Everyone in Seattle hates AI

I wonder if these feelings are what scribes and amanuenses felt when the printing press arrived.

I do enjoy programming, I like my job and take pride on it, but I actively try for it not to be the life-mean giving activity. I'm a just mercenary of my trade.

And I read a lot of articles about games that seem to love throwing a dig at AI even if it's not really relevant.

Personally, I can see why people dislike Gen AI. It takes people's creations without permission.

That being said, morality of the creation of AI tooling aside, there are still people who dislike AI-generated stuff. Like, they'd enjoy a song, or an image, or a book, and then suddenly when they find out it's AI suddenly they hate it. In my experience with playing with comfy ui to generate images, it's really easy to get something half decent, it's really hard to get something very high quality. It really is a skill in itself, but people who hate AI think it's just type a prompt and get image. I've seen workflows with 80+ nodes, multiple prompts, multiple masks, multiple loras, to generate one single image. It's a complex tool to learn, just like photoshop. Sure you can use Nano-Banana to get something but even then it can take dozens of generations and prompt iterations to get what you want.

[0] https://www.theverge.com/entertainment/827650/indie-develope...

That's probably me for a lot of people. The reality is a bit finer than this namely :

- I hate VC funded AI which is actually super shallow (basically OpenAI/Claude wrappers)

- I hate VC funded genuine BigAI that sells itself as the literal opposite of what it is, e.g. OpenAI... being NOT open.

- I hate AI that hides it's ecological cost. Generating text, videos, etc is actually fascinating, but not if making the shittiest video with the dumbest script is taking the same amount of energy I'd need to fly across the globe.

- I hate AI that hides it's human cost, namely using cheap labor from "far away" where people have to label atrocities (murders, rape, child abuse, etc) without being provided proper psychological support.

- I hate AI that embodies capitalist principles of exploitation. If somehow your entire AI business relies on an entire pyramid of everything listed above to capture a market then hike the price once dependency is entrenched you might be a brilliant business man but you suck as a human being.

etc... I could go on but you get the idea.

I do love open source public AI research though. Several of my very good friends are researchers in universities working on the topic. They are smart, kind and just great human beings. Not fucking ghouls riding the hype with 0 concern for our World.

So... yes maybe AI haters have a slightly more refined perspective but of course when one summarize whatever text they see in 3 words via their favorite LLM, it's hard to see.

AGI? No, although it's not there. LLMs? Yes, lots. The main benefit they can give is to sort-of-speed-up internet search, but I have to go and check the sources anyway so I'll revert back to 20+ years of experience of doing it myself. Any other application of machine learning such almost instant speech to text? No, it's useful.

In real life, I don't know anyone who genuinely wants to use AI. Most of them think it's "meh", but don't have any strong feelings about using it if it's more convenient - like Google shoving it in their face during a search. But would they pay for it, or miss it if it's gone? Nope, not a chance.

Is the siren song of "AI effect" so strong in your mind that you look at a system that writes short stories, solves advanced math problems and writes working code, and then immediately pronounce it "not intelligent"?

- At it's inception in 1955 it was "learning or any other feature of intelligence" simulated by a machine [1] (fun fact: both neural networks and computers using natural language were on the agenda back then)

- Following from that we have the "all machine learning is AI" which was the prevalent definition about a decade ago

- Then there's the academic definition that is roughly "computers acting in real or simulated environments" and includes such mundane and algorithmic things as path finding

- Then there's obviously AGI, or the closely related Hollywood/SciFi definition of AI

- Then there's just "things that the general public doesn't expect computers to be able to do". Back when chess computers used to be called AI this was probably the closest definition that fits. Clever sales people also used to love to call prediction via simple linear regression AI

Notably four out of five of them don't involve computers actually being intelligent. And just a couple years ago we still sold simple face detection as AI

1: https://www-formal.stanford.edu/jmc/history/dartmouth/dartmo...

Many people claim it's doing great because they have driven hundreds of kilometers, but don't particularly care whether they arrived at the exact place, and are happy with the approximate destination.

That's a great way to describe it. There's a good article that points out AI is the new aesthetic of fascism. And, of course, in Miyazaki's words, "I strongly feel that this is an insult to life itself."

