Trouble is afoot
I don't really drink alcohol. It is at times like these, however, that I wish I did. This is the kind of post that it'd be best to write while utterly shitfaced.
Over less than a week, about 10k EUR of value has been wiped out from my investments. While the wealthy Americans happily throw around 10x more money, to me that sum still matters a lot. Over here in Latvia, that is multiple months of salary, whereas in regards to the part of disposable income that you can realistically put in your savings, that's probably close to an entire year of making savings responsibly.
For all intents and purposes, it very much seems like the AI crash that many were rooting for is finally happening.
Yet, none of it makes any sense whatsoever. Here's how someone online put it:
AI stocks keep selling off. What's going on?
NVDA reported $115.8B in quarterly revenue, beating expectations, while Anthropic just pulled off a massive IPO at $83 per share. You'd think those two events would give AI stocks a boost, but the market still isn't buying it.
At this point, is it really about earnings anymore, or are investors finally starting to question how much they're paying for AI growth?
Nvidia had that scandal about Jensen Huang's infidelity that led to a change of CEO, alongside some other stuff, but it still doesn't feel like they stopped making chips overnight, nor does it seem that OpenAI or Anthropic ceased their operations either and no longer provide value to anyone.
And yet, there's headlines like this going about:

Consider that the gains of the entire year have been erased. That is absolutely insane - not just the tech sector, but the entire stock market, this much movement in a week is nowhere near normal or even easily comprehensible:

What could even bring something like that about?
I don't think it's a single CEO's fault, nor do I think it's the biggest IPO in the history, and certainly not the fact that AI as a technology wouldn't be capable. It recently appears to have been used to aid in solving the Riemann hypothesis, for crying out loud:

Maybe it's Nvidia's market meddling catching up to them:

Are the likes of Ed Zitron finally right, after being wrong so many times?
I could ponder whether the opposition against AI is truly organic, or whether there's external influences and groups of people being funded to stoke up the hate towards how the technology is implemented in the US, but I think that at this point it hardly matters - everyone loves to hate it.
Artists, because it stole their work. Developers, because it's both transforming their career and eliminating lots of positions. People in general, because the data centers are being built on the cheap and in locations where they interfere with the local population. In a word, the powers that be have made it easy to hate it.
They are all wrong at who they should be mad at. I previously pointed out how the critiques over power use and water use don't hold up and are purposefully exaggerated, even though the post itself was critical about how the AI orgs treat their customers. In a word, AI's impact is exaggerated.
Regardless of that, right this very moment there are people celebrating the supposed downfall:

There are people who are wishing that we could put the cat back in the bag or even outlaw the technology:

It matches what I've seen with people blaming AI for the state of the software security landscape right now, with how there are more and more vulnerabilities found, as if not knowing about them would have been better in any way, as if ignorance is preferable to knowing how bad things are:

That same opposition more or less matches the Democrats' goals after midterms in the US, with a proposed data center moratorium and even an outright RSI ban. It hurts me to see that, because it's not the technology that's at fault, it's everyone being shitty and just outright poor human beings along the way. I'm not even going to invoke China not slowing down as much as some sort of a boogeyman - they are also seeing trouble with funding the sheer amounts of compute AI needs.
What I want you to do is ask yourself, why do we see posts like this:

