responsibility is the trend that worries me the most, since it’s already how all mid-size organisations, businesses and institutions alike, operate by design, and LLMs are likely to make that much worse.
have worked closely with customer support teams, can confirm that the goal of any technical improvements that go in front of CS agents is to reduce ticket volume, and thus costs. of course they measure retention and satisfaction but ticket volume is always the big one. chatbots were big for this long before LLMs existed.
a fun side effect is that CS is also an early warning system for companies, so when you make it harder to get through to a human, you start throwing out info on your users' pain points. of course this only matters if people have a choice about whether to use your product, so that's gotta be an upside for insurance companies, etc.
I had a fun experience with my ISP where their chat bot couldn't help me (of course it couldn't, I don't call for "did you try turning it off and on again" problems), so it escalated me to a human agent. Said human agent was very obviously copy-pasting LLM output. I could tell because (1) the responses were nearly identical to what Claude already told me when I asked it before calling and (2) everyone once in a while I would get an uncharastically brief reply, without capitalization or punctuation, in Indian English.
I haven't a had a good experience since AT&T bought my previous ISP and forced me to switch to a different subsidiary.
Can I not shop for other insurance companies? I specifically chose my provider because I know there's an office I can call to talk to my agent or his secretary. The moment I have to interact with a chatbot, I take my business to someone else.
THEREFORE A COMPUTER MUST NEVER MAKE A MANAGEMENT DECISION
—IBM internal training,
1979
It took me a while to realise that the premise is saying the same thing as the reason why we have so many "Computer says no" experiences today.
The conclusion only follows if you want someone to be accountable.
If you want to avoid being accountable, computers should make all management decisions.
This has nothing to do with AI other than it provides another mechanism to do that.
People saying "I'd love to help you but the computer won't let me do that" has been happening for years now.
Websites develop abusive patterns because A/B testing lets a process decide based on the goal you want, It doesn't measure the repercussions so you have made no decision to allow them.
Management read it as
A COMPUTER CAN NEVER BE HELD ACCOUNTABLE
THEREFORE THERE CAN BE NO LIABILITY IF COMPUTERS MAKE ALL MANAGEMENT DECISIONS
I sent the entire series by Aphyr [1] to some friends. Two of them, independently, responded with a variant of, "TLDR, can you give a summary?"
I chat with these friends a lot but I rarely send articles that I suggest they read and that I think are profound, so I expected them to read it. These are smart people that have a history of reading lots of books.
They are both huge AI proponents now and use AI for nearly everything now. Debates on various topics with them used to be rich; now, they're shallow and they just send me AI summaries of points they're clearly just predisposed to. Their attention spans are dwindling.
maybe it means they were never really as smart as you thought?
Not meant to be snarky. It's been two decades now since my first wide-eyed entry into the workforce, moving for new opportunities, meeting new people. it's been great. There's a lot of smart people out there. I also realize that many people I seen as smart had more access to more content then i did. i still appreciated their sharing , it was enlightening to me. But after 20 years, I think back and it's literally quoting things from smart youtube videos. and regurgitating the latest thought leaders.
We all do this, but like you, what's meaningful to me is the chewing, the dissection and synthesis. coming together to share different perspectives and so on. i've had those friends too! it's just not 1:1
You might be right but they used to read much more and our arguments used to be deeper. The changes I'm seeing in them are highly correlated to their increased use of AI.
Maybe it's something like that AI allows them to indulge in their shallowness/laziness by giving them the impression that they're not doing that.
Or maybe they just don't want to read a long form analysis on something?
I also enjoy the series. But sometimes my friends send me things and I'm like, "not gonna read all of that."
Just because you're friends don't want to invest the same amount of time that you want to invest in your own personal enrichment doesn't mean they're getting stupid.
> Cognitive activity scaled down in relation to external tool use. …
> Self-reported ownership of essays was the lowest in the LLM group and the highest in the Brain-only group. LLM users also struggled to accurately quote their own work. While LLMs offer immediate convenience, our findings highlight potential cognitive costs. Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels.
I read one of his last week? and didn't like it that much. I read this one despite it because its quite high on hn for whatever reason.
I don't think everything is lies and i don't like how he thinks a LLM is just some bullshit machine.
Its also waaaay to early to even understand were this is going. We as humans have never had that much compute and used it this particular way. It could literlay be the road to a utopia or dystopia. But its very crazy to experience it.
His article series feels so negative and dismissive, that i'm not taking anything from it.
There is so much more research, money and compute behind this AI topic right now, every week or two weeks something relevant better/new comes out of this. From 2d, 3d models, new LLM versions, smaller LLms, faster inferencing (Nvidias Nemotron), we don't know how this will continue.
And the weird thing is that he clearly knows plenty about LLMs but it feels so negative dismissive, hard to put a finger to it.
I wouldn’t necessarily read a lengthy blog post either just because some friend recommended it to me, and conversely I wouldn’t expect a friend to necessarily read it if I was recommending it without being prompted for recommendations. There needs to be some additional incentive and/or interest.
Also, I’m reading this comment thread instead of TFA because I didn’t find the previous part I read that great. And I’m not an AI proponent, more of an AI skeptic.
I didn't provide much context but, 1) I've had deep conversations with these friends for years based on long articles or videos, and 2) I recommend maybe one or two long form items per year and they used to always review them without, "TLDR?"
So my main concern here is that my experience may be a microcosm of the shallowing of discussions correlated with some people's increased use of AI. That worries me.
It's more of a meta point to me. I get that this series isn't landing for some people, yourself included, but the meta-observation is that given something of roughly equal substantiveness as before, these friends' motivations for long form content and discussion seem to have atrophied, perhaps largely due to the addition of the AI summary reality cipher to their lives.
Of course, correlation isn't causation. Maybe they both just got older and more lazy, but given their reliance on AI summaries in other debates happening recently, I'm worried.
Is this something you're seeing personally? If so, how do you know it's because of the Online Safety Act? This is a personal blog and it doesn't seem to have any adult content that I can find. The homepage of the site isn't blocked when I check it here: https://www.blocked.org.uk/check
Regarding companies trying to block any contact with customer service and adding endless AI hurdles: In some countries, having a reachable means of contact is legally required. Is there a NOYB-style organization that specializes in enforcing this right (suing companies on behalf of consumers)?
For the "bureaucracy has royally fucked up and doesn't want to fix it", if it is something that can be fixed with money and isn't time sensitive (e.g. you need a refund rather than get the airline to actually provide you the ticket you already paid for and want to fly this weekend): In countries that have effective small claims courts, these can be a surprisingly convenient (less hassle than the "talk to the bot" wall of the company!) to resolve this kind of issue.
I hope that these resolution methods become more common - I think the tools to fight enshittification often already exist, we just don't use them enough. A welcome side effect would, of course, be that this would impose a real cost on the enshittifiers, creating an incentive to provide proper support.
> In countries that have effective small claims courts, these can be a surprisingly convenient (less hassle than the "talk to the bot" wall of the company!) to resolve this kind of issue.
