libraryofbabel

I actually like this article a lot. I do a bit of teaching, and I imagined the ideal audience for this as a smart junior engineer who knows SQL and has encountered transactions but maybe doesn’t really understand them yet. I think introducing things via examples of isolation anomalies (which most engineers will have seen examples of in bugs, even if they didn’t fully understand them) gives the explanation a lot more concreteness than starting with serializability as a theoretical concept as GP is proposing. Sure, strict serializability is a powerful idea that ties all this together and is more satisfying for an expert who already knows this stuff. But for someone who is just learning, you have to motivate it first.

If anything, I’d say it might be better to start with the lower isolation levels first, highlight the concurrency problems that can arise with them, and gradually introduce higher isolation levels until you get to serializability. That feels a bit more intuitive rather than downward progression from serializability to read uncommitted as presented here.

It also might be nice to see a quick discussion of why people choose particular isolation levels in practice, e.g. why you might make a tradeoff under high concurrency and give up serializability to avoid waits and deadlocks.

But excellent article overall, and great visualizations.

Is that actually true, if the charges differed at the 12th decimal place only? That’s non-obvious to me.

There’s a name for that: the Anthropic principle. And it is deeply unsatisfying as an explanation.

And does it even apply here? If the charge on the electron differed from the charge on the proton at just the 12th decimal place, would that actually prevent complex life from forming. Citation needed for that one.

I agree with OP. The unexplained symmetry points to a deeper level.

Slightly tangential, but I discovered recently that the famous literary critic Harold Bloom was a huge fan of Ursula Le Guin and rated her one of the great canonical writers of the 20th century, in all of literature not just sci-fi. Also, they never met but they struck up a polite friendship over email when they were both old and chatted back and forth.

Some might consider this raises the stature of Ursula Le Guin. I consider it rather as raising the stature of Harold Bloom. He recognized how she transcended genre and belongs alongside (or perhaps, above) writers of highbrow literary fiction.

From a blog post last month by the same author:

> Today, I would say that about 90% of my code is authored by Claude Code. The rest of the time, I’m mostly touching up its work or doing routine tasks that it’s slow at, like refactoring or renaming.

> I see a lot of my fellow developers burying their heads in the sand, refusing to acknowledge the truth in front of their eyes, and it breaks my heart because a lot of us are scared, confused, or uncertain, and not enough of us are talking honestly about it. Maybe it’s because the initial tribal battle lines have clouded everybody’s judgment, or maybe it’s because we inhabit different worlds where the technology is either better or worse (I still don’t think LLMs are great at UI for example), but there’s just a lot of patently unhelpful discourse out there, and I’m tired of it.

https://nolanlawson.com/2026/01/24/ai-tribalism/

If you're responding to this with angry anti-AI rants (or wild AI hype), might want to go read that post.

So many people responding to you with snarky comments or questioning your programming ability. It makes me sad. You shared a personal take (in response to TFA which was also a personal take). There is so much hostility and pessimism directed at engineers who simply say that AI makes them more productive and allows them to accomplish their goals faster.

To the skeptics: by all means, don't use AI if you don't want to; it's your choice, your career, your life. But I am not sure that hitching your identity to hating AI is altogether a good idea. It will make you increasingly bitter as these tools improve further and our industry and the wider world slowly shifts to incorporate them.

Frankly, I consider the mourning of The Craft of Software to be just a little myopic. If there are things to worry about with AI they are bigger things, like widespread shifts in the labor force and economic disruption 10 or 20 years from now, or even the consequences of the current investment bubble popping. And there are bigger potential gains in view as well. I want AI to help us advance the frontiers of science and help us get to cures for more diseases and ameliorate human suffering. If a particular way of working in a particular late-20th and early-21st century profession that I happen to be in goes away but we get to those things, so be it. I enjoy coding. I still do it without AI sometimes. It's a pleasant activity to be good at. But I don't kid myself that my feelings about it are all that important in the grand scheme of things.