Why do we see mass layoffs, that often get directly pinned on AI? Is it really because human employment is not necessary and AI is effectively replacing human beings?
Listen, I use the technology daily, in agentic workflows and have done more in weekends than others do in weeks, I've moved project development ahead months in weeks and not just some vibe coded slop either - but addressing real security flaws and architectural improvements and introducing more stringent testing to make software quality go up.
It is definitely a productivity multiplier, sitting somewhere between 2x and 10x depending on what kind of work you do. No, not just boilerplate code anymore, but large and complex tasks, sometimes being better than I am after over a decade of building software, as long as you guide and guardrail it enough (including custom linters for your project architecture and various other checks that historically weren't done). I'm not sure about claims of AGI, I'm not sure whether that even matters at this point, when the productivity and quality gains are very much real, as long as you don't do a half assed implementation of it.
And yet, me being more productive does not in any way mean we need fewer employees.
Your bosses are lying to you (and capitalism sucks)
You don't replace a tradesman with a power tool. If the same person can do more, your business can also do more. Not doing more and instead trying to cut down on how many people you employ is just an admission that you as a business are incapable of growing and taking on more work.
Yet, everyone ignores that. Companies out there are making record profits and still doing layoffs. All that matters is shareholder value. In other words, it's greed. They don't care about you as a human being that needs to make a living, they aren't optimizing for long term worker retention or improving the circumstances for the community either. They even stoop so low to enforce stack ranking and artificially create churn.
You aren't a human being to them. You are an expendable resource.
The powers that be will try to devalue human labor and if by some set of circumstances they'll go too far and end up with a labor shortage, they'll just try to hire back some of the people that they fired, but this time around with lower salaries. AI is just a convenient scapegoat and method at play here.
Shifting the blame from greed to AI is a brilliant move by the scumbags in power.
It's the same greed that leads to data centers being built on lower cost land that often will make the lives of the locals worse, it's the same kind of greed that leads to evaporative cooling being used instead of closed loop systems even in areas with poor water availability, that is also what leads to temporary gas turbines polluting the area illegally and so much more bullshit that frankly should have never been allowed.
Most of the people who made that happen should be fined into oblivion.
Companies took one look at this new technology that had untold power for good and went:
"Addressing our technical debt? Using it as a tool to do the things we didn't have enough resources for before? Improving the quality of our services for our clients and the market at large? Hah, fuck that, let's let go 25% of our workforce and make the remaining workers do more work for the same pay under the threat of them also being let go. Hell, if we do this at a large enough scale, everyone will be desperate for work and we can do the financial equivalent of giving a big middle finger to all of those white collar careers."
What's worse is that under capitalism they just don't have an incentive to make for a better world, to improve the local communities and the quality of life of the workforce. Instead, they attempt to maximize profits at the expense of all else, and in under-regulated capitalism coupled with a distracted and uneducated society, they are getting away with precisely all of it. People have so much brainrot that they look up to the billionaires and millionaires and believe that they themselves are just temporarily embarrassed rich people and some day will also be able to partake in the art of: "Fuck you, I got mine."
There is pretty much nothing that dictates that you must replace the people doing transcription and translation with LLMs - and yet, businesses realized that nobody would really put up that much resistance against the loss of human touch and introduction of untold amounts of slop (because obviously they'd half-ass both the implementation and use the lowest cost models they can) and went right ahead. Now you have tasteless idiots doing the same for graphic design and so many more careers - which is a damn shame because even in IT we need good UI/UX work more than ever!
On an individual level, due to how fucked the economy is, you also see a torrent of both book slop and game slop flooding every marketplace:

I don't have a problem with someone creating a game or doing work that they just couldn't previously - I myself am working on a fork of the Flax game engine and it's quite fun, however the difference is that I've spent months on it and only now have a decent basis for world generation and weather effects. A game at this rate would probably take 2-4 years to make, even if me doing a project of such scale solo would take 4-8 years, which feels increasingly less realistic.
That's not what the people rushing after profits are doing. Their goals more often than not are to churn out as much content as quickly as they can, in the hopes of some of those hitting it big and earning money. Knowing how attention focused our economy is, I can't wholly blame them - but the end result is that an ever increasing amount of games have that Claude UI feel and there are troves of books out there full of Claude-isms in their writing, with the same overused tropes and the same characters named Elara and the same positivity bias. A power tool does not write a good novel, I can tell you that much.
My whole argument is that you're buying into the bullshit: you hate AI too much, you don't hate your bosses enough. You don't hate this world humanity has built and the profit incentives enough. Thanks to capitalism, it feels like we're at one big rat race before it all falls apart, and humanity at large is losing here.
Blaming AI for the state of the world is nonsensical beyond belief. It's like all of the culture war bullshit in the US - a helpful distraction from how the rich are getting richer and are robbing everyone else blind. You are all barking up the wrong tree.
Just look at how disproportionately costs and income are rising in the US:

I can assure you, it's not much better in the EU either. Even with my currently good software development salary, unless I want to get a mortgage and go into debt, there's frankly no realistic way for me to even buy an apartment.
I cannot begin to imagine how much life sucks for teachers, policemen, firemen, or any other number of careers that provide equivalent if not much greater benefit to anything I do, yet receive horrendous pay. I do not envy those who have to support an entire family in the current economy either. We're past a point where it feels like if you work hard enough you'll be rewarded and will have a shot at a good life.
A singular scapegoat will not be sufficient to explain all of that, no matter how much the moneyed interests might try.
AI could genuinely be good and help us
For a second, let's put the question of copyright, unethical AI training and people at large currently not receiving any benefit from their specific data that was used to train the models aside, because we ain't solving that and I don't think anyone will.
Without all that, let me ask you a few relatively simple questions:
Do you want to die? Do you want to be trapped in a decaying flesh prison and watch time and illness take away any sort of enjoyment of life you might have had from you? Do you want to watch that happen to those around you, the people that you love and care about?
I'm not going to talk about some imaginary futurism and concepts like death escape velocity or suggest that anyone could live forever. I won't even say how it's a tragedy that human lifespan is limited to under 100 years for most of us, that there's only so much you can experience before you pass away, that erasure of consciousness and death itself is one of the biggest tragedies there are, I won't bother engaging with attempts to cope with it like religion (which might have some other societal benefits), or non-solutions like "living on in the memories of others".
I won't engage with the ignorance we have in regards to mortality as an engineering problem at large, or the fact that a human being that'd live a century might still be insanely immature, which can be surmised by looking around and seeing how much people are messing up the environment, knowing that they'll be dead by the time the effects become apparent.
No, I'm talking about something much plainer - healthy life expectancy (how many years of good health people are likely to have) and illness.
To me, it's self evident that at the absolute minimum, people deserve to have dignified lives, even when they're older. That we as humanity should do what we can to treat both those who get sick and our seniors as well as possible, that we should help them. We should treat whatever cancers there are, because statistically, if something else doesn't kill you first, cancer eventually will. We should look for any ways to treat cardiovascular disease and other issues that arise from the average way human beings live - advocate for healthy living, sure, but don't expect a sentient bag of chemical reactions to make the best long term choices that go against the pressures and cravings of evolution all the time.
If there's even a 5% increase in our odds of beating any type of cancer and other types of illnesses that befall humanity, we should take it.
AI is one of the tools that could help us get there, as proven by its capability to solve novel maths issues, and with enough work it will work across various different domains. AGI or not, I don't care, as long as the results are real.
We are finally at a point where those results are very much becoming real:
Using this benchmark, the system achieved 83.2% accuracy on New England Journal of Medicine clinicopathological conference cases and 80.3% on MIMIC-IV-ED emergency department admissions, outperforming generalist physicians (36.1 and 46.8%, respectively).
That is even before you consider how overworked (and sometimes plain bad) your average physician will be, when compared to a system that can be trained with more data than the human brain will be able to remember, has more working memory and better recall, alongside not getting tired, bored, as well as not being as prejudiced as humans can be (I've heard from so many women that their health concerns are straight up ignored).
Even with the fact that we suck at building AI assisted systems due to our collective incompetence (even web development is like the Wild West out there, software engineering is the fakest type of "engineering", unless you're working on planes or spacecraft; I would never step on a bridge if it were built like software is), even with the variety in how LLMs predict text, the variety in performance there will still be better than the difference between a well educated, brilliant and attentive doctor, and someone who's been warming their chair for a decade and just wants you out of their office.
The same could as well apply to pretty much any industry out there, where AI becomes good enough - software development is obviously another example. There should be nothing pushing you to ship vibe coded slop, instead if given the proper resources and time, you could eliminate most of the glaring vulnerabilities and issues with your architecture, actually work on making the software provably correct and test it thoroughly, improve the infrastructure, utility scripts and overall developer experience throughout, alongside better user experience, better products, better resource utilization and resiliency.
Or you could partake in this capitalist hellscape and ship another sloppy SaaS that won't turn a profit in an effort to get at least one of them to eventually make it big, so that you may stand on your pile of dead shitty projects and feel like a less of a wage slave.
I don't expect some mythical moment of "singularity", rather I want the world to suck less as things improve more with each decade. Sadly, I don't think anything can overcome and resolve human greed, hate, ignorance and bigotry, this little hell on earth that we've built ourselves, but we should just go for whatever improvements we can make.
It's lonely here
Yet, even something as simple as being generally pro-AI makes me politically disfranchised, because while generally I'd consider myself left leaning, this one question is where my views are the complete opposite of most of the left leaning people that I know - which, given all of the above, makes precisely zero sense. It is lonely, to see the potential here but only confronting hate online. Seeing people treat a new type of computation as nothing more than a pump & dump to enrich themselves, even if they have to make the global markets and economy crash in doing so.
It's like people are advocating for faster horses instead of cars, and thinking that we should all outlaw gasoline cars and their development due to the pollution that they cause - not even something that's intrinsic to the degree that the industrial revolution saw, but rather a consequence due to nobody caring to measure and limit it. The solution was to build better cars until something better came along, instead of outlawing them altogether. On the flip side, those that say we shouldn't embrace public transportation are also wholly wrong - in many cases, trains and trams are buses are indeed that "something better".
I guess what I feel the world needs more of is utilitarianism and less of dogma.
So yeah, have fun making the stock market crash, however I seriously doubt that it will lead to affordable PC parts anytime soon, the same way how I feel like Anthropic's IPO will inevitably lead to more stringent token limits and people being able to do less with the technology. It could have been a slow adjustment without likely impacting the livelihoods of millions in the coming months. I hope your triumph and seeing red feels good to you, though it feels like you might actually be fucking over all of us if that leads to a large scale economic recession.
Whatever the complex circumstances at play may be, I hate it here.