Idk where we fall on the scale of “effectiveness” vs our peers, but I do read more people’s stories of Small Claims that are positive than negative. But I’ve never used this. I suspect it would be difficult to press a claim against a random large “company” just based on how slippery their identities even are. “Oh, Apple Inc. isn’t responsible for that, it’s a different subsidiary based in Ireland for tax reasons. Go serve them.” I think most people would have to be out more money (maybe more than the S.C. limit?) before being motivated to engage with the chronically overextended legal system, sadly.
Also, if the effective tools do exist, count on American companies using the American bribery-based political system to change the laws to dull those tools or to eliminate them.
Again, if you live in some consumer-friendly country good for you, I’m just saying how it plays out in this one society. I’ll stipulate that it’s all our fault blah blah blah.
I've been enjoying these articles by 'aphyr and I think they raise important points. Primarily though, they read to me as polemics of a curiously American nature.
The pattern goes something like this:
- this development is bad
- companies will be unrestrained in their use of this development
- there will be no rules so they can do whatever they want
- we are all fucked as a result
But then...propose that we make some laws to put rules around this stuff, also known as regulations and everybody goes "whoa hold up hold up hold up...I dunno about that part."
Dear friends - America has always been this way. Study your 19th and 20th century history. Companies will exploit the shit out of us unless we put some rules in place to prevent it. Yes, that might mean making less money in the short term as regulations cause friction. But in the long term it means we can have a better and actually livable society.
(For what it's worth I'm an American and not an uppity European or Australian taking potshots from across the pond; no offense to Euros or Aussies intended, love you guys)
I mean, whenever I read these types of articles and responses, my reaction is almost always the same: what is this ideal world people are hoping for?
The future of everything is lies... sure, well so was the past, what's your point?
When we say "snake oil salesmen" we literally had people selling these mineral oils, fraudulently, for a century. People yearn for a time when there wasn't an antagonistic relationship between buyer and seller, but it has never existed. There is only one way have that relationship, and it's by having a personal -- repeatable -- relationship with your seller, and that's usually expensive, even if that expense is not monetary. It just the game theory of repeated games.
The easiest way to establish this behavior is to be a regular at a bar while being a good patron that creates a positive experience for everyone else. There are multiple places I frequent where I have no doubt in my mind that if there was an emergency, they might ask me to help out in some way because I'm trustworthy. And they could hand me an envelope of money, knowing I'd not steal it, because it's not worth it for me to take a bit of money and ruin my welcome at this place.
That's not going to happen with online sellers, and it's not going to happen with most corporations. The promise was always "save money through economies of scale, such that you still win even though there is no relationship." For the most part, that's proven true. But as everything has been outsourced to China, and anyone can make a basic version of everything, we're running up against the limits of those benefits. And I truly believe that brands are going to start mattering a lot more going forward for marketplaces.
Shop at Costco because they care about their customers. In other areas where companies don't, you're going to have to do the legwork to find out who is going to treat you right. That sucks, and it'll probably be more expensive, but that's the antagonistic relationship that buyers and sellers have naturally. Regulation helps, yes, but it's not some cure all. The expectation that you can just go with the person with the lowest price and get quality services is not a thing that's every existed.
It’s not that people are against regulations totally, but that the structure of society is broken. People with wealth, including corporations, can influence and control everything with money. Legislators are easily bribed. Lawsuits are expensive and take years. It’s hard to make anything happen unless you’re already rich or connected enough to access the right people.
The real issue is new amendments are needed. But that’s hard. You need 75% of states ratifying. And that seems impossible today on any topic.
I’d andd another layer - for American tech workers, regulation also reduces profits. This hurts salaries, stocks options and career growth.
Incentives make the world go round, so even if people recognize the issue, they would rather it become someone else’s problem, than willingly harm their own future.
But then...propose that we make some laws to put rules around this stuff, also known as regulations and everybody goes "whoa hold up hold up hold up...I dunno about that part."
Of course, what those who call for regulation of AI or other nascent technologies are really saying is, "Unqualified and/or biased and/or corrupt and/or dull-witted people should make decisions that affect us all, based on incomplete, misleading or rapidly-evolving information, with the power to enforce them at gunpoint."
And then they wonder why other countries beat their own.
I wonder if there's also less of a stigma and sense of wrongdoing about tricking an LLM versus tricking an employee
We intuitively know that an employee will be punished and may get fired if we trick them. Many of us won't try to trick human employees as a result, because we would feel bad if they had bad consequences as a result of our trickery
There is likely no such hesitation around tricking LLMs. I know I personally wouldn't feel bad about it at all. Mostly because any computerized customer service process is annoying so anything I can do to limit my time dealing with it is a win in my books
„Agentic commerce means handing your credit card to a Large Language Model” - this is simply not true. LLMs/Agents will never get any credit/debit card details, they will be just an interface.
> People are very excited about “agentic commerce”. Agentic commerce means handing your credit card to a Large Language Model, giving it access to the Internet, telling it to buy something, and calling it in a loop until something exciting happens.
I think you're confusing this for the other side of things. The article talks about how some people already use OpenClaw and the variations, give them access to bunch of stuff including cards to purchase things (sometimes virtual and limited cards), I think that's what the article talks about when they say "agentic commerce".
Obviously a intentional simplification in the language the author uses, but I think it gets the point across at least.
Ok, in this case I misunderstood author's point here. "Agentic commerce" for me is a framework that networks (MC, Visa) and other big players in the ecosystem are working on. So it's far from "i'll give my credit cards details to openclaw and hope for the best". Obviously I know that a lot of people do that, unfortunately...
This is doomerism. Yes, everything will get worse. But everything will also get better. Such is progress. (for every one of these examples of annoyances, I can think of two ways to use AI to get around the annoyance. not clever programmer things, but things an average person who learns to use Codex or Claude Desktop to operate their desktop will know)
Most of these annoyances are also things that existed before AI, and will continue to exist after, because consumerist capitalism. The good little obedient consumers get abused because they don't stand up for themselves. Customer service is an enfuriating maze? Yeah, because you voted with your dollars (and political indifference) to allow companies to make customer service (the thing you pay for) worse. We bring these problems on ourselves. It's pointless to complain if you aren't willing to do anything to change it. (And if you think you can't change it, there's other nations to look at, as well as the fact that you live in a democracy - for now - unlike the rest of the world)
Hell, we already have companies whose sole purpose is to manage your subscriptions for you because you're too lazy to do it yourself. You could look at this and say, man, the world is terrible! Or you could look at this and say, man, how great is my life that I can not only subscribe to a lot of things without going bankrupt, but I have extra cash left over to pay a company to manage my subscriptions?
Don't let the hedonic treadmill and complacency trick you into A) accepting a worse life, or B) convincing yourself your life is bad when it's actually better than most people's.
>(for every one of these examples of annoyances, I can think of two ways to use AI to get around the annoyance. not clever programmer things, but things an average person who learns to use Codex or Claude Desktop to operate their desktop will know)
As the author said:
>I suspect that like the job market, everyone will wind up paying massive “AI” companies to manage the drudgery they created.
It could also lead to a massive crash of capitalism and reevaluation of how our society functions.
It could lead to significant progress in every single research area.
I'm at least very impressed about the amount of open models and that it doesn't hold up that the gap between public and private diverges massivly. Public is probably one year behind.
