One of the things that makes me very sad about AI is the quality dropping all over the place and nobody really seems to care? At least no where near the amount of people I would have thought.
There are people I’ve worked with for years that were very good at their job. They were always proud of everything they did and took the time to do stuff right instead of half-assing a job. But then they started using AI and just press the easy button on everything they do. I’m pretty sure some of them no longer have the skills they once had and would take 5-10 years to get back to where they were. The quality of the output has been dropping steadily and my pushback on the quality hasn’t been welcome and AI has been blamed for mistakes. Like my dude, just because you pressed the AI button doesn’t mean you are no longer responsible. The AI didn’t fuck up, you fucked up, you are supposed to check the AI output. It’s still your work, even though you let the AI do all of the work. Management has been praising the increased output, even though I feel the only reason there’s more output is because the quality has dropped. Instances where a 4 hour job takes 20 mins are heralded as the best thing since sliced bread. But there are so many instances where a job gets sent back multiple times, because it was simply wrong and terrible. But that’s labeled as me being picky or hating on AI just because it’s AI. And there have also been times someone was hard stuck because the AI couldn’t do what we needed and I needed to take over the task.
I feel like the odd one out, not praising AI and pushing back. I wouldn’t have thought so many people, especially people I used to respect a lot, would become so in love and addicted to AI. In a couple of years our code base is going to be totally fucked and nobody is going to be able to get it back.
Tech has been my passion my entire life and for the first time I feel like I might not be a fit for tech anymore. I want to be proud at something I created, not take credit for pushing the “Do the thing” button. I want to craft my own code, not be a sloperator or “prompt engineer”. Hardware has become so expensive, my current computer looks like my last computer for at least the foreseeable future. And it seems like I’m the crazy one out, with everybody being happy with the slop the AI shits out.
I think this article goes into some details why what you describe is happening: https://hermit-tech.com/blog/ai-mania-is-eviscerating-global-decisionmaking
And yes, LLM written code is the asbestos of our time that will take decades to clean up.
As expected from an article that attributes “insanity” to people, it misses the root for why it is happening. It talks about board members, CEOs, and regular personnel, but it fully ignores shareholders.
AI promises to be the final conquest of capital: a form of labor that is privately owned and wants to be privately owned, which can be removed from existence when it is not needed and which needs comparatively little to exist. No revolution, no union, no secret loathing.
In a world that is rapidly going up in smoke, capital owners have few places to invest. The stock market is a bubble where speculation is far more important than return on investment. AI is just about the only hope shareholders have for the continuation of their wealth.
All the “insanity” of corporations is them trying to cater to shareholders’ actual need. Not profit, and definitely not production, but the development of a form of labor that will not rebel as billions perish.
If AI not working means the rich will get eaten, the rich have no incentive to care about realities in which AI does not work, no matter how likely. AI fanaticism is perfectly sane, from their perspective.
Very well said. Agreed with those material realities or at least fantasies.
I also think there is a more immediate, less existential incentive, that lets individuals pretend like they’re just acting out the (bad) efficiencies of the rational new market while ignoring the awful existential endgame of continuing to enact the will of capital (I won’t say the will of shareholders because many of them would also claim their hands are tied).
They can pretend it’s just one more of their money tricks where they extract more profits or externalise more costs, like with the Metaverse or Web3 or NFTs before it.
I like how Rushkoff puts this: “going meta” as owners’ way to keep the grift rolling, just keep going more meta, more abstract, more virtual. Until the market is so far removed from the reality of trees and rivers and human bodies and clean air… well anyway I’m on a rant.
But my point is I hate how this fucked over financialized worldview makes it seem like every step along the road to apocalypse is rational and inevitable. Like we’re powerless to stop it.
It’s an economic theory, not an asteroid hurtling towards us. It literally runs on people.
Thanks for the link, interesting article.
And yes, LLM written code is the asbestos of our time that will take decades to clean up.
My fear is: Nobody will clean it up, this is just how tech will be from now on. It might not be a fully rational fear, but I feel like we’re not living in a rational world and haven’t been for a while now.
I feel like that’s a fear for everybody that lived through Y2k. We came so dang close to everything breaking, and I think we came away from that with hubris and not foresight.
Yeah I remember those days.
I made bank back then tho. I did a small project for a local company to get their stuff ready voor Y2K and they had a group of techs working on a lot of stuff. I started running with that crew and did job after job, mostly doing the exact same thing we did for the previous company. And the money was very good, since a lot of companies were scared they wouldn’t be ready.
For me it did feel very different, because everyone was afraid everything would break. But I was actively working on preventing stuff from going wrong. And in the end we did a very good job, all of our clients got through without any issues.
Now I feel like I’ve been thrown in the deep end of the pool and no matter how much I swim, I keep sinking further down.
