Crypto is a lifeline for me, as I cannot open a bank account in the country I live in, for reasons I can neither control nor fix. So I am happy if crypto is useless for you. For me and for millions like me, it is a matter of life and death.
As for LLMs — once again, magic for some, reliable deterministic instrument for others (and also magic). Just classified and sorted a few hundreds of invoices. Yes, magic.
Similar argument to https://www.baldurbjarnason.com/2025/trusting-your-own-judge..., but I like this one better because at least it doesn’t try to pull the rhetorical trick of slipping from “we can’t know whether LLMs are helping because we haven’t studied the question systematically” to “actually we do know, and they’re shit”.
I think that the disparity comes between people that are too in the weeds believing that their use cases apply to everyone. The reality is this world is made up of people with a wide array of different needs and AI is yet to proliferate into all usage applications.
Sure some of this comes from a lack of education.
But similar to crypto these movements only have value if the value is widely perceived. We have to work to continue to educate, continue to question, continue to understand different perspectives. All in favor of advancing the movement and coming out with better tech.
I am a supporter of both but I agree with the reference in the article to both becoming echo chambers at times. This is a setback we need to avoid.
Ok. Claude Code produces most code at Anthropic. Theres an enterprise code base, with acute real needs. There are real, experienced SWEs. How much babysitting and reviewing is undetermined; but the Ants seem to tremendously prefer the workflow.
Even crypto people didn’t dogfood their crypto like that, on their own critical path.
The best way I’ve heard this described: AI (LLMs) is probably 90% of the way to human levels of reasoning. We can probably get to about 95% optimizing current technology.
Whether or not we can get to 100% using LLMs is an open research problem and far from guaranteed. If we can’t, it’s unclear if it will ever really proliferate the way things hope. That 5% makes a big difference in most non-niche use cases…
I follow Emily Bender on LinkedIn. She cuts through the AI hype and is also the author of The AI Con book - https://thecon.ai/
Of course people will either love AI or hate AI - and some don’t care. I am cautious especially when people say ‘AI is here to stay’. It takes away agency.
One thing I find frustrating is that management where I work has heard of 10x productivity gains. Some of those claims even come from early adopters at my work.
But that sets expectation way too high. Partly it is due to Amdahl's law: I spend only a portion of my time coding, and far more time thinking and communicating with others that are customers of my code. Even if does make the coding 10x faster (and it doesn't most of the time) overall my productivity is 10-15% better. That is nothing to sneeze at, but it isn't 10x.
I'm a retired programmer. I can't imagine trusting code generated by probablities for anything mission critical. If it were close and just needed minor tweaks I could understand that. But I don't have experience with it.
My comment is mainly to say LLMs are amazing in areas that are not coding, like brainstorming, blue sky thinking, filling in research details, asking questions that make me reflect. I treat the LLM like a thinking partner. It does make mistakes, but those can be caught easily by checking other sources, or even having another LLM review the conclusions.
You don't trust the code coming out of the probabilistic machine. You build a validation cage around it with hard interfaces and you also review the output.
The thing is, the questions such as “are they an expert in the domain” … “are they good at coding to being with” … and so on only really apply to the folks claiming positive results from LLMs. On the flip side, someone not getting much value - or dare I say, a skeptic - pushes back because they _can see_ what the LLM gave them is wrong. I’m not providing any revelatory comment here, but the simple truth is: people who are shit to begin with think this is all amazing/magic/the future.
I have to say I’m in the exact camp the author is complaining about. I’ve shipped non trivial greenfield products which I started back when it was only ChatGPT and it was shitty. I started using Claude with copying and pasting back and forth between the web chat and XCode. Then I discovered Cursor. It left me with a lot of annoying build errors, but my productivity was still at least 3x. Now that agents are better and claude 4 is out, I barely ever write code, and I don’t mind. I’ve leaned into the Architect/Manager role and direct the agent with my specialized knowledge if I need to.
I started a job at a demanding startup and it’s been several months and I have still not written a single line of code by hand. I audit everything myself before making PRs and test rigorously, but Cursor + Sonnet is just insane with their codebase. I’m convinced I’m their most productive employee and that’s not by measuring lines of code, which don’t matter; people who are experts in the codebase ask me for help with niche bugs I can narrow in on in 5-30 minutes as someone whose fresh to their domain. I had to lay off taking work away from the front end dev (which I’ve avoided my whole career) because I was stepping on his toes, fixing little problems as I saw them thanks to Claude. It’s not vibe coding - there’s a process of research and planning and perusing in careful steps, and I set the agent up for success. Domain knowledge is necessary. But I’m just so floored how anyone could not be extracting the same utility from it. It feels like there’s two articles like this every week now.
More anecdata: +1 for “LLMs write all my production code now”. 25+ years in industry, as expert as it’s possible to be in my domain. 100% agree LLMs fail hilariously badly, often, and dangerously. And still, write ~all my code.
