Hacker News

AI coding is gambling on slot machines, managing developers is betting on race horses.

I think the addiction angle seems to make AI coding more similar to gambling. Some people seem to be disturbingly addicted to agentic coding. Much more so than traditional programming. To the point of doing destructive things like waking up in the middle of the night to check agents. Or giving an agent access to their bank account.

You (in theory) have more control over the quality of the team you are managing, than the quality of the models you are using.

And the quality of code models puts out is, in general, well below the average output of a professional developer.

It is however much faster, which makes the gambling loop feel better. Buying and holding a stock for a few months doesn't feel the same as playing a slot machine.

Framing anything with a common blanket concepts usually fails to apply the same framing to related areas. A lot of things include some gambling, you need to compare how it was before was 'gambling', and how 'not using AI' is also 'gambling', etc.

As @m00x points out "coding is gambling on slot machines, managing developers is betting on race horses."

I ssk an AI to play hangman with me and looked at it's reasoning. It didn't just pick a secret word and play a straightforward game of hangman. It continually adjusted the secret word based on the letters I guessed, providing me the "perfect" game of hangman. Not too many of my guesses were "right" and not too many "wrong" and I after a little struggle and almost losing, I won in the end.

It wasn't a real game of hangman, it was flat out manipulation, engagement farming. Do you think it's possible that AI does that in any other situations?

I don‘t think so. A project manager can give feedback, train their staff, etc. An AI coding model is all you get, and you have to wait until your provider trains a new model before you might see an improvement.

This must be it. So many of our colleagues have been burnt by bad coworkers that they would rather burn everything down than spend another day working with them.

> AI models are non-deterministic, and humans are also non-deterministic. You could assign the same task to two different developers and end up with entirely different results.

Except, one can explain themselves (humans) and their actions can be held to account in the case of any legal issue whereas an AI cannot; making such an entity completely unsuitable for high risk situations.

This typical AI booster comparison has got to stop.

As a human, you generally have the opportunity make decent headway in understanding the other humans that you're working with and adjusting your instructions to better anticipate the outputs that they'll return to you. This is almost impossible with AI because of a combination of several factors:

>You are not an AI and do not know how an AI "thinks".

>Even if you come to be able to anticipate an AI's output, you will be undermined by the constant and uncontrollable update schedule imposed on you by AI platforms. Humans only make drastic changes like this under uncommon circumstances, like when they're going through large changes in their life, not as a matter of course.

>However, without this update schedule, problems that were once intractable will likely stay so forever. Humans, on the other hand, can grow without becoming completely unpredictable.

It's a Catch-22. AI is way closer to gambling.

One size never fits all. I am old enough to remember what a game changer Spreadsheets (VisiCalc) where. They made the personal computer into a SwissArmy knife for many people that could not justify investing large sums of money into software to solve a niche problem. Until that time PCs simply were not a big thing.

I believe AI will do something similar for programming. The level of complexity in modern apps is high and requires the use of many technologies that most of us cannot remotely claim to be expert in. Getting an idea and getting a prototype will definitely be easier. Production Code is another beast. Dealing with legacy systems etc will still require experts at least for the near future IMHO.

> but I finally feel like I'm _good_ at programming, which is insane, because I literally haven't written a line of code myself in months

This is exactly the sort of mentality that makes me hate this technology

You finally feel good at programming despite admitting that you aren't actually doing it

Please explain why anyone should take this seriously?

I really hope that was your creativity and not AI

> When you’re staring at the diff view,

Bold assumption that people are looking at the diffs at all. They leave that for their coworkers agents.

You're a gamblin' man, I see...

Good sir, have you heard the Good Word of the Waterfall development process? It sounds like that's what you are describing

> properly defining the spec

Why do you often need to re-prompt things like "can you simplify this and make it more human readable without sacrificing performance?". No amount of specification addresses this on the first shot unless you already know the exact implementation details in which case you might as well write it yourself directly.

I often have to put in a prompt like this 5-10 times before the code resembles something I'd even consider using as a 1st draft base to refactor into something I would consider worthy of being git commit.

I sometimes use AI for tiny standalone functions or scripts so we're not talking about a lot of deeply nested complexity here.

That can't be the whole story, right? Because there are an arbitrarily large number of (e.g.) Rust programs that will implement any given spec given in terms of unit tests, types, and perhaps some performance benchmarks.

But even accounting for all these "hard" constraints and metrics, there are clearly reasons to prefer some possible programs over others even when they all satisfy the same constraints and perform equally on all relevant metrics.