It was already tough to run flagship-class local models and it's only getting worse with the demand for datacenter-scale compute from those specific big players. What happens when the model that works best needs 1TB of HBM and specialized TPUs?

AI computation looks a lot like early Bitcoin: first the CPU, then the GPUs, then the ASICs, then the ASICs mostly being made specifically by syndicates for syndicates. We are speedrunning the same centralization.

If it's ever to be economically viable to run a model like this, you basically need to run it non-stop, and make money doing so non-stop in order to offset the hardware costs.

Creating a search engine index requires several orders of magnitude less computing power then creating the weights of an LLM model. Like it is theoretically possible for somebody with a lot of money to spare to create a new search index, but only the richest of the rich can do that with an LLM model.

And search engines are there to fulfill exactly one technical niche, albeit an important one. LLMs are stuffed into everything, whether you like it or not. Like if you want to use Zoom, you are not told to “enrich your experience with web search”, you are told, “here is an AI summary of your conversation”.

The hardest part with that IMO will be democratizing the hardware so that everybody can afford it.

And he's right. LLMs are fancy text query engines and work very well as such.

The problem is when people try to shoehorn everything into LLMs. That's a disaster yet being perused vigorously by some.

AI meeting notes are great! After you spend twice as long editing out the errors, figuring out which of the two Daves was talking each time, and removing all the unimportant side-items that were captured in the same level of detail as the core decision.

AI summaries are great - if you're the sort of person that would use a calculator that's wrong 10% of the time. The rest of us realize that an hour spent reading something is more rewarding and useful than an hour spent double checking an AI summary for accuracy.

AI as Asbestos isn't even an apt comparison, both are toxic and insidious, but Asbestos at least had a clear and compelling use case at the time. It solved some problems both better and cheaper than available alternatives. AI solves problems poorly and at higher cost, and people call you "threatened" if you point that out.

About the only problem here is the increase of surveillance and you can avoid that by running your own models, which are getting better and better by the day. The fact that people are so willing to accept these criticisms without much scrutiny is really just indicative of prior bias

AI is just not that good. If it really made me more productive, why wouldn't I use it all the time? I'd get everything done before lunch and go home. Or I'd use it all day to do the work of 3 people and be on the fast track to promotions.

The problem is simply that it gets in the way. For things I know nothing about, AI is excellent. For things that I'm good at and have literally been doing for a decade+, I can just do it better and faster myself, and I'm tired of people who know nothing about my profession gaslighting me into thinking that LLMs do the same thing. And I'm really tired of people saying "oh AI is not good today, but it'll be good tomorrow so just start using them" -- fine, wake me up when it's good because I've been waiting and patiently testing every new SOTA model since 2023.

Just get the facts right, that's all I ask of tech execs. Why has AI become a religion?

You're absolutely right, this is another face of this AI coin... We people are taught to do things and love doing them and we're scared it's going to be taken away from us. This is what I thought when writing about the man who cancelled CS course. He apparently predicted that learning algorithms solving won't make him happy because AI will do it for him

The main point is that the tools need to be of a certain quality/maturity for dogfooding to be effective.

Being forced to use a shit tool because <some other department somewhere in the company wants your feedback>, while your deadlines haven't been adjusted for all this wasted time is not acceptable behaviour. It's the kind of authoritarian horseshit that's that's so often pushed by unproductive parasites onto people who do actual work.

As J. R. "Bob" Dobbs once said, "I don't practice what I preach because I'm not the kind of person I'm preaching to." ( see https://en.wikiquote.org/wiki/J._R._%22Bob%22_Dobbs )

Maybe the engineers complaining about dogfooding vibe-coding tools aren't the kind of developers you should have vibe-coding.

Regarding dog fooding, Project Reunion was also a victim of all engines AI, now the damage is done and only the Windows team cares, because their job depends on using it.

Funnily, Apple also has an unspoken "deal" (pay a bit low but treat really well) and they stuck to it even through the layoff era.

Thankfully I am technology mercenary, polyglot, and use whatever the clients need, regardless of my point of view, but it is sad to see the human part behind those decisions being affected.

In those cases the actual "new" technology (ie, not the underlying ai necessarily) is not as substantive and novel (to me at least) as a product whose internals are not just an (existing) llm.