No agenda here, not selling anything. Just sitting here towards the later part of my career, no need to prove anything to anyone, stating the view from a grey beard.
Crypto hype was shill from grifters pumping whatever bag holding scam they could, which was precisely what the behavioral economic incentives drove. GenAI dev is something else. I’ve watched many people working with it, your mileage will vary. But in my opinion (and it’s mine, you do you), hand coding is an apocryphal skill. The only part I wonder about is how far up and down the system/design/architecture stack the power-tooling is going to go. My intuition and empirical findings incline towards a direction I think would fuel a flame war. But I’m just grey beard Internet random, and hey look, no evidence just more baseless claims. Nothing to see here.
Disclosure: I hold no direct shares in Mag 7, nor do I work for one.
ChatGPT can write research papers in about 20 minutes—its the “Deep Research” tool. These are not original papers, but it can perform complex tasks that require multiple steps that would normally take a person hours. No its not a magic superintelligence, but it will transform a lot of white collar labor.
_So much_ work in the 'services' industries globally comes down to really a human transposing data from one Excel sheet to another (or from a CRM/emails to Excel), manually. Every (or nearly every) enterprise scale company will have hundreds if not thousands of FTEs doing this kind of work day in day out - often with a lot of it outsourced. I would guess that for every 1 software engineer there are 100 people doing this kind of 'manual data pipelining'.
So really for giant value to be created out of LLMs you do not need them to be incredible at OCaml. They just need to ~outperform humans on Excel. Where I do think MCP really helps is that you can connect all these systems together easily, and a lot of the errors in this kind of work came from trying to pass the entire 'task' in context. If you can take an email via MCP, extract some data out and put it into a CRM (again via MCP) a row at a time the hallucination rate is very low IME. I would say at least a junior overworked human level.
Perhaps this was the point of the article, but non-determinism is not an issue for these kind of use cases, given all the humans involved are not deterministic either. We can build systems and processes to help enforce quality on non deterministic (eg: human) systems.
Finally, I've followed crypto closely and also LLMs closely. They do not seem to be similar in terms of utility and adoption. The closest thing I can recall is smartphone adoption. A lot of my non technical friends didn't think/want a smartphone when the iPhone first came out. Within a few years, all of them have them. Similar with LLMs. Virtually all of my non technical friends use it now for incredibly varied use cases.
Everything? As a lawyer, I’m producing 2x - with fewer errors. Admittedly, law is a field that mostly involves shuffling words around so it may be the best case scenario, but much of the skepticism comes off as cope.
Software methodologies and workflows are not engineering either, yet we spend a fair amount of time iterating and refining those. You can very much become better at prompt engineering. There is a huge differential between individuals, for example.
The code coming out of LLMs is just as deterministic as code coming out of humans, and despite humans being feckle beings, we still talk of software engineering.
As for LLMs, they are and will forever be "unknowable". The human mind just can't comprehend what a billion parameters trained on trillions of tokens under different regimes for months corresponds to. While science has to do microscopic steps towards understanding the brain, we still have methods to teach, learn, be creative, be rigorous, communicate that do work despite it being this "magical" organ.
With LLMs, you can be pretty rigorous. Benchmarks, evals, and just the vibes of day to day usage if you are a programmer, are not "wishful thinking", they are reasonably effective methods and the best we have.
Loosely related, but I find the use of AGI (and sometimes even AI) as terms annoying lately. Especially in scientific papers, where I would imagine everything to be well defined. If at least in how it is used in that paper.
So, why can't we just come up with some definition for what AGI is? We could then, say, logically prove that some AI fits that definition. Even if this doesn't seem practically useful, it's theoretically much more useful than just using that term with no meaning.
Instead it kind of feels like it's an escape hatch. On wikipedia we have "a type of ai that would match or surpass human capabilities across virtually all cognitive tasks". How could we measure that? What good is this if we can't prove that a system has this property?
Bit of a rant but I hope it's somewhat legible still.
This reads like the author is mad about imprecision in the discourse which is real but to be quite frank more rampant amongst detractors than promoters, who often have to deal with the flaws and limitations on a day to day basis.
The conclusion that everything around LLMs is magical thinking seems to be fairly hubristic to me given that in the last 5 years a set of previously borderline intractable problems have become completely or near completely solved, translation, transcription, and code generation (up to some scale), for instance.
I think it is more like googling: when the search engine appeared, everybody had to learn how to write a good query, even though the expectation was that everybody could use them.
With LLMs, it's quite similar: you have to learn how to use them. Yes, they are non-deterministic, but if you know how to use them, you can increase your chances of getting a good result dramatically. Often, this not only means articulating a task, but also looking at the bigger picture and asking yourself what tasks you should assign in the first place.