We do treat programs as efficient causes[1] of side effects in computing systems: a file is written, a block of memory is updated, etc. and the program is the cause of this.

But we also treat them as statements of a theory of the problem being solved[2]. And this latter treatment is often more important socially and economically. It is irrational to be indifferent to the theory of the problem the program expresses.

[1]: https://en.wikipedia.org/wiki/Four_causes#Efficient

[2]: https://pages.cs.wisc.edu/~remzi/Naur.pdf

Then pulling the lever until it works! You can also code up a little helper to continuously pull the lever until it works!

Well it’s more how much we care about those.

Which with the advent of LLMs just lowered our standards so we can claim success.

AI: "Yes, the specs are perfectly clear and architectural standards are fully respected."

[Imports the completely fabricated library docker_quantum_telepathy.js and calls the resolve_all_bugs_and_make_coffee() method, magically compiling the code on an unplugged Raspberry Pi]

AI: "Done! The production deployment was successful, zero errors in the logs, and the app works flawlessly on the first try!"

I had a CIO tell me 15 years ago with Agile I was wasting my time with specs and design documents.

That was always the easy part.

The endless next steps of "and add this feature" or "this part needs to work differently" or "this seems like a bug?" or "we must speed up this part!" is where 98% of the effort always was.

Is it different with AI coding?

Personally, I get a huge rush of dopamine seeing LLMs build out complex features very quickly to the point that it will keep me up all night wanting to push further and further.

That's where the gambling metaphor really resonates. It's not whether or not the output is correct, I've been building software for many years and I know how direct LLMs pretty well at this point. But I'm also an alcoholic in recovery and I know that my brain is wired differently than most. And using LLMs has tested my ability to self-regulate in ways that I haven't dealt with since I deleted social media years ago.

Poker is a skill-based game where your actions influence your success, but many people who play it are gambling.

One way it works is if you think of cognitive debt as the "house". As in "the house always wins".

Yeah, I don't think the metaphor applies exactly but I definitely see similarities from my personal experience

1/ Dependency -- Once I got used to agentic coding, I almost always reached out to it even for small changes (e.g. update a yaml config)

2/ Addiction -- In the initial euphoria phase, many people experience not wanting to "waste" any time agent idle and they'd try to assign AI agents task before they go to sleep

3/ You trust your judgement less and less as agent takes over your code

4/ "Slot machine" behavior -- running multiple AI agents parallel on same task in hope of getting some valuable insight from either

5/ Psychosis -- We have all met crypto traders who'd tell you how your 9-5 is stupid and you could be making so much trading NFTs. Social media if full of similar anecodotes these days in regards to vibecoding with people boasting their Claude spend, LOC and what not

Slot machines have very controlled results. They are regulated to a high precision of reliability.

> The gambling metaphor often applied to vibecoding implies that the outcome

The important part of the not-really-a-metaphor is the relationship between user and machine, and how it affects the user's mind.

What the machine outputs on "wins" doesn't matter as much, addictive gambling can still happen even when the payouts are dumb.

it can give better results more consistently with the proper checks and balances.

You can get more consistent results from a slot machine with a bunch of magnets and some swift kicks. It's still gambling.

Exactly. It's not gambling if you win most of the time. This is like saying driving a car is gambling. I mean sure, I guess if you think any amount of risk equals gambling.

I don't know where I'd draw the line personally, but wherever you draw it there's a problem. If you give increasingly more advanced tasks to it, you will eventually cross the line.

we're winning so much we started complaining "I can't handle so much winning"

It’s just how discussion on the internet works, for basically anything at all worth discussing. It’s exhausting, but I can hardly blame HN specifically.

> Everybody pretty much agrees that it works for some use cases

That isn't true, which is the exact reason why people have a binary mindset. More than once on Hacker News I've had people accuse me of being an AI booster just because I said I had success with agents and they did not.

VIM vs Emacs vs IDE vs..., Tabs vs Spaces, Procedural vs OOP vs Functional.

We love a good holy war for sure.

The nuance is lost, and the conversations we should be having never happen (requirements, hiring/skills, developer experience).

I like the analogy but which 2 is AI coding?

Fast & Cheap (but not Good?) - I wouldn't really say that AI coding is "cheap"

Cheap & Good (but not Fast) - Again, not really "cheap"

Fast & Good (but not Cheap) - This seems like maybe where we're at? Is this a bad place?

> The cost of AI is low

If we're only talking about money spent on prompting AI, maybe. The damage to online trust is massive imo. So is the damage done by looting the commons to build them.