> Yes, everything will get worse. But everything will also get better.
That is not known. Things could easily just get worse, and IMO that is far more likely. Every civilisation has collapsed, ours is clearly in decline, and AI could likely accelerate that decline.
I'm not exactly seeing the progress here. AI helps us write some software a bit faster? Doesn't seem revolutionary to me. Is it having any significant impact on peoples lives other than the various economic forces? I'm not seeing it.
> Yeah, because you voted with your dollars
In our system? No. In crony capitalism the companies who win do it through manipulating the political system. And when the government inflates the currency and destroys people buying power they simply cannot afford to "vote with their dollars". This is hilariously naive.
> That is not known. Things could easily just get worse, and IMO that is far more likely. Every civilisation has collapsed, ours is clearly in decline, and AI could likely accelerate that decline.
This is the definition of doomerism. Of all the ways each society has adapted to industrializaion, of all the technologies we've adapted to, and our continued use of all of them despite any negative impacts, you think an LLM is going to end things? An overcomplicated autocomplete?
There is no evidence whatsoever that suggests our society would collapse under AI. An imagination is not evidence. I mean, it's just a tool! Like the car, the telephone, the genetically engineered seed. Humans always freak out when they see something they don't understand. But it turns out there's no tool we have created that we can't control. We're stupid, but we're not idiots.
I read the first couple of posts in the series. The essay is full of criticism of LLMs, and in a couple of places the author distances himself, as if he himself isn't using them ("some people I respect tell me that...").
It's certainly worth discussing the fact that the entire industry is starting to outsource large amounts of our thinking and writing work to non-sentient statistical algorithms, but this discussion needs to honestly confront the extent to which they are successfully completing useful tasks today.
Lots of blaming LLMs but I think the root cause lies elsewhere, I’m not even sure whether dismissing it as “capitalism” or “profit motives” would do it justice, because in general it feels more like the world that we live in lacks humanity.
Even in a capitalist world, a company could take a stance and decide not to purposefully screw people over, but in the world that we live in instead they look for ways to better screw over people and extract more money from them. It doesn’t matter whether your customer support is handled by someone from India, a crappy telephone tree or some voice model, when the incentive is the same - to do the bare minimum for customer “support” (in practice, just getting you to fuck off). Same for handling insurance claims and “dynamic pricing” of things - it doesn’t matter whether it’s some proprietary algorithm or just an LLM making crap up when the goal is to screw you over.
Blaming “AI” for all of this would be barking up the wrong tree (without that tech they’d just find other ways), though one can definitely acknowledge that this technology provides another convenient scapegoat, same as how you can lay employees off and just say cause it’s because of AI when in actuality it’s just greed and wanting to make your books look better.
In a capitalist world, the company that does decide to screw people over gets rich and the one that doesn't goes out of business.
It would be great if people chose not to do business with the former, but many simply do not care. They may think only other people get screwed. They may not take the time to think about it, especially if the company spends a ton of money obfuscating their misbehavior. Quite a few actively defend the right of companies to screw them.
Technology multiplies that like a lever. We weren't prepared for capitalism before LLMs and we're massively under-prepared now.
At various previous companies I've worked at product managers, executives, and engineers love bandying about the idea of "building for nontechnical users" as a way to make their widgets more "friendly". But it's just another way to otherize and denigrate "those people" who are the out group. They might, through a metacognitive defect or simple sociopathy, actually believe they're "doing good" by considering the poor creature's plight and making compassionate decisions on their behalf. But it's all crap. All they're actually doing is confirming their biases. LLMs are the divine nectar to these people, an enshittification accelerant par excellence.
Payment processing, is better than it was in 2000, but still not good.
Micropayments: this is obnoxiously expensive to do.
Discovery, and discoverability: again here we have better but not good solutions (and many of the ones that were once good are enshitified).
Pricing: this is a problem everywhere, and frankly we need the law to change in a way that is pro consumer. Publishing prices, disclosure of fees, in both services and for payment processing (that 3 percent back from visa looks a lot less attractive when it's part of a 5 percent mark up).
Customer service: well there are already companies promoting models where they cut you off and send you into a black hole (google is a prime example). Good customer service will become a differentiator, and maybe a "paid for" service as well.
> Good customer service will become a differentiator
This does not matter without antitrust, which is why customer service became bad in the first place. 30 years ago, the low quality of customer service we complain about now simply didn't exist, at any size or professional level of business, and never had.
If a company back then had the customer service of the average company now, or even the average government agency now, people would have suspected that it was a covert front for criminals or spies.
If a company doesn't have to compete, it can cut everything until it only has the ghost of a product and a billing department. You don't boycott monopolies, monopolies boycott you. If three companies put you on a list to not have internet, phone service, a bank account or a credit card, etc., you just can't have them. You've become a European human rights judge.
To lie requires recognition of the truth and an intention to deceive. LLM’s don’t have such abilities. They are systems that generate plausible sequences of symbols based on training inputs, alignments, reinforcement, and inference. These systems don’t know or care what truth is and therefore cannot lie.
It’s already bad. I’m not looking forward to the future. These systems are terrible. It’s a future without people that they want for some reason. I’d rather deal with people incompetent, tired, annoyed people than an LLM.
An important distinction to make, and I whole heartedly agree.
It’s not LLMs replacing workers, it’s people. People who have a lot of money and don’t sell their labour for a paycheque. And the systems that compel them to such actions.
LLM when it came out, was perfect as an interface between a system and a normal human.
So many people call customer support for issues they could in theory fix themselves. If that LLM system can understand me well enough, its an okay interface.
In worst case you have to escalate anyway. My mum actually told me that she talked to some AI.
And yes normal systems are also not correct often enough. With AI/LLM software will get cheaper which should incresase quality overall.
I dont think ai/llm in this case will change anything.
Relevant change will happen due to the fact that humans can be replaced by AI/LLMs. It was not even imaginable a few years back how a good ai system would even look like. Translaters lost their jobs, basic arists lost their jobs. Small contracts for basic things are gone. The restaurant poster no one cares? AI. The website translation for some small business? no one cares.
Do you want to add any argument so we can discuss this?
I mean, did you not write with ChatGPT and were surprised how well it response?
I'm schocked how well i can talk to an AI through some app like Gemini or ChatGTP. A few years ago i couldn't imagine building such a generic system which such high quality of understanding.
I was playing around with dragon naturally speaking and similiar dictation tools 10 years ago and it was horrible. And that software is expensive.
If you look how normal people use a computer, they are slow just because they don't understand basic drag and drop. Or they are unable to just create some java or php script to convert some data or clean up some data. I would just write a php script reading some csv file and converting stuff around and was faster than everyone around me.
Tool calling is bonkers.
And i tried to break GPT-3, i can literaly write an english sentence and just dropin german words, it was already that good.
Its often enough shitty in doing exactly what i want, but the quality is massive to everything we had before. Massive.