Oh sure. We agree on all you said. I wouldn’t call it cheap either. :)

The best you can say is “high cost but positive ROI investment.” Although I don’t think that’s true beyond a certain point either, certainly not outside special cases like small startups with a lot of funding trying to build a product quickly. You can’t just spew tokens about and expect revenue to increase.

That said, I do reserve some special scorn for companies that penny-pinch on AI tooling. Any CTO or CEO who thinks a $200/month Claude Max subscription (or equivalent) for each developer is too much money to spent really needs to rethink their whole model of software ROI and costs. You’re often paying your devs >$100k yr and you won’t pay $2k / yr to make them more productive? I understand there are budget and planning cycle constraints blah blah, but… really?!

I do not really agree with the below, but the logic is probably:

1) Engineering investment at companies generally pays off in multiples of what is spent on engineering time. Say you pay 10 engineers $200k / year each and the features those 10 engineers build grow yearly revenue by $10M. That’s a 4x ROI and clearly a good deal. (Of course, this only applies up to some ceiling; not every company has enough TAM to grow as big as Amazon).

2) Giving engineers near-unlimited access to token usage means they can create even more features, in a way that still produces positive ROI per token. This is the part I disagree with most. It’s complicated. You cannot just ship infinite slop and make money. It glosses over massive complexity in how software is delivered and used.

3) Therefore (so the argument goes) you should not cap tokens and should encourage engineers to use as many as possible.

Like I said, I don’t agree with this argument. But the key thing here is step 1. Engineering time is an investment to grow revenue. If you really could get positive ROI per token in revenue growth, you should buy infinite tokens until you hit the ceiling of your business.

Of course, the real world does not work like this.

It’s a cute aphorism, and useful sometimes, but when you look closely or from different angles hedgehogs often have foxy attributes, or vice versa. Some moments in a person’s life they look more like a fox, other times like a hedgehog. Perhaps the distinction applies better to specific ideas or to pieces of work than as a fixed essence of a person.

Einstein might seem like a quintessential hedgehog (surely the principle of relativity is a Big Hedgehog Idea if ever there was one). Then you learn he once invented a refrigerator. Tolstoy looks like an obvious fox earlier in his writing career, but increasingly a hedgehog towards the end of his life. And slightly less exaltedly, I feel like a fox in some contexts and a hedgehog in others. It might change day to day, or depend on who I’m talking too.

(People are complicated. All aphorisms are wrong, but some are useful I guess. I still quote this one sometimes.)

Thanks for elaborating. So what is the empirically-testable assertion behind this… that an LLM cannot create a (sufficiently complex) system without examples of the source code of similar systems in its training set? That seems empirically testable, although not for compilers without training a whole new model that excludes compiler source code from training. But what other kind of system would count for you?

I wonder too. It might be that progress on the underlying models is going to plateau, or it might be that we haven't yet reached what in retrospect will be the biggest inflection point. Technological developments can seem to make sense in hindsight as a story of continuous progress when the dust has settled and we can write and tell the history, but when you go back and look at the full range of voices in the historical sources you realize just how deeply nothing was clear to anyone at all at the time it was happening because everyone was hurtling into the unknown future with a fog of war in front of them. In 1910 I'd say it would have been perfectly reasonable to predict airplanes would remain a terrifying curiosity reserved for daredevils only (and people did); or conversely, in the 1960s a lot of commentators thought that the future of passenger air travel in the 70s and 80s would be supersonic jets. I keep this in mind and don't really pay too much attention to over-confident predictions about the technological future.

I mostly mentioned Claude Code because it's what Mitchell first tried according to his post, and it's what I personally use. From what I hear Codex is pretty comparable; it has a lot of fans. There are definitely some differences and strengths and weaknesses of both the CLIs and the underlying LLMs that others who use more than one tool might want to weight in on, but they're all fairly comparable. (Although, we'll see how the new models released from Anthropic and OpenAI today stack up.) Codex and Gemini CLI are basically Claude Code clones with different LLMs behind them, after all.

This is such a lovely balanced thoughtful refreshingly hype-free post to read. 2025 really was the year when things shifted and many first-rate developers (often previously AI skeptics, as Mitchell was) found the tools had actually got good enough that they could incorporate AI agents into their workflows.