(And I do want to clarify that, to me personally, this tendency towards 'thin-shell' products is kind of an inherent flaw with the current state of ai. Having a very flexible llm with broad applications means that you can just put Chatgpt in a lot of stuff and have it more or less work. With the caveat that what you get is rarely a better UX than what you'd get if you'd just prompted an llm yourself.

When someone isn't using llms, in my experience you get more bespoke engineering. The results might not be better than an llm, but obviously that bespoke code is much more interesting to me as a fellow programmer)

LLMs are not an intelligence, and people who treat them as if they are infallible Oracles of wisdom are responsible for a lot of this fatigue with AI

I've tried implementing features with Claude Code Max and if I had let that go on for a week instead of just a couple of days I would've lost a week's worth of work (it was pretty immediately obvious that it was too slow at doing pretty much everything, and even the slightest interaction with the LLM caused very long round-trips that would add additional time, over and over and over again). It's possible people simply don't do the kind of things I do. On the extreme end of that, had I spent my days making CRUD apps I probably would've thought it was magic and a "game changer"... But I don't.

I actually don't have a problem believing that there are people who basically only need to write 25% of their code now; if all you're doing for work is gluing together libraries and writing boilerplate then of course an LLM is going to help with that, you're probably the 1000th person that day to ask for the same thing.

The one part I would say LLMs seem to help me with is medium-depth questions about DirectX12. Not really how to use it, but parts of the API itself. MSDN is good for learning about it, but I would concede that LLMs have been useful for just getting more composite knowledge of DX12.

P.S.:

I have found that very short completions, 1-3 lines, is a lot more productive for me personally than any kind of "generate this feature", or even function-sized generation. The reason is likely that LLMs just suck at the things I do, but they can figure out that a pattern exists in the pretty immediate context and just spit out that pattern with some context clues nearby. That remains my best experience with any and all LLM-assisted coding. I don't use it often because we don't allow LLMs for work, but I have a keybind for querying for a completion when I do side projects.

My company is basically writing blank cheques for "AI" (aka LLM, I hate the way we've poisoned AI as a term))tooling so that people can use any and all tooling they want and see what works and doesn't. This is a company with ~1500ish engineers, ranging from hardware engineers building POS devices to the junior frontenders building out our simplest UIs. There's also a whole lot more people who aren't technical, and they're also encouraged to use any and all AI tooling they can.

Despite the entire company trying to figure out how to use these effectively precisely because we're trying to look at things objectively and separate out the hype from the reality, the only people I've seen with any kind of praise so far (and this has been going on since the early ChatGPT days) have been people in Marketing and Sales, because for them it doesn't matter if the AI hallucinates some pure bullshit since that's 90% of their job anyway.

We have spent god knows how much time and resources trying to get these tools doing anything more useful than simple demos that get thrown out immediately, and it's just not there. No one is pushing 100x the code or features they were before, projects aren't finishing any faster than they were before, and nobody even bothers turning on the meeting transcription tools either anymore because more often than not it'll interpret things said in the meeting just plain wrong or even make up entire discussion points that were never had.

Just recently, like last week recently, we had some idiotic PR review bot from coderabbit or some other such company be activated. I've never seen so many people complain all at once on Slack, there was a thread with hundreds of individuals all saying how garbage it was and how much it was distracting from reviews. I didn't see a single person say they liked the tool, not 1 single person had anything good to say about it.

So as far as I'm concerned, it's just a MASSIVE fucking hype bubble that will ultimately spawn some tooling that is sorta useful for generating unimportant scripts, but little else.

If you haven’t had a mind blown moment with AI yet, you aren’t doing it right or are anchoring in what you know vs discovering new tech.

I’m not making any case for anything, but it’s just not that hard to get excited for something that sure does seem like magic sometimes.

Edit: lol this forum :)

There are many valid critiques of AI, but “there’s not much there” isn’t one of them.

To me, any software engineer who tries an LLM, shrugs and says “huh, that’s interesting” and then “gets back to work” is completely failing at their actual job, which is using technology to solve problems. Maybe AI isn’t the right tool for the job, but that kind of shallow dismissal indicates a closed mind, or perhaps a fear-based reaction. Either way, the market is going to punish them accordingly.

