I mean, I’m glad he said it but anyone who’s been paying attention already knows this. The entire country is collectively waiting for the bubble to burst as it is. You also have to separate things like Machine Learning from LLMs. Two totally different use cases. They’ve not even clearly articulated any end goals. Even NASA had ‘put a man on the moon’, not just ‘more space’. From what I can tell their goals boil down to either ‘make money’ or ‘make God’. The former works great for individuals but not a whole class of technology and the latter would surely not garner trillions in investment dollars. LLMs are certainly great productivity tools but what is their goal really for a CEO to be able to fire their entire staff so they can sit down at a computer and type in ‘make my business better’ and they think that’ll work?
It’s changed the price of RAM, that much is assured.
The whole world is lying because that’s how we get fucking money. What a stupid hole we’ve dug ourselves.
Here’s a link to the exact same post on Cory Doctorow’s (tracker-free) blog: https://pluralistic.net/2026/08/01/dare-snot/#i-will-fucking-piledrive-you-if-you-mention-ai-again
AI has been helpful for my job, sure, but the cost in terms of societal damage is so great it makes me wish AI were never created.
Work keeps insisting we use it and I keep telling them to fuck off.
I get more done than my colleagues BECAUSE i don’t use the not-actually-AI “AI” tool.
In my academic field of sustainability, AI still fails to explain what economic sustainability means. It’s very simple.
I lecture and explain to my students that economic sustainability is about how we assess and manage our knowledge, innovations, wealth, and other man-made capital. I stress that it is ABSOLUTELY NOT profit and revenue.
Sure enough, my students submit assignments clearly written by AI that state economic sustainability is profit and revenue.
AI is regurgitating decades of green-washing and even bad academic articles that state economic sustainability is about profits. I use that AI god mode site to test several AI’s at once and they’ve consistently gotten this wrong for 3 years.
The story itself is pretty ridiculous. He starts with a conclusion, and then looks for stories to support that narrative. What should happen, is first establish criteria to measure if an AI project has been successful or not and then measure projects against that criteria. Instead the author uses anecdotal stories of projects he has heard failed, so it must all be wrong.
Not one mention of the advances in medical research using AI, because it goes against his narrative.
I’m so happy everyone in the comments is engaging in a thoughtful and informed manner and not just spouts uninformed nonsensical talking points from both sides to see who can shout the loudest.
AI is changing everything. Its not a lie, its absolute truth.
The problem is the changes are fuckawful terrible for everyone and everything.
I typed into Google “potatoes pressure cooker” to get a reminder of how long to cook them. The AI told me, and correctly said that I should use the steamer sieve, but also told me to salt the water. The water that would be under the potatoes and that wouldn’t touch them.
I asked why, and the AI said that little splashes of salt water from the boiling would land on the surface of the potatoes and salt them gently, that the salt would increase the boiling temperature, and that the vapour would be aromatic.
I pointed out that the amount of salt landing on the surface of the potatoes would be negligible, that the vapour is distilled water in gas form, and that the increase in boiling temperature from salting the water is less than 1°C. The AI said, oh yeah, you’re right on all counts. Don’t salt the water.
What it has been doing was exactly what an LLM does. It gave me the received wisdom from the internet.
That aligns with what Cory Doctorow was saying here, namely that all the AI successes were from people who already knew what they were doing and could use the AI as a tool. Not from people who had no clue and just let AI do the job.
I can’t say AI hasn’t changed how I work, it has. Instead of going to Google, vendor docs, or the manpage for whatever code or system I am working on I will ask Claude or Gemini to take a stab at building the initial pass of the code or to look up what some cryptic error code means for me. Or even a better way to write some code that I think can be more concise.
The key I think is I never let it make the final decisions for me. The workflow is always look up information > shows it to me > ask for permission to do the thing > do the thing.
But none of that is “changing everything” in the way executives and the Sam Altman’s of the world want to profess. It is just a sometimes more efficient for me way of doing the same thing I was doing before.
A lot of people I have seen that crow about how they are using AI to change their work lives are just doing really fancy Github workflows and actions to automate something they used to still do manually for some reason. They rarely are doing anything that couldn’t be done with the time to research and test without AI.
In other words, corporate leadership is starting from the premise that AI has (or will) radically change the business, and they’re working backwards from that premise to find the evidence to support this article of faith.