Typical privatize the profits socialize the costs bullshit

I doubt gambling is in nature. Investments based on reason pay off. Evolution shapes for sensical moves.

Humans invented gambling as a rigged game that mimics what's in nature, perversed for profit.

The first anthropomorphization of AI which is actually useful.

My understanding is that waymo has gone on the record to say that they have human operators that remotely drive the vehicle in scenarios where their automated system is confused.

Which I assert is semantically equivalent to saying: Human drivers (even when operating at the diminished capacity of not even being present in the car) are less likely to make errors driving a car than AIs.

If I said I had a machine where I put in "tokens", watch it spin, and either get nothing or something valuable (with which I get being largely chance), you'd presume it's some kind of slot machine. The important things IMO are the random chance of getting something and being able to keep retrying so rapidly.

You can't keep paying to play the "refinancing game" until you get a good rate (at least not like pulling the lever again and again, you have to wait a long time, you won't call the same bank again and again, and suddenly they have an amazing rate), it's a different experience and the psychology is different.

That’s very true. But interns aren’t supposed to be doing useful work. The purpose of interns is training interns and identifying people who might become useful at a later date.

I’ve never worked anywhere where the interns had net productivity on average.

1. Interns learn. LLMs only get better when a new model comes out, which will happen (or not) regardless of whether you use them now.

2. Who here thinks that having interns write all/almost all of your code and moving all your mid level and senior developers to exclusively reviewing their work and managing them is a good idea?

There are two major mistakes here.

The first is equating human and LLM intelligence. Note that I am not saying that humans are smarter than LLMs. But I do believe that LLMs represent an alien intelligence with a linguistic layer that obscures the differences. The thought processes are very different. At top AI firms, they have the equivalent of Asimov's Susan Calvin trying to understand how these programs think, because it does not resemble human cognition despite the similar outputs.

The second and more important is the feedback loop. What makes gambling gambling is you can smash that lever over and over again and immediately learn if you lost or got a jackpot. The slowness and imprecision of human communication creates a totally different dynamic.

To reiterate, I am not saying interns are superior to LLMs. I'm just saying they are fundamentally different.

And, if we're being honest, the way people talk about interns is weirdly dehumanizing, and the fact that they are always trotted out in these AI debates is depressing.

exactly where my mind went as well. There aren't really levels to pulling a lever on a slot machine, other than the ability for each pull to result in more "plays" of the same potential outcome.

The reason i think this metaphor keeps popping up, is because of how easy it is to just hit a wall and constantly prompt "its not working please fix it" and sometimes that will actually result in a positive outcome. So you can choose to gamble very easily, and receive the gambling feedback very quickly unlike with an intern where the feedback loop is considerably delayed, and the delayed interns output might simply be them screaming that they don't understand.

The only similarity is that they both say "you’re absolutely right" when you point out their obvious mistakes

Drawing parallels between AI and interns just shows you're a misanthrope

You should value assigning tasks to human interns more than AI because they are human

You generally don’t assign work to an intern just for the output, though.

An intern can be taught. If you try to 'teach' a craps table, they'll drag you out of the casino.

As someone who has worked with interns for year, expect feedback and reiterations always, be surprised if they get it the first time... which merits feedback as well!

But looks like the intern mafia is bombarding you with downvotes.

Addiction for the mere purpose of satisfying a compulsion, rather than to achieve a reward or physical "high."

> it's not gambling if the most expected outcome actually occurs.

> you have to spend a lot of tokens to keep the error/hallucination rate low

Ironically, I find your comment more effective at convincing me AI coding is gambling than the original article. You're talking about it the exact same way that gamblers do about their games.

You remind me of those guys who swear they have a "system" at the casino.

> If, on the other hand, you treat it like a hyper-competent collaborator, and follow good project management and development practices, you're golden.

I am consistently using 100% of my weekly $200 max plan. I know how this thing works, I know how to get value out of it, and I wish what you said were true.

If you do all of these things? You are in a better spot. You are in a far better spot than if you hadn't! Setting up hooks to ensure notes get written? Massive win! Red-green TDD? Yes, please! But in terms of just ... well, being able to rely on the damn thing?

https://github.com/ctoth/claude-failures

_hyper-competent collaborator who may completely make things up occasionally and will sometimes give different answers to the same question*_

Life is full of variable reward schemes. Probably why we evolved to be so enamoured by them.

In a healthy environment. We are harmed more by being totally risk adverse. Than by accepting risk as part of life and work.

> literally forbidden to touch a single line if code.

That is extremely stupid. What does that ban get you? I reqct to this because a friend mentioned exactly this. And I was dumbfounded.