Not the OP, but you wrote “LLM when it came out, was perfect as an interface between a system and a normal human”. That’s a specific and very encompassing claim. I can only think of very simplistic systems (like a microwave oven maybe) where a current LLM could function perfectly as the sole command interface, much less when LLMs first became available. For systems of any significant complexity, it tends to turn into an exercise in frustration and failure modes when the LLM is your only interface (and frequently even when it isn’t).
An LLM can enhance the interface of a system and can be really useful in that despite its imperfections. But that’s a very different claim.
It was a significant jump from whatever we had before to a quality unseen before.
As i mentioned, i threw english and german at it.
How many people can change the time on their microwave?
How many people can ask an LLM through voice or text to change the time of the microwave?
A LLM is an interface to a service if you add a MCP Server. Now i can ask Jira things like "hey whats my current task? And what do i need to do?"
Its also an interface to documentation. I asked it to help me build up a hugo templating based website because just reading the hugo docs did not help me as much as the LLM did (and that was 2 years ago).
In best case, as long as an LLM is not AGI or ASI, we have good tools with validation behind the LLMs before the LLM becomes the system itself.
You're on a forum with a disproportionate number of people who are trying to profit from AI and have an interest in promoting that it's a worthwhile time and resource investment. It is a different universe than other places outside this bubble.
I added talking points like the one were i state that people call support just to fix issues they could fix themselves.
My point with my mum should imply that it was successful but for sure at least you are pointing something out and now we can talk about it: My mum talked to an AI and it helped her.
So basically more ways of trying to make people buy things, do things, think things than before? I feel like our whole world more and more circulates around manipulation and the absence of truth and discourse.
Then again, I do think LLMs are an incredible technological achievement. The issue is not so much what they do or that they exist, but how they are utilized. Right now, they are utilized to further the class divide between rich and poor.
Who are we to trust in the future? Not big companies, not the state, not LLMs. Time to organize around groups and collectives that we know we can trust and that we know have our wellbeing in mind.
> The issue is not so much what they do or that they exist, but how they are utilized
This is exactly how we got here though. Technology is not passive. It changes incentives, procedures, ideas and shapes the world. If we don't structurally limit what and how it's used, then we are not in control, no matter what are choices personally are.
A major problem is that if we structurally limit what technologies do, we are still not in control. Now whoever we empowered to control and limit the technology is in control. Who keeps them accountable?
You’ll probably get one of three outcomes: regulatory capture by monopolies, self dealing by bureaucrats to enrich themselves or gain power, or regulatory capture by self absorbed ideologues who halt all progress or force it down some ideologically approved path.
In none of those scenarios is anything aligned with the best interest of the people.
I don’t disagree. A consumer oriented democracy is not well equipped for the challenge.
I hate to tell you this but nobody has ever been in control. To think you can is to think you can unring a bell.
Right, and that's why we all died in a nuclear war.....
The disincentives to nuclear war are glaringly obvious enough that even politicians (and their masters) get it.
AI isn't like that. One problem is that it's rather generally misunderstood at this point. "AI" is not "intelligence". It's intelligence-adjacent, and something like LLMs is part of our psyche...the subconscious facility that allows us to form sentences without really thinking about it.
At any rate, I have to agree with most of the points the blog author brings up.
The majority of human history has been written by the ruling class of the day. Transparency only seems to follow in the wake of their inevitable fall, usually at great cost in retrospective research via the oft thankless unraveling of threads of truth from their more copious fictions. Much like the machines we construct in our likeness, we too seem to get stuck in endless regressive cycles.
Folks in the "now" have always had a tendency to cling to their fictions as if they were truth for whatever reason; like nationalist exceptionalism, racial superiority, or religions rooted in "othering", etc. Humans seem to have an innate desire to fool themselves and trust in things they should not. Perhaps it's simply a sort of existential coping mechanism of living in a cold, unforgiving reality. We seek the comfort of lies.
Organizing around groups of trust, tends to lead to factionalism and conflicts. Knowing and trusting are sadly very different things in our species.
> I feel like our whole world more and more circulates around manipulation
Hate to break it to you but it's always been this way, and it was easier in the past when information was so much more expensive to distribute.
> Time to organize around groups and collectives that we know we can trust
I’ve had the same thoughts, but if you look deeper, it all circles back to what we already had: (open, transparent) public institutions, society, and government by the people. The foundation wasn't the problem; the environment was.
Along the way, social media noise, engagement-optimisation and Kardashian-style "entertainment news" infecting real news made an attention economy where, no matter how scandalous you are, attention can be minted into dollars. That is what polluted our infosphere and lead to the lack of trust.
Now, nobody trusts these previously mentioned public entities any more - sometimes due to state-actor or ad-tech disinformation, and sometimes for good reason like when the poisoned public allowed these 80s-style telemarketer-style political weirdos and their cronies to take over public administration.
> Right now, they are utilized to further the class divide between rich and poor.
Ironically this was the main reason LLMs were introduced in the first place, not to benefit the poor, but to widen the gap between the rich and the poor.
Our society, pre internet, built systems to manage trust. The conditions that allowed those systems to exist (the speed of transmission of data, the ratio of content generation to verification, the ability to shape consensus), have changed.
You are ringing the clarion call for community and cooperation, and it will not work. Not because people don’t want community or the better things, but because incentives make the world go round.
The choice between making some money at the cost of polluting the information commons is no choice at all. That degradation of the commons means no one can escape. No community you form, no group you build, dodges the fallout when someone decides to set fire to shared infrastructure.
We are moving into the dark forest era of the information economy. As models improve, inference costs drop, and capacity increases, the primary organism creating content online will be the bot.
Instead of building communities of people, build collections based on rules of engagement. Participants - be it bots or humans - must follow proscribed rules of conflict and debate.
That way it doesn’t matter if you are talking to a machine or a person. All that matters is that the rules were followed.
Local models and powerful consumer HW and an informed populace that doesn't hate STEM, but that's not good for the shareholder value so you get expensive everything everywhere all at once instead. And if you dare question the mindset of hating on STEM whilst being addicted to its fruits, that just means you're another one of those maximally SV-aligned sociopaths so why bother? Evolve and let the chips fall where they may because I don't see any other options that play out in the idiocracy craving for strong confidently wrong leadership.
Self inflating nipple shaped balloons that generate their own lift without any helium would be an incredible achievement but that doesn't mean it's useful beyond being novel. Chatbots are ultimately just predictive text on steroids, and only complete fools would base their business, or entire economy around it.
I don't need to conduct 1000 transactions per day. I don't forsee a world in which it will be some sort of fatal inconvenience to need to approve all purchases. I certainly don't plan on ever just handing over my credit card to an LLM, due to its fundamental architectural issues with injection, and I still don't anticipate handing it over to any future AI architecture anytime soon because I struggle to imagine what benefits could possibly be worth the risk of taking down such a basic, cheap barrier.
All that stuff about support, though, inevitable.
Agreed. My only real complaint with this article is it frames needing to argue with a machine as though this is a new, freshly annoying thing. I already do this constantly.
Every time I call the Costco pharmacy, I just hit 0 immediately because: Phone. Trees. Suck. They have always sucked, it's just an awful, grindingly slow way to accomplish ANYTHING, and it's so, so much easier to, when I need help, get a person on the line who can figure out what's gone wrong and sort it.