Sounds familiar. The proof-of-work blockchain scheme which was the fad a while ago is also often described as a solution in search of a problem.
I feel like block chain as a public ledger has found quite a few purposeful uses, and was a solution to a handful of problems most people rarely if ever encounter. Then fake money got involved
The core issue feels like the dead torrent problem.
Its distributed, so it needs to be hosted, by several locations. Which becomes probitive over time as.it becomes huge. Its also massively inefficient for what its trying to do.
At some point, its basically just one server farm somewhere hosting a glorified MySWL database that doubles as a space heater for the entire planet every time you need to add a row.
I was thinking that with legal backing, blockchain could be a good record of ownership for deeds. Like NFTs but for physical property.
The history of ownership and the public visibility are good features there.
Lack of reliable, authoritative attribution of changes and no rollback or compensatory transactions are all dealbreakers.
Yeah who doesn’t want a ledger you cannot change, sounds like a good idea. /s
In the case of deeds, i think it is desirable because it gives you an immutable history of ownership.
There are other problems but i don’t believe that is one.
No it doesn’t. It only works because 51% doesn’t change the ledger.
You already have this with a simple database for ownership. It already exists, and ledgers made out of paper predates that.
But with good legal backing, all you need is a shared database and maybe basic cryptographic signatures to prove ownership in case of malicious actors. So basically, traditional banking.
Yes, exactly.
Block chain would have been an amazing solution to most democratic vote validation processes.
Well no, because we don’t have a problem that blockchain solves (better than open software, if it must be digital), and there are loads of problems that the blockchain doesn’t solve. Like how am I supposed to vote with that thing under no coercion.
I’m not really sure what you’re getting at. The blockchain is a good mechanism to maintain a source of truth that is inspectable and verifiable by anyone. It would only serve to certify that official counts can’t be tampered with in certain ways.
But it doesn’t replace the whole democratic voting process. It doesn’t protect against coercive voting, for instance.
My point originally was that blockchain can be of use to us, but tech bros ruined its rep. Same with AI. I’ve been waiting for routers with ai-powered (really, just machine learning) firewalls that can adapt rules to environments. But instead of that, everyone is obsessed with generative stuff that isnt very good.
Yeah I hear you, but the 51% attack exists. Like if you use paper ballots, you’re safer and it consumes less energy. There is just nothing better by using a blockchain, and paper ballots (for example) can be verified by anyone, the blockchain not so much.
Aren’t votes meant to be anonymous to avoid retribution?
Isn’t the public ledger identifying information to be used by authoritarian violence on every single voter in that case?
Proposed ideas for block chain to maintain election integrity would only use a block chain ledger to verify counts, not store voter information.
It can still be used that way. While most people talk about how corrupt politicians enrich themselves with fake currencies, more legal and legitimate uses can still use decentralized ledgers for useful things. Of course these won’t make headlines.
There are far better global ledger approaches than blockchain.
I believe you, but can you name one so we can all learn more about the alternatives out there?
SWIFT is the easy one. Fun fact they looked into blockchain at one point and went “yeah, nah”
I thought Swift is currently Block-Chain based. Do you mean this Swift?
Which purposeful uses did you have in mind?
Nothing that a centralized server wouldn’t be able to handle better.
Or several other solutions for different use cases. I always remember this article about alternatives: https://gist.github.com/joepie91/a90e21e3d06e1ad924a1bfdfe3c16902
I’m glad you specifically pointed out proof-of-work blockchains.
They’re very inefficient (economically and ecologically) by design. In almost all cases this design isn’t warranted.
Alas there are other designs and while most of them are rubbish as well, a few ones are doing things quite right.I hope that just like other schemes than proof-of-work were thought of regarding blockchains, there will be AI models that are way, way more efficient and ideally can be run locally - for those cases that can use AI…
Wasn’t Nano predicated on those lines? I remember reading transactions could be done locally, even off-line, validated beetween the parts and only later added to the global ledger.
Nano uses “proof of stake” instead of proof of work to decide the order of transactions, and who receives block rewards.
The main problem with PoS is that it’s essentially the same as Federal Reserve bonds: all the new money goes towards people with extra money to freeze. This is part of why inequality has spiraled out of control since the Nixon Shock. Proof of work literally burns most of the profits because of difficulty adjustment.
A more specific problem with Nano (formerly RaiBlocks) is that the entire supply was centrally issued, with a pinky promise from this private organization that they only issued coins by CAPTCHA. If they were lying, then they could have issued 51% of the supply to themselves for permanent control. The only way we’d be able to detect it is if the price kept going down for years.