For example, I can ask the LLM to write software directly, or I can ask it to write user stories or prototypes and then take a multi-step approach to develop the software. This can make a huge difference in reliability.
And to be clear, I don't mean that every bad result is caused by not correctly handling the LLM (some models are simply poor at specific tasks), but rather that it is a significant factor to consider when evaluating results.
1. he talks about what he's shipped, and yet compares it to crypto – already, you're in a contradiction as to your relative comparison – you straight up shouldn't blog if you can't conceive that these two are opposing thoughts
2. this whole refrain from people of like, "SHOW ME your enterprise codebase that includes lots of LLM code" – HELLO, people who work at private companies CANNOT just reveal their codebase to you for internet points
3. anyone who has actually used these tools has now integrated them into their daily life on the order of millions of people and billions of dollars – unless you think all CEOs are in a grand conspiracy, lying about their teams adopting AI
Crypto and NFT situation happened because of our society, media and vc/startup landscape who hype things up a lot for their own reasons. We treat massive technologies as new brands of bottled water. Or, actually, as a new hype toy as fidget spinners or pop it toys. This tech is massively more complex and you have to invest time to learn about its abilities, limitations and potential developments. Almost nobody actually does this, it's easier to follow hype train and put money into something that grows up and looks cool without obvious cons. Crypto is cool for some stuff. On the other hand, where's your Stepn (and move to earn in general), decentraland cities, Apes that will make a multimedia universe? Where's "you'll be paying using crypto for everything"?
Same for LLMs and AI: it is awesome for some things and absolutely sucks for other things. Curiously tho, it feels like UX was solved by making chats, but it actually still sucks enormously, as with crypto. It is mostly sufficient for doing basic stuff. It is difficult to predict where we'll land on the curve of difficult (or expensive) vs abilities. I'd bet AI will get way more capable, but even now you can't really deny its usefulness.
59 comments
[ 5.4 ms ] story [ 74.1 ms ] threadCrypto is a lifeline for me, as I cannot open a bank account in the country I live in, for reasons I can neither control nor fix. So I am happy if crypto is useless for you. For me and for millions like me, it is a matter of life and death.
As for LLMs — once again, magic for some, reliable deterministic instrument for others (and also magic). Just classified and sorted a few hundreds of invoices. Yes, magic.
Sure some of this comes from a lack of education.
But similar to crypto these movements only have value if the value is widely perceived. We have to work to continue to educate, continue to question, continue to understand different perspectives. All in favor of advancing the movement and coming out with better tech.
I am a supporter of both but I agree with the reference in the article to both becoming echo chambers at times. This is a setback we need to avoid.
Even crypto people didn’t dogfood their crypto like that, on their own critical path.
Whether or not we can get to 100% using LLMs is an open research problem and far from guaranteed. If we can’t, it’s unclear if it will ever really proliferate the way things hope. That 5% makes a big difference in most non-niche use cases…
Of course people will either love AI or hate AI - and some don’t care. I am cautious especially when people say ‘AI is here to stay’. It takes away agency.
includes the 3rd law, which reads, and seems on topic,
"Any sufficiently advanced technology is indistinguishable from magic."
But that sets expectation way too high. Partly it is due to Amdahl's law: I spend only a portion of my time coding, and far more time thinking and communicating with others that are customers of my code. Even if does make the coding 10x faster (and it doesn't most of the time) overall my productivity is 10-15% better. That is nothing to sneeze at, but it isn't 10x.
My comment is mainly to say LLMs are amazing in areas that are not coding, like brainstorming, blue sky thinking, filling in research details, asking questions that make me reflect. I treat the LLM like a thinking partner. It does make mistakes, but those can be caught easily by checking other sources, or even having another LLM review the conclusions.
I started a job at a demanding startup and it’s been several months and I have still not written a single line of code by hand. I audit everything myself before making PRs and test rigorously, but Cursor + Sonnet is just insane with their codebase. I’m convinced I’m their most productive employee and that’s not by measuring lines of code, which don’t matter; people who are experts in the codebase ask me for help with niche bugs I can narrow in on in 5-30 minutes as someone whose fresh to their domain. I had to lay off taking work away from the front end dev (which I’ve avoided my whole career) because I was stepping on his toes, fixing little problems as I saw them thanks to Claude. It’s not vibe coding - there’s a process of research and planning and perusing in careful steps, and I set the agent up for success. Domain knowledge is necessary. But I’m just so floored how anyone could not be extracting the same utility from it. It feels like there’s two articles like this every week now.
No agenda here, not selling anything. Just sitting here towards the later part of my career, no need to prove anything to anyone, stating the view from a grey beard.