There's a lot of disconnected-from-reality hustling (a.k.a lying) going on. For instance, that's practically Elon Musk's entire job, when he's actually doing it. A lot of people see those examples, think it's normal, and emulate it. There are a lot of unearned superlatives getting thrown around automatically to describe tech.

"Engineers don't try" doesn’t refer to trying out AI in the article. It refers to trying to do something constructive and useful outside the usual corporate churn, but having given up on that because management is single-mindedly focused on AI.

One way to summarize the article is: The AI engineers are doing hype-driven AI stuff, and the other engineers have lost all ambition for anything else, because AI is the only thing that gets attention and helps the career; and they hate it.

You touched on some of the reasons; it doesn't take much skill to call an API, the technology is in a period of rapid evolution, etc.

And now with almost every company trying to adopt "AI" there is no shortage of people who can put AI experience on their resume and make a genuine case for it.

This right here is the real thing which AI is deployed to upset.

The Enlightenment values which brought us the Industrial Revolution imply that the disparity between the people-persuasion-plane and the real-world-facts-plane should naturally decrease.

The implicit expectation here is that as civilization as a whole learns more about how the universe works, people would naturally become more rational, and thus more persuadable by reality-compliant arguments and less persuadable by reality-denying ones.

That's... not really what I've been seeing. That's not really what most of us have been seeing. Like, devastatingly not so.

My guess is that something became... saturated? I'd place it sometime around the 1970s, same time Bretton Woods ended, and the productivity/wages gap began to grow. Something pertaining to the shared-culture-plane. Maybe there's only so much "informed" people can become before some sort of phase shift occurs and the driving force behind decisions becomes some vague, ethically unaccountable ingroup intuition ("vibes", yo), rather than the kind of explicit, systematic reasoning which actually is available to any human, except for the weird fact how nobody seems to trust it very much any more.

The whole cryptocurrency world requires evangelical buy-in. But there is no directly created functional value other than a historic record of transactions and hypothetical decentralization. It doesn’t directly create value. It’s a store of it - again, assuming enough people continue to buy into the narrative so that it doesn’t dramatically deflate when you need to recover your assets. States and other investors are helping make stability happen to maintain it as a value store, but you require the story propagating to achieve those ends.

The values of bitcoin are:

- easy access to trading for everyone, without institutional or national barriers

- high leverage to effectively easily borrow a lot of money to trade with

- new derivative products that streamline the process and make speculation easier than ever

The blockchain plays very little part in this. If anything it makes borrowing harder.

Bitcoin is probably unkillable. Even if were to crash, it won't be hard to round up enough true believers to boost it up again. But it's technically stagnant.

We're conditioned to do so, in large part because this kind of work ethic makes exploitation easier. Doesn't mean that's our natural state, or a desirable one for that matter.

"AI-based economy" is too broad a brush to be painting with. From the Marxist perspective, the question you should be asking is: who owns the robots? and who owns the wealth that they generate?

These aren't meant to be gotcha rhetorical questions, just parts of my professional life where AI _isn't_ desirable by those in power, even if they're some of the only real world use cases where I'd want to use it. As someone said upthread, I want AI to do my dishes and laundry so I can focus on leisure and creative pursuits (or, in my job, writing code). I don't want AI doing creative stuff for me so I can do dishes and laundry.

I have mostly seen people on HN criticizing the few people in tech who have attached themselves to the hype and senselessly push it everywhere, not "the masses." The masses don't particularly like AI. It seems like it's only people hyping it that think everyone but Luddites are into it.

You're both painting a narrative that anti-AI sentiment is a popular bandwagon everyone is doing to be cool, as well as not that big actually and everyone is loving AI. Which is it?

As for the other points, are the LLMs wrong sometimes, yes. But so are humans so it's not really a novel thing to point out. The question is, are they more correct than humans? I have seen they can be more accurate, less biased, etc... and we are driving toward higher accuracy and other ways to make them right.

And the fix later attitude is not great toward everything and I was referring to the accuracy issues that people often point out as why AI is hype. The things you mention are side effects and those should be controlled because the cat is out of the bag. You can spend your time yelling at the clouds or try to do something to make it better. I assure you, capitalism is a tough enemy. This is no different than another type of combustable engine that was created that has negative consequences on the environment in different ways.

I'm not disagreeing with you... mostly just saying: the hype is warranted

I think our tooling is holding us back more than the actual models, and even if they never advance at all from here (unlikely), we'll still get years of improvement and innovation.