It's a shame that AI coding tools have become such a polarizing issue among developers. I understand the reasons, but I wish there had been a smoother path to this future. The early LLMs like GPT-3 could sort of code enough for it to look like there was a lot of potential, and so there was a lot of hype to drum up investment and a lot of promises made that weren't really viable with the tech as it was then. This created a large number of AI skeptics (of whom I was one, for a while) and a whole bunch of cynicism and suspicion and resistance amongst a large swathe of developers. But could it have been different? It seems a lot of transformative new tech is fated to evolve this way. Early aircraft were extremely unreliable and dangerous and not yet worthy of the promises being made about them, but eventually with enough evolution and lessons learned we got the Douglas DC-3, and then in the end the 747.

If you're a developer who still doesn't believe that AI tools are useful, I would recommend you go read Mitchell's post, and give Claude Code a trial run like he did. Try and forget about the annoying hype and the vibe-coding influencers and the noise and just treat it like any new tool you might put through its paces. There are many important conversations about AI to be had, it has plenty of downsides, but a proper discussion begins with close engagement with the tools.

Nobody disputes that the LLM was drawing on knowledge in its training data. Obviously it was! But you'll need to be a bit more specific with your critique, because there is a whole spectrum of interpretations, from "it just decompressed fuzzily-stored code verbatim from the internet" (obviously wrong, since the Rust-based C compiler it wrote doesn't exist on the internet) all the way to "it used general knowledge from its training about compiler architecture and x86 and the C language."

Your post is phrased like it's a two sentence slam-dunk refutation of Anthropic's claims. I don't think it is, and I'm not even clear on what you're claiming precisely except that LLMs use knowledge acquired during training, which we all agree on here.

Umm, Sonnet 4.5 has a 1m context window option if you are using it through the api, and it works pretty well. I tend not to reach for it much these days because I prefer Opus 4.5 so much that I don't mind the added pain of clearing context, but it's perfectly usable. I'm very excited I'll get this from Opus now too.

Some of that is true, sure, but nobody who claims LLMs can code and reason about problems is claiming that they operate like humans. Can you give concrete examples of actual specific coding tasks that LLMs can’t do and never will be able to do as a consequence of all that?

I mean, yes and. Companies may do things for broadly marketing reasons, but that can have positive consequences for users and companies can make committed decisions that don't just optimize for short term benefits like revenue or share price. For example, Apple's commitment to user privacy is "just marketing" in a sense, but it does benefit users and they do sacrifice sources of revenue for it and even get into conflicts with governments over the issue.

And company execs can hold strong principles and act to push companies in a certain direction because of them, although they are always acting within a set of constraints and conflicting incentives in the corporate environment and maybe not able to impose their direction as far as they would like. Anthropic's CEO in particular seems unusually thoughtful and principled by the standards of tech companies, although of course as you say even he may be pushed to take money from unsavory sources.

Basically it's complicated. 'Good guys' and 'bad guys' are for Marvel movies. We live in a messy world and nobody is pure and independent once they are enmeshed within a corporate structure (or really, any strong social structure). I think we all know this, I'm not saying you don't! But it's useful to spell it out.

And I agree with you that we shouldn't really trust any corporations. Incentives shift. Leadership changes. Companies get acquired. Look out for yourself and try not to tie yourself too closely to anyone's product or ecosystem if it's not open source.

Similar point applies to being able to use a sextant for celestial navigation in bluewater sailing. GPS is great, until you lose power or your equipment malfunctions. Of course, you can have double or triple redundancy to make such cases vanishingly rare, but still — it’s nice to have a backup that relies on nothing outside your control.

Some of it is pure nostalgia, though, I’ll admit. It a way to honor how people solved similar problems in the past. In the 18th century a sextant plus accurate chronometer or lunar distance table was one of the pivotal technologies of the age; you could use it to pinpoint your location on the boundless ocean within a few miles. That demands respect, and it’s also just really cool it was possible in an era before electricity and radio.