Crypto hype was shill from grifters pumping whatever bag holding scam they could, which was precisely what the behavioral economic incentives drove. GenAI dev is something else. I’ve watched many people working with it, your mileage will vary. But in my opinion (and it’s mine, you do you), hand coding is an apocryphal skill. The only part I wonder about is how far up and down the system/design/architecture stack the power-tooling is going to go. My intuition and empirical findings incline towards a direction I think would fuel a flame war. But I’m just grey beard Internet random, and hey look, no evidence just more baseless claims. Nothing to see here.
Disclosure: I hold no direct shares in Mag 7, nor do I work for one.
A bit suspicious, wouldn’t you agree?
_So much_ work in the 'services' industries globally comes down to really a human transposing data from one Excel sheet to another (or from a CRM/emails to Excel), manually. Every (or nearly every) enterprise scale company will have hundreds if not thousands of FTEs doing this kind of work day in day out - often with a lot of it outsourced. I would guess that for every 1 software engineer there are 100 people doing this kind of 'manual data pipelining'.
So really for giant value to be created out of LLMs you do not need them to be incredible at OCaml. They just need to ~outperform humans on Excel. Where I do think MCP really helps is that you can connect all these systems together easily, and a lot of the errors in this kind of work came from trying to pass the entire 'task' in context. If you can take an email via MCP, extract some data out and put it into a CRM (again via MCP) a row at a time the hallucination rate is very low IME. I would say at least a junior overworked human level.
Perhaps this was the point of the article, but non-determinism is not an issue for these kind of use cases, given all the humans involved are not deterministic either. We can build systems and processes to help enforce quality on non deterministic (eg: human) systems.
Finally, I've followed crypto closely and also LLMs closely. They do not seem to be similar in terms of utility and adoption. The closest thing I can recall is smartphone adoption. A lot of my non technical friends didn't think/want a smartphone when the iPhone first came out. Within a few years, all of them have them. Similar with LLMs. Virtually all of my non technical friends use it now for incredibly varied use cases.
The code coming out of LLMs is just as deterministic as code coming out of humans, and despite humans being feckle beings, we still talk of software engineering.
As for LLMs, they are and will forever be "unknowable". The human mind just can't comprehend what a billion parameters trained on trillions of tokens under different regimes for months corresponds to. While science has to do microscopic steps towards understanding the brain, we still have methods to teach, learn, be creative, be rigorous, communicate that do work despite it being this "magical" organ.
With LLMs, you can be pretty rigorous. Benchmarks, evals, and just the vibes of day to day usage if you are a programmer, are not "wishful thinking", they are reasonably effective methods and the best we have.
So, why can't we just come up with some definition for what AGI is? We could then, say, logically prove that some AI fits that definition. Even if this doesn't seem practically useful, it's theoretically much more useful than just using that term with no meaning.
Instead it kind of feels like it's an escape hatch. On wikipedia we have "a type of ai that would match or surpass human capabilities across virtually all cognitive tasks". How could we measure that? What good is this if we can't prove that a system has this property?
Bit of a rant but I hope it's somewhat legible still.
The conclusion that everything around LLMs is magical thinking seems to be fairly hubristic to me given that in the last 5 years a set of previously borderline intractable problems have become completely or near completely solved, translation, transcription, and code generation (up to some scale), for instance.
Google Translate, Whisper and Code Generators (up to some scale) have existed for quite some time without using LLMs.
With LLMs, it's quite similar: you have to learn how to use them. Yes, they are non-deterministic, but if you know how to use them, you can increase your chances of getting a good result dramatically. Often, this not only means articulating a task, but also looking at the bigger picture and asking yourself what tasks you should assign in the first place.
For example, I can ask the LLM to write software directly, or I can ask it to write user stories or prototypes and then take a multi-step approach to develop the software. This can make a huge difference in reliability.
And to be clear, I don't mean that every bad result is caused by not correctly handling the LLM (some models are simply poor at specific tasks), but rather that it is a significant factor to consider when evaluating results.
Millions of beginner developers running with scissors in their hands, millions of investment in the garbage.
I don't think this can be reversed anymore, companies are all-in and pot commited.
1. he talks about what he's shipped, and yet compares it to crypto – already, you're in a contradiction as to your relative comparison – you straight up shouldn't blog if you can't conceive that these two are opposing thoughts
2. this whole refrain from people of like, "SHOW ME your enterprise codebase that includes lots of LLM code" – HELLO, people who work at private companies CANNOT just reveal their codebase to you for internet points
3. anyone who has actually used these tools has now integrated them into their daily life on the order of millions of people and billions of dollars – unless you think all CEOs are in a grand conspiracy, lying about their teams adopting AI
Same for LLMs and AI: it is awesome for some things and absolutely sucks for other things. Curiously tho, it feels like UX was solved by making chats, but it actually still sucks enormously, as with crypto. It is mostly sufficient for doing basic stuff. It is difficult to predict where we'll land on the curve of difficult (or expensive) vs abilities. I'd bet AI will get way more capable, but even now you can't really deny its usefulness.