In English, I’d pronounce it very similar to “wonderful”.

It makes a lot of sense that someone casually coming in to use chatgpt for 30 minutes a week doesn't have any reason to think more deeply about what using that tool 'means' or where it came from. Honestly, they shouldn't have to think about it.

Among actual people (i.e. not managers) there seems to be a bit of a generation gap - my younger friends (Gen Z) are almost disturbingly enthusiastic about entrusting their every thought and action to ChatGPT; my older friends (young millennials and up) find it odious.

There's maybe a slow trend towards "that's not true, you should know better than to trust AI for that sort of question" in discussions when someone says something like "I asked AI how [xyz was done]" but it's definitely not enough yet to keep anyone from going to it as their first option for answering a question.

At least, that's my wife's experience working on a contract with a state government at a big tech vendor.

I think actually a lot about it is the sort of crass, unthinking, default-American-college-student manner about the way ChatGPT speaks. It's so American and we can feel it. But AI generated art and music is hugely unpopular, AI chatbots replacing real customer service is something we loathe.

Generally speaking I would say that AI feels like something that is being done to us by a handful of powerful Americans we profoundly distrust (and for good reason: they are untrustworthy and we can see through their bullshit).

I can tell you that this is so different to the way the internet was initially received even by older people. But again, perhaps this is in part due to our changing perspectives on America. It felt like an exciting thing to be part of, and it helped in the media that the Web was the brainchild of a British person (even if twenty years later that same media would have liked to pretend he wasn't at a European research institution when he did it).

The feeling about AI is more like the feeling we have about what the internet eventually did to our culture: destroying our high streets. We know what is coming will not be good for what makes us us.

I can assure you, living in Seattle I still encounter a lot a AI boosters just as much as I encounter AI haters/skeptics

The world is not at all about fairness of benefits and impacts to all people it is about a populist mass and what amuses them and makes their life convenient, hopefully without attending the relevant funerals themselves.

Those slops already existed, but AI scales them by an order of magnitude.

I guess the same can be said of any technology, but AI is just a more powerful tool overall. Using languages as an example - lets say duck typing allowed a 10% productivity boost, but also introduced 5% more mistakes/problems. AI (claims to) allow a 10x productivity boost, but also ~10x mistakes/problems.

So between the debugging and many times you need to reprompt and redo (if you bother at all, but then that adds debugging time) is any time actually saved?

I think the dust hasn’t settled yet because no one has shipped mostly AI generated code for a non-trivial application. They couldn’t have with its current state. So it’s still unknown whether building on incredibly shaky ground will actually work in real life (I personally doubt it).

pretty sure the process I've seen most places is more like: one junior approves, one senior approves, then the owner manually merges.

so your process seems inadequate to me, agents or not.

also, was it tagged as generated? that seems like an obvious safety feature. As a junior, I might be thinking: 'my senior colleague sure knows lots of this stuff', but all it would take to dispel my illusion is an agent tag on the PR.

As you're also saying this is the worst it will ever be. There is only one direction, the question is the acceleration/velocity.

Where I'm not sure I agree is with the perception this automatically means we're all going to be out of a job. It's possible there would be more software engineering jobs. It's not clear. Someone still has to catch the bad approaches, the big mistakes, etc. There is going to be a lot more software produced with these tools than ever.

In general I think -> if I was not personally invested in this situation (i.e. another man on the street) what would be my immediate reaction to this? Would I still become a software engineer as an example? Even if it is doesn't come to past, given what I know now, would I take that bet with my life/career?

I think if people were honest with themselves sadly the answer for many would probably be "no". Most other professions wouldn't do this to themselves either; SWE is quite unique in this regard.

I'm extremely skeptical of the argument that this will end up creating jobs just like other technological advances did. I'm sure that will happen around the edges, but this is the first time thinking itself is being commodified, even if it's rudimentary in its current state. It feels very different from automating physical labor: most folks don't dream of working on an assembly line. But I'm not sure what's left if white collar work and creative work are automated en masse for "efficiency's" sake. Most folks like feeling like they're contributing towards something, despite some people who would rather do nothing.