This has been a problem for us too. Sometimes they reach for skills, sometimes they don’t and just try to do the thing on their own. It’s annoying.

I think this is (mostly) a solvable problem. The current generation of SotA models wasn’t RLVR-trained on skills (they didn’t exist at that time) and probably gets slightly confused by the way the little descriptions are all packed into the same tool call schema. (At least that’s how it works with Claude Code.) The next generation will have likely been RLVRed on a lot of tasks where skills are available, and will use them much more reliably. Basically, wait until the next Opus release and you should hopefully see major improvements. (Of course, all this stuff is non-deterministic blah blah, but I think it’s reasonable to expect going from “misses the skill 30% of the time” to “misses it 2% of the time”.)

Yeah. At this point, at the start of 2026, people that are taking these sorts of positions with this sort of tone tend to have their identity wrapped up in wanting AI to fail or go away. That’s not conducive to a reasoned discussion.

There are a whole range of interesting questions here that it’s possible to have a nuanced discussion about, without falling into AI hype and while maintaining a skeptical attitude. But you have to do it from a place of curiosity rather than starting with hatred of the technology and wishing for it to be somehow proved useless and fade away. Because that’s not going to happen now, even if the current investment bubble pops.

There is so much confusion on this topic. Please don't spread more of it; the answers are just a quick google away. To spell it out:

1) AI companies make money on the tokens they sell through their APIs. At my company we run Claude Code by buying Claude Sonnet and Opus tokens from AWS Bedrock. AWS and Anthropic make money on those tokens. The unit economics are very good here; estimates are that Anthropic and OpenAI have a gross margin of 40% on selling tokens.

2) Claude Code subscriptions are probably subsidized somewhat on a per token basis, for strategic reasons (Anthropic wants to capture the market). Although even this is complicated, as the usage distribution is such that Anthropic is making money on some subscribers and then subsidizing the ultra-heavy-usage vibe coders who max out their subscriptions. If they lowered the cap, most people with subscriptions would still not max out and they could start making money, but they'd probably upset a lot of the loudest ultra-heavy-usage influencer-types.

3) The biggest cost AI companies have is training new models. That is the reason AI companies are not net profitable. But that's a completely separate set of questions from what inference costs, which is what matters here.

I agree with you that everything is changing and that we don’t know what’s coming, but I think you really have to stretch things to imagine that it’s a likely scenario that AI-assisted coding will “dry up and blow away.” You’ll need to elaborate on that, because I don’t think it’s likely even if the AI investment bubble pops. Remember that inference is not really that expensive. Or do you think that things shift on the demand side somehow?

Not OP, but I would imagine (or hope) that this attitude is far less common amongst peer CS educators. It is so clear that AI tools will be (and are already) a big part of future jobs for CS majors now, both in industry and academia. The best-positioned students will be the ones who can operate these tools effectively but with a critical mindset, while also being able to do without AI as needed (which of course makes them better at directing AI when they do engage it).

That said I agree with all your points too: some version of this argument will apply to most white collar jobs now. I just think this is less clear to the general population and it’s much more of a touchy emotional subject, in certain circles. Although I suppose there may be a point to be made about being more slightly cautious about introducing AI at the high school level, versus college.

This point about "some people already said six months ago that it was better than it was six months ago" is regularly trotted out in threads like this like it's some sort of trump card that proves AI is just hype. It doesn't make sense to me. What else do you expect people to be saying about a rapidly-improving technology? How does it help you to distinguish technologies that are hype from those that are not?

I'm sure people were saying similar things about, say, aviation all through the first decades of the 20th century, "wow, those planes are getting better every few years"... "Until recently planes were just gimmicks, but now they can fly across the English channel!"... "I wouldn't have got in one of those death traps 5 years ago, but now I might consider it!" And different people were saying things like that at different times, because they had different views of the technology, different definitions of usefulness, different appetites for risk. It's just a wide range of voices talking in similar-sounding terms about a rapidly-developing technology over a span of time.

This is just how people are going to talk about rapidly-improving technologies for which different people have different levels of adoption at different times. It's not a terribly interesting point. You have to engage with the specifics, I'm afraid.