The only people benefiting from cutting that down are the scum class (combo of shareholders and executives) and who's shocked, really. Everything is being ruined nearly at all times to benefit the scum class.
At least phone trees are deterministic and there's still (usually) an option to get to a person for matters that aren't covered by the multiple choice options. Talking to AI is a much worse experience and the hope of the industry is that there won't need to be a human as a fallback anymore because (they believe) the AI is intelligent enough to handle anything.
I'm surprised to find so many people who consider human-based customer support a good experience. I wasted an hour on the phone last month with a series of polite support agents who I'm sure were wonderful people in their personal lives. They kept saying they'd like to try one more thing, making me wait 5 minutes (just short enough that I can't get anything done in the interim!), and then asking for one more pointless permutation of the workflow that did not work because their website was not showing me a button the support scripts said should be there. Talking to an LLM would have let me realize a lot faster that we weren't getting anywhere.
you're part of the scum class btw (we all hold shares)
My father just changed his car key battery with the help of ai and he likes that. He also consulted it about about car insurance regulations and he got more out of it than searching the web himself.
For most simple mainstream questions I just ask ai instead of googling shitty results.
Most of the time ai is good enough and often better than the status ante.
People do not care if it is a stupid token prediction machine as long as the job gets done.
[delayed]
> Perhaps we’ll see distributed boycotts where many people deploy personal models to force Burger King’s models to burn through tokens at a fantastic rate.
Given how many people hate AI in general, I'm surprised there hasn't been anything like this happening. They could even get around the irony of using "AI" themselves, I bet low-tech language models like Markov chains could provide sufficient time wasting potential (I'd love to see it done with an old fashioned AIML chatbot). Asymmetric chatbot warfare.
Most people who hate AI have been completely dis-enfranchised by the system. The media won't amplificate their voices, any viable political leader that is seem as threat will be completely and utterly destroyed by the parties and the PAC machine still on the primaries.
It is an incredibly vexing situation to see whatever you're an AI hater or enthusiast.
I, for one, welcome our trillion parameters multiple layers overlords.
If your complaints about AI are largely about the industrial energy use, the poor quality of service, and the displacement of human labor, wasting more CPU time doesn't seem like a viable or useful protest. The lesson Burger King would take away from your DDoS protest isn't that they should provide better customer service, but that they shouldn't provide any customer service. You'd be giving them free cover to blame consumers for making customer service too expensive.
The worst thing is that non-technical people, and actually a lot of technical people without experience in ML, will tend to overstimate the capabilities of those systems, neither the nuances of probabilitic thinking to properly integrate their outputs in a decision.
Remember that the polygraph still exists, now we will be dealing with a massive portion of the decision makers will treat as artificial inteligence not in the technical sense we use, but as real inteligence, maybe even super-inteligence.
The erosion and further diffusion of responsibility is the trend that worries me the most, since it’s already how all mid-size organisations, businesses and institutions alike, operate by design, and LLMs are likely to make that much worse.
Just need your own LLMs to exhaust them. The future seems like it will be owned by whoever can automate with LLMs in whatever problem domain there is.
My agent will be in touch with yours, I guess.
have worked closely with customer support teams, can confirm that the goal of any technical improvements that go in front of CS agents is to reduce ticket volume, and thus costs. of course they measure retention and satisfaction but ticket volume is always the big one. chatbots were big for this long before LLMs existed.
a fun side effect is that CS is also an early warning system for companies, so when you make it harder to get through to a human, you start throwing out info on your users' pain points. of course this only matters if people have a choice about whether to use your product, so that's gotta be an upside for insurance companies, etc.
I had a fun experience with my ISP where their chat bot couldn't help me (of course it couldn't, I don't call for "did you try turning it off and on again" problems), so it escalated me to a human agent. Said human agent was very obviously copy-pasting LLM output. I could tell because (1) the responses were nearly identical to what Claude already told me when I asked it before calling and (2) everyone once in a while I would get an uncharastically brief reply, without capitalization or punctuation, in Indian English.
I haven't a had a good experience since AT&T bought my previous ISP and forced me to switch to a different subsidiary.
Can I not shop for other insurance companies? I specifically chose my provider because I know there's an office I can call to talk to my agent or his secretary. The moment I have to interact with a chatbot, I take my business to someone else.
A COMPUTER CAN NEVER BE HELD ACCOUNTABLE
THEREFORE A COMPUTER MUST NEVER MAKE A MANAGEMENT DECISION
—IBM internal training, 1979
It took me a while to realise that the premise is saying the same thing as the reason why we have so many "Computer says no" experiences today.
The conclusion only follows if you want someone to be accountable.
If you want to avoid being accountable, computers should make all management decisions. This has nothing to do with AI other than it provides another mechanism to do that.
People saying "I'd love to help you but the computer won't let me do that" has been happening for years now.
Websites develop abusive patterns because A/B testing lets a process decide based on the goal you want, It doesn't measure the repercussions so you have made no decision to allow them.
Management read it as
A COMPUTER CAN NEVER BE HELD ACCOUNTABLE
THEREFORE THERE CAN BE NO LIABILITY IF COMPUTERS MAKE ALL MANAGEMENT DECISIONS
You're misinterpreting the implication. A better phrasing might be:
A computer can never be held accountable. Therefore, since all management decisions must have accountability, a computer must never make them.
I sent the entire series by Aphyr [1] to some friends. Two of them, independently, responded with a variant of, "TLDR, can you give a summary?"
I chat with these friends a lot but I rarely send articles that I suggest they read and that I think are profound, so I expected them to read it. These are smart people that have a history of reading lots of books.
They are both huge AI proponents now and use AI for nearly everything now. Debates on various topics with them used to be rich; now, they're shallow and they just send me AI summaries of points they're clearly just predisposed to. Their attention spans are dwindling.
[1] https://aphyr.com/data/posts/411/the-future-of-everything-is...
maybe it means they were never really as smart as you thought?
Not meant to be snarky. It's been two decades now since my first wide-eyed entry into the workforce, moving for new opportunities, meeting new people. it's been great. There's a lot of smart people out there. I also realize that many people I seen as smart had more access to more content then i did. i still appreciated their sharing , it was enlightening to me. But after 20 years, I think back and it's literally quoting things from smart youtube videos. and regurgitating the latest thought leaders.
We all do this, but like you, what's meaningful to me is the chewing, the dissection and synthesis. coming together to share different perspectives and so on. i've had those friends too! it's just not 1:1
You might be right but they used to read much more and our arguments used to be deeper. The changes I'm seeing in them are highly correlated to their increased use of AI.
Maybe it's something like that AI allows them to indulge in their shallowness/laziness by giving them the impression that they're not doing that.
That's interesting; have you talked to your friends about their changes in behaviour? Is it something they've noticed themselves?
Friends don’t send friends AI summaries
Coworkers, too.
Or maybe they just don't want to read a long form analysis on something?
I also enjoy the series. But sometimes my friends send me things and I'm like, "not gonna read all of that."
Just because you're friends don't want to invest the same amount of time that you want to invest in your own personal enrichment doesn't mean they're getting stupid.
MIT actually has a paper on how ChatGPT use impacted cognitive skills for essay writing.
> Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task
> https://arxiv.org/abs/2506.08872
> Cognitive activity scaled down in relation to external tool use. …
> Self-reported ownership of essays was the lowest in the LLM group and the highest in the Brain-only group. LLM users also struggled to accurately quote their own work. While LLMs offer immediate convenience, our findings highlight potential cognitive costs. Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels.
You might just overhype this blog.
I read one of his last week? and didn't like it that much. I read this one despite it because its quite high on hn for whatever reason.
I don't think everything is lies and i don't like how he thinks a LLM is just some bullshit machine.
Its also waaaay to early to even understand were this is going. We as humans have never had that much compute and used it this particular way. It could literlay be the road to a utopia or dystopia. But its very crazy to experience it.
His article series feels so negative and dismissive, that i'm not taking anything from it.
There is so much more research, money and compute behind this AI topic right now, every week or two weeks something relevant better/new comes out of this. From 2d, 3d models, new LLM versions, smaller LLms, faster inferencing (Nvidias Nemotron), we don't know how this will continue.
And the weird thing is that he clearly knows plenty about LLMs but it feels so negative dismissive, hard to put a finger to it.
The author uses a lot of words and references to make critical conclusions that they do disclaim aren't expert.
Rather than dismissive, I see it as effort intensive. The conclusions can be negative, but they've spawned so much discussion which i think is great.
I wasn't even hyping it though. I shared it among friends to spark discussion. Sure, there's some hyperbole, but I found it thought provoking.
(FYI, I didn't downvote your comment)
I wouldn’t necessarily read a lengthy blog post either just because some friend recommended it to me, and conversely I wouldn’t expect a friend to necessarily read it if I was recommending it without being prompted for recommendations. There needs to be some additional incentive and/or interest.
Also, I’m reading this comment thread instead of TFA because I didn’t find the previous part I read that great. And I’m not an AI proponent, more of an AI skeptic.
I didn't provide much context but, 1) I've had deep conversations with these friends for years based on long articles or videos, and 2) I recommend maybe one or two long form items per year and they used to always review them without, "TLDR?"
So my main concern here is that my experience may be a microcosm of the shallowing of discussions correlated with some people's increased use of AI. That worries me.
It's more of a meta point to me. I get that this series isn't landing for some people, yourself included, but the meta-observation is that given something of roughly equal substantiveness as before, these friends' motivations for long form content and discussion seem to have atrophied, perhaps largely due to the addition of the AI summary reality cipher to their lives.
Of course, correlation isn't causation. Maybe they both just got older and more lazy, but given their reliance on AI summaries in other debates happening recently, I'm worried.
https://archive.is/c2sBh
... is there any reason why I shouldn't be visiting Aphyr's site directly?
Blocked in the UK. “Online safety act”
Is this something you're seeing personally? If so, how do you know it's because of the Online Safety Act? This is a personal blog and it doesn't seem to have any adult content that I can find. The homepage of the site isn't blocked when I check it here: https://www.blocked.org.uk/check
https://aphyr.com/posts/395-geoblocking-multiple-localities-...
Thanks for this! Seems like a bold stance… but the Online Safety Act also seems like a poor piece of legislation.
It tells you when you visit.
Regarding companies trying to block any contact with customer service and adding endless AI hurdles: In some countries, having a reachable means of contact is legally required. Is there a NOYB-style organization that specializes in enforcing this right (suing companies on behalf of consumers)?
For the "bureaucracy has royally fucked up and doesn't want to fix it", if it is something that can be fixed with money and isn't time sensitive (e.g. you need a refund rather than get the airline to actually provide you the ticket you already paid for and want to fly this weekend): In countries that have effective small claims courts, these can be a surprisingly convenient (less hassle than the "talk to the bot" wall of the company!) to resolve this kind of issue.
I hope that these resolution methods become more common - I think the tools to fight enshittification often already exist, we just don't use them enough. A welcome side effect would, of course, be that this would impose a real cost on the enshittifiers, creating an incentive to provide proper support.
> In countries that have effective small claims courts, these can be a surprisingly convenient (less hassle than the "talk to the bot" wall of the company!) to resolve this kind of issue. Idk where we fall on the scale of “effectiveness” vs our peers, but I do read more people’s stories of Small Claims that are positive than negative. But I’ve never used this. I suspect it would be difficult to press a claim against a random large “company” just based on how slippery their identities even are. “Oh, Apple Inc. isn’t responsible for that, it’s a different subsidiary based in Ireland for tax reasons. Go serve them.” I think most people would have to be out more money (maybe more than the S.C. limit?) before being motivated to engage with the chronically overextended legal system, sadly.
Also, if the effective tools do exist, count on American companies using the American bribery-based political system to change the laws to dull those tools or to eliminate them.
Again, if you live in some consumer-friendly country good for you, I’m just saying how it plays out in this one society. I’ll stipulate that it’s all our fault blah blah blah.
I've been enjoying these articles by 'aphyr and I think they raise important points. Primarily though, they read to me as polemics of a curiously American nature.
The pattern goes something like this:
- this development is bad
- companies will be unrestrained in their use of this development
- there will be no rules so they can do whatever they want
- we are all fucked as a result
But then...propose that we make some laws to put rules around this stuff, also known as regulations and everybody goes "whoa hold up hold up hold up...I dunno about that part."
Dear friends - America has always been this way. Study your 19th and 20th century history. Companies will exploit the shit out of us unless we put some rules in place to prevent it. Yes, that might mean making less money in the short term as regulations cause friction. But in the long term it means we can have a better and actually livable society.
(For what it's worth I'm an American and not an uppity European or Australian taking potshots from across the pond; no offense to Euros or Aussies intended, love you guys)
I mean, whenever I read these types of articles and responses, my reaction is almost always the same: what is this ideal world people are hoping for?
The future of everything is lies... sure, well so was the past, what's your point?
When we say "snake oil salesmen" we literally had people selling these mineral oils, fraudulently, for a century. People yearn for a time when there wasn't an antagonistic relationship between buyer and seller, but it has never existed. There is only one way have that relationship, and it's by having a personal -- repeatable -- relationship with your seller, and that's usually expensive, even if that expense is not monetary. It just the game theory of repeated games.
The easiest way to establish this behavior is to be a regular at a bar while being a good patron that creates a positive experience for everyone else. There are multiple places I frequent where I have no doubt in my mind that if there was an emergency, they might ask me to help out in some way because I'm trustworthy. And they could hand me an envelope of money, knowing I'd not steal it, because it's not worth it for me to take a bit of money and ruin my welcome at this place.
That's not going to happen with online sellers, and it's not going to happen with most corporations. The promise was always "save money through economies of scale, such that you still win even though there is no relationship." For the most part, that's proven true. But as everything has been outsourced to China, and anyone can make a basic version of everything, we're running up against the limits of those benefits. And I truly believe that brands are going to start mattering a lot more going forward for marketplaces.
Shop at Costco because they care about their customers. In other areas where companies don't, you're going to have to do the legwork to find out who is going to treat you right. That sucks, and it'll probably be more expensive, but that's the antagonistic relationship that buyers and sellers have naturally. Regulation helps, yes, but it's not some cure all. The expectation that you can just go with the person with the lowest price and get quality services is not a thing that's every existed.