That requires a global ledger. Blockchain is an appallingly bad way to implement one.
What I understood then was that the local ledger of a given user of the coin would authorize and register the transactions done on/off the recorded balance and, when network available, would broadcast such transactions to the global ledger.
The global ledger was the total sum of transactions, not imperatively required to authorize transaction by transaction.
Funny, that you even know about Nano!
…it’s one of the often overlooked projects because it doesn’t have a ton of fuck-off-money and instead tries to focus on a solid protocol.Nano has a lot of interesting attributes, but I fail to see how what you describe would work in practice.
If both parties want to make sure there are no shenenigans at play, they need to know about the most recent state of the respective account chains, which essentially requires them to be online for agreeing on said transaction.
But overall Nano is very fast and efficient by design and only a failure in terms of “gainz for Lambo”.As we’re here in a thread about AI I should remark that machine-to-machine-payments - in this case: agent-to-agent-payments - would work pretty well with Nano as currency because of the transaction finality (typically less than 1 second) and the feeless nature of transactions.
If AI agents are looking for the most viable way to transfer tiny amounts of value fast and without fees they might find Nano and use it - who knows…
…and just like Nano showed that efficient ways to create digital money are possible I’m hoping for efficient AI models that are economically and ecologically worthwhile.What I understood then was that in person transactions could be done, as the local ledger of each user would authorize and record operations on and off the available balance and wait until network availability to syncronize with the global record.
Returning to the subject at hand: I can imagine very specialized “AI” being useful for scientifical research, where very knowledgeable people use it as a tool to facilitate processes but are nonetheless capable of reviewing whatever results it produces.
Not gigantic datacenters required for this but small, purpose made and perfected, locally run, even if on higher specifications hardware to do so, but machines built for a given task and purpose. The economic viability on it be damned; it’s a tool for research, it is not made to earn money.
To put it bluntly: if you do an offline transaction, you’re prone to fraud.
If you expect the senders account chain to have balance x (because that’s your offline record for that) and the sender has sent all funds to a different address after you synced that account chain, you receive money that isn’t there - kind of like an invalid cheque.
I have no clue how that would work in practice, because to know a random account in advance, you’d have to sync the whole amount of account chains there is (called block lattice in Nano’s case).
With a mobile device that’s hardly feasible and without a mobile device I don’t see how you’d get in contact with people to make such an offline transaction.That kind of specialized AI is what I imagine to be a use case for locally run AI, too.
After all you don’t want to build processes on an AI, where you have zero control over what happens behind the curtains.
That includes feeding potentially sensitive data back to it as well as being unable to control the training data set, its learning, version numbers, etc.
That’s top down thinking for you.
AI would be great if(currently its mixed for me,and whats the best thing you can do ATM ):
- The Datacenters didnt drain like 3 oceans and half of the power grid(or used solar energy + air cooling only,the only way to solve this is Local LLMs currently or specific AI solutions that isnt ChatGPT or Gemini and so on. )
- the LLM is not trained on copyrighted data(non-public domain) and not Paying the artists or not ask the original creator to train on their work.(Local LLMS partially solve this,and this requires specific LLMS too.)
- People did not use AI mindlessly to write Code for you(well i think this requires the user to use their brain)
and so on…
and AI should stay as a “TOOL”, not a full replacement for human work like most sane people do.
edit 1:
i forgot to mention AI is a big contributer to the AI crisis.I mean, the article actually says that they are not lying.
But the truth they are telling is that this kind of tools is like Yuuzhan-Vong assistants in Star Wars NJO era. It’s just further in the direction of auto-complete and recommendations.
That’s good. It also means that a real job can’t be “done by AI”.
It’s another kind of data compression, in a very fundamental and general way.
It accelerates the most things that shouldn’t be done at all. Boilerplate code that’s written again and again because of wrong processes, which are taught as standard way of thinking and doing development. Other things that could be a higher-level programming/markup language or a library. Just a frontal way of solving things that were done by hand.
That’s useful, one can’t do everything by hand.
Just - things that require real human decisions, understanding, responsibility, choice, or semantic and semiotic context that has formed in the course of one’s own life for like 20 years - that they can’t do. And that’s what forms a professional. They can extrapolate stuff that doesn’t matter.