I was pretty young when it was going on, so I didn’t really understand it until later. I just worry that we’ve, culturally, come away from it with a “oh, I guess it wasn’t that big of a deal after all, we can do this again next time” mentality and not a “it took a lot of hard work to keep this from happening, we should prepare better in the future” outlook.
I’ve got the same worry about the lessons we learned after the pandemic, now that I think about it, lol.
I know that odd one out feeling when it comes to AI too. My last role I frequently said the AI output was bad and usually wrong. Others would blindly copy/paste its output and send it to the customer, seemingly without reading any of the customers message or the AI output.
Hear me out, it became bad once the business people realized how much money is in there. They started squeezing.
Never let business people touch anything.
Besides suffering from a heavy ‘bullshit jobs’ problem tech is the new ‘I was just following orders’.
‘I was just writing code. I had no idea I was writing code for the baddies!’
I used to teach Computer Science at university, long before the “AI boom”. Even back then, we were overloaded with students who had no true passion for technology or science but were there only because they had been told it’s a good career move. They were convinced by their parents or schoolteachers that tech is a high paying, high status job like law or medicine.
When they graduated and actually went to get a job they realized the market was already saturated with others like them - who had rushed into an IT degree and a career that they didn’t really enjoy.
Some will say: good, fuck ‘em. In the 2010s, Knowledge Workers told everyone to learn to code while they sipped kombucha and played Xbox in beanbag chairs. Then, Knowledge Workers sat inside during the pandemic while frontline workers risked their health to make Amazon and Uber Eats deliveries. Now, those same Knowledge Workers are building AI that threatens to eliminate work for humans across industries and make a very few people wealthy beyond imagination. And these same assholes want pity now?
No coders I knew in the 2010s were telling others to “learn to code” because they already knew the above - tech was already oversaturated with graduates in the 2000s. We were telling students to do what you are talented and/or passionate about. It’s only pundits and posers like Musk who tell people to “learn to code”.
On the other hand, the people who actually do have a passion for science and technology are burnt out because of the commercialization of tech. They want to work on interesting tech that makes people’s lives better, not add features to products that will sell more licenses. The article’s discussion of “Workism” encapsulates some of this, but all the problems it describes were still there many years ago before the “threat of AI”.
Still true. I’m a recent-ish CS grad and I started programming when I was a tween so I could mod games. So many people in my classes were former business majors who switched cause apparently CS is trendy. These people had no idea what they were doing and even when using AI to cheat, they still had a hard time completing undergrad CS assignments. I know because I would often finish early and the professor would ask me to help cause so many people kept needing help like constantly. I’m still unemployed and honestly becoming a farmer is seeming more and more enticing to me these days.
No coders I knew in the 2010s were telling others to “learn to code”
Maybe it’s just because I’m on the west coast, or maybe it’s because I’m not that into tech, but I’ve seen it.
I went into bio/psych during college (graduated HS in '06 for reference), but never finished, and I eventually specced into caregiving for developmentally disabled adults. It only paid a dollar more than minimum wage, but it was highly rewarding until the agency decided they were going to start doing end-of-life care for some of our older clients. Even though science has known for a while that the gene that codes for the plaques and tangles of Alzheimer’s is on the duplicated gene that causes Down’s Syndrome, which causes 100% of people with Down’s Syndrome to develop early-onset Alzheimer’s, there was zero professional guidance on how to help them through it. I was working 60hr weeks, the insurance provided wouldn’t pay for more than a dozen trips to therapy each year, and was being physically assaulted at work by a 50yo guy who was seeing the hat man because Lewy Bodies.
Some “friends” were less than sympathetic, because if I wanted a good job, I should have just learned how to code like them instead of “caring about r****ds”. I hope they’ve all lost their nice jobs. If they’d said this to my face, I would have punched them and not stopped punching until my arms got tired.
I quit being a caregiver in 2018 after a car accident, and was just getting back into the workforce when the pandemic hit, but even through COVID, I couldn’t log into social media without seeing people patting themselves on the back for being smart enough to learn programming so they didn’t have to be be on the front lines.
he article’s discussion of “Workism” encapsulates some of this, but all the problems it describes were still there many years ago before the “threat of AI”.
Yes, that is kinda the point of the article, no? All these problems are not new, but the AI situation might boil it over to the point where people quit all together and never look back in mass (as opposed to the occasional burnout that caused this so far).
Cuz ram is expensive!
In their vision for the future, work goes from a messy experience of learning and exploring to something more akin to assembly-line production.
I found it hard to reconcile this vision with today’s LLM technology stack until I realized that the people who fall for this are being deeply fucking stupid, and don’t have any real friends to tell them how deeply fucking stupid they are being.