It’s not that people are against regulations totally, but that the structure of society is broken. People with wealth, including corporations, can influence and control everything with money. Legislators are easily bribed. Lawsuits are expensive and take years. It’s hard to make anything happen unless you’re already rich or connected enough to access the right people.
The real issue is new amendments are needed. But that’s hard. You need 75% of states ratifying. And that seems impossible today on any topic.
I’d andd another layer - for American tech workers, regulation also reduces profits. This hurts salaries, stocks options and career growth.
Incentives make the world go round, so even if people recognize the issue, they would rather it become someone else’s problem, than willingly harm their own future.
But then...propose that we make some laws to put rules around this stuff, also known as regulations and everybody goes "whoa hold up hold up hold up...I dunno about that part."
Of course, what those who call for regulation of AI or other nascent technologies are really saying is, "Unqualified and/or biased and/or corrupt and/or dull-witted people should make decisions that affect us all, based on incomplete, misleading or rapidly-evolving information, with the power to enforce them at gunpoint."
And then they wonder why other countries beat their own.
I'm sorry...wut?
What regulations would you suggest?
> Since LLMs are unpredictable and vulnerable to injection attacks, customer service machines must also have limited power
Haha yes. I interacted with a bank one. It was like press 5 for mortgages but with a text to speech front end.
At the end of the day the LLM can be tricked into doing anything.
I wonder if there's also less of a stigma and sense of wrongdoing about tricking an LLM versus tricking an employee
We intuitively know that an employee will be punished and may get fired if we trick them. Many of us won't try to trick human employees as a result, because we would feel bad if they had bad consequences as a result of our trickery
There is likely no such hesitation around tricking LLMs. I know I personally wouldn't feel bad about it at all. Mostly because any computerized customer service process is annoying so anything I can do to limit my time dealing with it is a win in my books
D^HLying is easy, it's comedy that's hard...
So, providing actual customer service becomes a market differentiator?
"Yes, we cost more, but your get what you pay for" can be a good play.
„Agentic commerce means handing your credit card to a Large Language Model” - this is simply not true. LLMs/Agents will never get any credit/debit card details, they will be just an interface.
Full quote:
> People are very excited about “agentic commerce”. Agentic commerce means handing your credit card to a Large Language Model, giving it access to the Internet, telling it to buy something, and calling it in a loop until something exciting happens.
I think you're confusing this for the other side of things. The article talks about how some people already use OpenClaw and the variations, give them access to bunch of stuff including cards to purchase things (sometimes virtual and limited cards), I think that's what the article talks about when they say "agentic commerce".
Obviously a intentional simplification in the language the author uses, but I think it gets the point across at least.
Ok, in this case I misunderstood author's point here. "Agentic commerce" for me is a framework that networks (MC, Visa) and other big players in the ecosystem are working on. So it's far from "i'll give my credit cards details to openclaw and hope for the best". Obviously I know that a lot of people do that, unfortunately...
Millions of users, and you think exactly zero of them will hand over their credit card info?
my wife is one, so definitely not zero (she uses cards from privacy.com so there’s that…)
Excellent essay. I see some of this is already happening imo
This is doomerism. Yes, everything will get worse. But everything will also get better. Such is progress. (for every one of these examples of annoyances, I can think of two ways to use AI to get around the annoyance. not clever programmer things, but things an average person who learns to use Codex or Claude Desktop to operate their desktop will know)
Most of these annoyances are also things that existed before AI, and will continue to exist after, because consumerist capitalism. The good little obedient consumers get abused because they don't stand up for themselves. Customer service is an enfuriating maze? Yeah, because you voted with your dollars (and political indifference) to allow companies to make customer service (the thing you pay for) worse. We bring these problems on ourselves. It's pointless to complain if you aren't willing to do anything to change it. (And if you think you can't change it, there's other nations to look at, as well as the fact that you live in a democracy - for now - unlike the rest of the world)
Hell, we already have companies whose sole purpose is to manage your subscriptions for you because you're too lazy to do it yourself. You could look at this and say, man, the world is terrible! Or you could look at this and say, man, how great is my life that I can not only subscribe to a lot of things without going bankrupt, but I have extra cash left over to pay a company to manage my subscriptions?
Don't let the hedonic treadmill and complacency trick you into A) accepting a worse life, or B) convincing yourself your life is bad when it's actually better than most people's.
>(for every one of these examples of annoyances, I can think of two ways to use AI to get around the annoyance. not clever programmer things, but things an average person who learns to use Codex or Claude Desktop to operate their desktop will know)
As the author said:
>I suspect that like the job market, everyone will wind up paying massive “AI” companies to manage the drudgery they created.
It could also lead to a massive crash of capitalism and reevaluation of how our society functions.
It could lead to significant progress in every single research area.
I'm at least very impressed about the amount of open models and that it doesn't hold up that the gap between public and private diverges massivly. Public is probably one year behind.
> Yes, everything will get worse. But everything will also get better.
That is not known. Things could easily just get worse, and IMO that is far more likely. Every civilisation has collapsed, ours is clearly in decline, and AI could likely accelerate that decline.
I'm not exactly seeing the progress here. AI helps us write some software a bit faster? Doesn't seem revolutionary to me. Is it having any significant impact on peoples lives other than the various economic forces? I'm not seeing it.
> Yeah, because you voted with your dollars
In our system? No. In crony capitalism the companies who win do it through manipulating the political system. And when the government inflates the currency and destroys people buying power they simply cannot afford to "vote with their dollars". This is hilariously naive.
> That is not known. Things could easily just get worse, and IMO that is far more likely. Every civilisation has collapsed, ours is clearly in decline, and AI could likely accelerate that decline.
This is the definition of doomerism. Of all the ways each society has adapted to industrializaion, of all the technologies we've adapted to, and our continued use of all of them despite any negative impacts, you think an LLM is going to end things? An overcomplicated autocomplete?
There is no evidence whatsoever that suggests our society would collapse under AI. An imagination is not evidence. I mean, it's just a tool! Like the car, the telephone, the genetically engineered seed. Humans always freak out when they see something they don't understand. But it turns out there's no tool we have created that we can't control. We're stupid, but we're not idiots.
AI on AI warfare
I read the first couple of posts in the series. The essay is full of criticism of LLMs, and in a couple of places the author distances himself, as if he himself isn't using them ("some people I respect tell me that...").
It's certainly worth discussing the fact that the entire industry is starting to outsource large amounts of our thinking and writing work to non-sentient statistical algorithms, but this discussion needs to honestly confront the extent to which they are successfully completing useful tasks today.
> ML models will hurt innocent people.
Lots of blaming LLMs but I think the root cause lies elsewhere, I’m not even sure whether dismissing it as “capitalism” or “profit motives” would do it justice, because in general it feels more like the world that we live in lacks humanity.