To me it is clear that this is going to have negative effects on SWE and DS labor, and I'm unsure if I'll have a career in 5 years despite being a senior with a great track record. So, agreed. Save what you can.

I've been reading this since 2023 and yet it hasn't really improved all that much. The same things are still problems that were problems back then. And if anything the improvement is slowing down, not speeding up.

I suspect unless we have real AGI we won't have human-level coding from AIs.

I'm also of the mindset that even if this is not true, that is, even if current state of LLMs is best that it ever will be, AI still would be helpful. It is already great at writing self contained scripts, and efficiency with large codebases has already improved.

> I would imagine the chance of many of us being on the losing side of this within the next decade is non-trivial.

Yes, this is worrisome. Though its ironic that almost every serious software engineer at some point in time in their possibly early childhood / career when programming was more for fun than work, thought of how cool it would be for a computer program to write a computer program. And now when we have the capability, in front of our eyes, we're afraid of it.

But, one thing humans are really good at is adaptability. We adapt to circumstances / situation -- good or bad. Even if the worst happens, people loose jobs, for a short term it will be negatively impactful for the families, however, over a period of time, humans will adapt to the situation, adapt to coexist with AI, and find next endeavour to conquer.

Rejecting AI is not the solution. Using it as any other tool, is. A tool that, if used correctly, by the right person, can indeed produce faster results.

I have to challenge this one, the research on natural language generation and machine learning dates back to the 50s, it just it only recently came together at scale in a way that became useful, but tons of the hardest progress was made over many decades, and very little innovation happened in the last 5 years. The innovation has mostly been bigger scale, better data, minor architectural tweaks, and reinforcement learning with human feedback and other such fine tuning.

I sure hope so. But until the hallucination problem is solved, there's still going to be a lot of toxic waste generated. We have got to get AI systems which know when they don't know something and don't try to fake it.

This is the ultimate hypester’s motte to retreat to whenever the bailey of claimed utility of a technology falls. It’a trivially true of literally any technology, but also completely meaningless on its own.

I find people mostly prefer what they are used to, and if your preference was so superior then how could so many people build fantastic software using the method you don't like?

AI isn't like that. AI is a bunch of people telling me this product can do wonderful things that will change society and replace workers, yet almost every time I use it, it falls far short of that promise. AI is certainly not reliable enough for me to jeopardize the quality of my work by using it heavily.

It’s because they know it works better every day and the people controlling it are gleefully fucking over the rest of the world because they can.

The plainly stated goal is TO ELIMINATE ALL HUMAN EMPLOYEES, with no plan for how those people will feed, clothe, or house themselves.

The reactions the author was getting was the reaction of a horse talking to someone happily working for the glue factory.

I think the feeling stems from the exaggeration of the value it provides combined with a large number of internal corporate LLMs being absolute trash.

The overvaluation is seen in effect everywhere from the stock market, the price of RAM, the cost of energy as well as IP theft issues etc etc. AI has taken over and yet it still feels like just a really good fuzzy search. Like yeah I can search something 10x faster than before but might get a bad answer every now and then.

Yeah its been useful (so have many other things). No it's not worth building trillion dollar data centers for. I would be happier if the spend went towards manufacturing or semiconductor fabs.

More precisely:

In one side, it's the "tools that build up critical mass" philosophy. AI firmly resides here.

On the other, it's the "all you need is brain and plain text" philosophy. We don't see much AI in this camp.

One thing I learned is that you should never underestimate the "all you need is brain and plain text" camp. That philosophy survived many, many "fatal blows" and has come up on top several times. It has one unique feature: resilience to bloat, something that the current smart tools camp is obviously overlooking.

It feels awkward living in the "LLMs are a useful tool for some tasks" experience. I suspect this is because the two tribes are the loudest.

It's a sad commentary on the state of search results and the Internet now that ChatGPT is superior, particularly since pre-knowledge-panel/AI-overview Google was superior in several ways (not hallucinating, for one, and being able to triangulate multiple sources to tell the truth).

Working on our mega huge code basis with lots of custom tooling and bleeding edge stuff hasn't been the best for for AI generated code compared to most companies.

I do think AI as a rubber ducky / research assistant type has been overall helpful as a SWE.

When I started at my post-Google job, I felt so vindicated when my new TL recommended that I use an LLM to catch up if no one was available to answer my questions.