There is a gem of a sci.math thread where Archimedes Plutonium claims to have proved the Riemann Hypothesis and Terrance Tao (before he was quite as famous, but still!) replies to him pointing out his errors.

Also TIL Archimedes Plutonium was actually his legal name.

> Having been born in Königsberg in 1724, he never left the small German city, dying there in 1804 aged 79 never having once gone further than the city’s limits.

Totally false! Slander! He once went as far as the village of Jarnołtowo, a whole 60 miles from Königsberg![0]

But yeah. Maybe not one of history’s greatest travelers.

[0] https://en.wikipedia.org/wiki/Jarno%C5%82towo

You’re right to call out consumer use is what’s eating all the tokens. I suppose what was behind my way of putting it is: I haven’t seen much in the way of truly transformative products with LLMs in the consumer space. Sure, there’s a few power users doing some cool things, and lots of promises along the lines of “AI will plan and book you a whole vacation!”, but basically for the median consumer we have an “improved search”, and “fun image generation”, with people using it a couple times a week. So what is the product that changes the world and makes 500M people pay $100 a month? I don’t think it’s really here yet. This feels a bit like we’re in 1996 where people are still trying to figure out what the internet is for, and we don’t know if OpenAI will be 1996’s Amazon, or its AOL.

None of this is to deny how remarkable the underling LLM tech is. I never expected to see something like this in my lifetime; it feels far far more strange and new than when the iPhone came along. I use a coding agent daily and it’s dramatically changed how I work. But I still think we’re in a bubble here.

All of these things can be simultaneously true (and I would say, are true):

1) We are in a huge investment bubble right now and it's going to burst.

2) LLMs are extremely useful right now for certain niche tasks, especially software engineering.

3) LLMs have the potential to transform our world long-term (~10 yr horizon), on the order of the transformations wrought by the internet and mobile.

4) LLM's don't lead directly to AGI (no continuous learning), and we're not getting AGI any time soon.

This is an extremely obvious point, but bears repeating. I feel the assumption of an implicit link (in both truth or falsehood) between these fairly independent assertions can cause people to talk past each other about the really important questions in play here.

Regarding The Great Bubble, I am very very bearish about OpenAI in particular. They've had a good run for three years with consumer mindshare due to their first-mover advantage, but they have no moat, trouble monetizing most of their users, not much luck building out products that stick among consumers that aren't chatbots, and their models are no better than Anthropic's, Google's, or even the best Chinese open weight models 6 months later.

My bet would be on Google and Apple together (with Gemini powering Siri, for now) destroying OpenAI in the consumer AI market over the next 2-3 years. Google has first-rate models... but more than that, both Google and Apple have the enormous advantage of owning underlying platforms that they can use to put their own AI chat in front of consumers. Google has a mobile OS, the leading browser, and search. Apple has the premium hardware and the other, premium, mobile OS. They also have the advantage of the current regulatory climate being less antitrust than it was. And they don't have to monetize their AI offerings (no ads in gemini; ChatGPT is adding them) and can run them at a loss for as long as it takes to eat up OpenAI's market share. If they partner up, as they seem to be doing, OpenAI should very very afraid.

In fact it peaked in size in 1920. But you’re broadly correct about the sense of decline post-WWI and the sense that America would be the dominant 20th-century power.

It’s funny, from his books I always imagined he spent years up there in the Yukon. But it turns out he trekked in (by far the hardest experience of the Yukon Gold Rush rush was just getting there with the mandated 1 year of food and all your mining equipment), staked an unprofitable claim, talked to a lot of people in bars in Dawson City, spent an uncomfortable winter in a small cabin with some other gold-rushers eating just bread and beans and bacon, got scurvy from eating just bread and beans and bacon, and then got the hell out and went back to San Francisco.

I say this not to minimize the depth or the hardship of his experience (it sounds like a nightmare) but more in amazement at all the compressed experiences he had and the folder for stories he amassed during that one year. Certain years in life flash by (or they seem that way to me) and others are formative and seem to last forever. Clearly this was the latter for him.