Even in a capitalist world, a company could take a stance and decide not to purposefully screw people over, but in the world that we live in instead they look for ways to better screw over people and extract more money from them. It doesn’t matter whether your customer support is handled by someone from India, a crappy telephone tree or some voice model, when the incentive is the same - to do the bare minimum for customer “support” (in practice, just getting you to fuck off). Same for handling insurance claims and “dynamic pricing” of things - it doesn’t matter whether it’s some proprietary algorithm or just an LLM making crap up when the goal is to screw you over.
Blaming “AI” for all of this would be barking up the wrong tree (without that tech they’d just find other ways), though one can definitely acknowledge that this technology provides another convenient scapegoat, same as how you can lay employees off and just say cause it’s because of AI when in actuality it’s just greed and wanting to make your books look better.
A lot of this has been going on for a long time and I've been sensitive to it. LLMs may not be solely responsible but they're a massive escalation.
The articles didn’t blame LLMs, they talked about how they would get used, precisely through the lens of systems, incentives and culture.
In a capitalist world, the company that does decide to screw people over gets rich and the one that doesn't goes out of business.
It would be great if people chose not to do business with the former, but many simply do not care. They may think only other people get screwed. They may not take the time to think about it, especially if the company spends a ton of money obfuscating their misbehavior. Quite a few actively defend the right of companies to screw them.
Technology multiplies that like a lever. We weren't prepared for capitalism before LLMs and we're massively under-prepared now.
At various previous companies I've worked at product managers, executives, and engineers love bandying about the idea of "building for nontechnical users" as a way to make their widgets more "friendly". But it's just another way to otherize and denigrate "those people" who are the out group. They might, through a metacognitive defect or simple sociopathy, actually believe they're "doing good" by considering the poor creature's plight and making compassionate decisions on their behalf. But it's all crap. All they're actually doing is confirming their biases. LLMs are the divine nectar to these people, an enshittification accelerant par excellence.
Everything that is old is new again.
Payment processing, is better than it was in 2000, but still not good.
Micropayments: this is obnoxiously expensive to do.
Discovery, and discoverability: again here we have better but not good solutions (and many of the ones that were once good are enshitified).
Pricing: this is a problem everywhere, and frankly we need the law to change in a way that is pro consumer. Publishing prices, disclosure of fees, in both services and for payment processing (that 3 percent back from visa looks a lot less attractive when it's part of a 5 percent mark up).
Customer service: well there are already companies promoting models where they cut you off and send you into a black hole (google is a prime example). Good customer service will become a differentiator, and maybe a "paid for" service as well.
> Good customer service will become a differentiator
This does not matter without antitrust, which is why customer service became bad in the first place. 30 years ago, the low quality of customer service we complain about now simply didn't exist, at any size or professional level of business, and never had.
If a company back then had the customer service of the average company now, or even the average government agency now, people would have suspected that it was a covert front for criminals or spies.
If a company doesn't have to compete, it can cut everything until it only has the ghost of a product and a billing department. You don't boycott monopolies, monopolies boycott you. If three companies put you on a list to not have internet, phone service, a bank account or a credit card, etc., you just can't have them. You've become a European human rights judge.
To lie requires recognition of the truth and an intention to deceive. LLM’s don’t have such abilities. They are systems that generate plausible sequences of symbols based on training inputs, alignments, reinforcement, and inference. These systems don’t know or care what truth is and therefore cannot lie.
It’s already bad. I’m not looking forward to the future. These systems are terrible. It’s a future without people that they want for some reason. I’d rather deal with people incompetent, tired, annoyed people than an LLM.
Ill-thought out logic.
The company that deployed the LLM is lying to you. The people who made that decision are the ones who are culpable.
We both agree that it’s terrible.
I think it’s important to have an enforcement mechanism to force companies to do what they are responsible for doing. An Anti-Kafka Law, so to speak.
An important distinction to make, and I whole heartedly agree.
It’s not LLMs replacing workers, it’s people. People who have a lot of money and don’t sell their labour for a paycheque. And the systems that compel them to such actions.
Don't agree with this.
LLM when it came out, was perfect as an interface between a system and a normal human.
So many people call customer support for issues they could in theory fix themselves. If that LLM system can understand me well enough, its an okay interface.
In worst case you have to escalate anyway. My mum actually told me that she talked to some AI.
And yes normal systems are also not correct often enough. With AI/LLM software will get cheaper which should incresase quality overall.
I dont think ai/llm in this case will change anything.
Relevant change will happen due to the fact that humans can be replaced by AI/LLMs. It was not even imaginable a few years back how a good ai system would even look like. Translaters lost their jobs, basic arists lost their jobs. Small contracts for basic things are gone. The restaurant poster no one cares? AI. The website translation for some small business? no one cares.
>LLM when it came out, was perfect as an interface between a system and a normal human.
Statements like this make me feel like I live in a different universe with a different implementation of LLMs than other internet commenters.
Do you want to add any argument so we can discuss this?
I mean, did you not write with ChatGPT and were surprised how well it response?
I'm schocked how well i can talk to an AI through some app like Gemini or ChatGTP. A few years ago i couldn't imagine building such a generic system which such high quality of understanding.
I was playing around with dragon naturally speaking and similiar dictation tools 10 years ago and it was horrible. And that software is expensive.
If you look how normal people use a computer, they are slow just because they don't understand basic drag and drop. Or they are unable to just create some java or php script to convert some data or clean up some data. I would just write a php script reading some csv file and converting stuff around and was faster than everyone around me.
Tool calling is bonkers.
And i tried to break GPT-3, i can literaly write an english sentence and just dropin german words, it was already that good.
Its often enough shitty in doing exactly what i want, but the quality is massive to everything we had before. Massive.
Not the OP, but you wrote “LLM when it came out, was perfect as an interface between a system and a normal human”. That’s a specific and very encompassing claim. I can only think of very simplistic systems (like a microwave oven maybe) where a current LLM could function perfectly as the sole command interface, much less when LLMs first became available. For systems of any significant complexity, it tends to turn into an exercise in frustration and failure modes when the LLM is your only interface (and frequently even when it isn’t).
An LLM can enhance the interface of a system and can be really useful in that despite its imperfections. But that’s a very different claim.
It was a significant jump from whatever we had before to a quality unseen before. As i mentioned, i threw english and german at it.
How many people can change the time on their microwave?
How many people can ask an LLM through voice or text to change the time of the microwave?
A LLM is an interface to a service if you add a MCP Server. Now i can ask Jira things like "hey whats my current task? And what do i need to do?"
Its also an interface to documentation. I asked it to help me build up a hugo templating based website because just reading the hugo docs did not help me as much as the LLM did (and that was 2 years ago).
In best case, as long as an LLM is not AGI or ASI, we have good tools with validation behind the LLMs before the LLM becomes the system itself.
You're on a forum with a disproportionate number of people who are trying to profit from AI and have an interest in promoting that it's a worthwhile time and resource investment. It is a different universe than other places outside this bubble.
And it's a one day old account.
> My mum actually told me that she talked to some AI.
You have no argument here. Make an argument then we can talk. Right now it’s going in circles.
I added talking points like the one were i state that people call support just to fix issues they could fix themselves.
My point with my mum should imply that it was successful but for sure at least you are pointing something out and now we can talk about it: My mum talked to an AI and it helped her.