Perhaps they only call you guys in when middle management sucks, so you have a jaundiced view. I’ve had some managers who were real idiots, and also some managers who knew what they were doing, including one who built the group from not having any respect anywhere to being known and respected throughout our industry,
Do you think Musk’s minions would have any clue as to which of the people doing the bullets were dead weight and which were vital? I don’t. If I had been making the bullet list I’d have put things I think they would find valuable, not things that really were valuable.
Remember, the guy in the weather service in Texas who was forced out was the guy who had contacts with local law enforcement, who believed him when he said this flood alert was real. You think that part of his job showed up on the bullet list? (He wasn’t fired, he retired, but IIRC because Musk and Trump treated them as useless.) So, when the flood came no one alerted the cops properly, they didn’t do warnings, and bunch of girls died.
Most of the people I worked with were pretty good, but I worked in high prestige tech companies.
It has made me a much, much lazier thinker. I never write even the simplest code anymore, just ask AI to do it, whether it’s python or an easy VBA routine, the AI does it better, faster, with better documentation and better error checking than I ever did.
When it comes to presentations, I just feed the docs to an AI and out pops a fantastic set of slides, better formatted than I ever did, with interesting insights that I never thought of.
Front end to a database backend? The AI will make a beautiful front end, with options I didn’t think of, and connect faster and more efficiently that I knew how to do.
All of our old spreadsheets are being replaced with python/HTML/Javascript front ends, completely vibe coded.
I don’t even try and figure out if there’s some Windows environment variable that’s off – if the AI has a problem trying to put something together, I tell it to figure it out, download what it needs, modify what it has to, and it just takes care of it.
I honestly don’t know how I would learn anything in college right now – the temptation to have the AI do all my thinking for me would be just too great. You mention calculators – I’m not going to bother with long division, since the calculator can do it for me. Well, for coding, Powerpoint generation, etc., that’s the same with the current AIs – why would I bother doing it, if the AI can do it better and faster?
Anyone who is saying this isn’t some kind of existential threat is just not using it right. And, that’s leaving aside the real existential threats mentioned above, putting the foremost nuclear and biological brains into the LLM and making that knowledge and skill base available to bad actors.
I’d also like to challenge the perspective of seeing people as fungible labor units, like GPUs that could be trivially reassigned to different tasks whenever the need arose.
Reskilling is hard, and not everyone has the resources, home situation, ability, or willpower for it, especially on the cutthroat timeline of modern technology, and especially when there’s no clear role to reskill to.
It’s one thing to say that work in a few decades might look very different, but that’s different than the question of what happens to millions of workers and families in the interim.
I don’t think the answer is therefore “AI is evil and must be stopped”, but the problem shouldn’t be altogether ignored either just because AI happens to be an effective force multiplier for a few people.
The “make do” work may be inefficient, but sometimes that’s still more beneficial than not having the work at all. Keeping people occupied on seemingly trivial tasks at least gives them something to do that isn’t meth, provides a subsistence living, and keeps them active in the workforce should better opportunities come along. Just abandoning them to decay into the next former coal or logging town isn’t great for anyone except a few capitalists who stand to get even richer while society picks up the pieces.
I’m one of those who happened to be lucky enough to have transitioned from natural resources to solar to tech to AI. I was cutting redwoods with a crosscut saw one year, hooking up solar inverters the next, and learning Javascript the next. In that same period, I also saw friends succumb to poverty, spiral into alcoholism, and commit suicide.
The Dope skews old, smart, rich, and white. A lot of the country isn’t like that, and a lot of the world is even worse off. Life is hard enough for them already that I don’t want to see them get shoved even further into forgotten crevices, all in the name of “the show must go on”.
Even as someone who often straddles the line between the haves and have nots, I think the AI divide is going to be different both in scope and magnitude. And not in a good way, unless we are able to exercise some democratic checks and balances over it and have transition support for the billions left behind. That problem doesn’t magically go away just because it makes Excel or Javascript a little less tedious for some of us.
OK but that’s still an improvement. My navigation skills may have become rusty because of the availability of Google Maps and GPS, but it’s still an undeniable benefit. Have my walking and running skills become eroded because I have a car to drive places? My arithmetic may be eroded because I have a calculator to perform 39876 divided by 114 instead of having to spend minutes doing it myself, but aren’t we better off for having calculators? What’s the solution - intentionally make ourselves primitive for the sake of…being more manual about our work?
Are you now saying that AI hasn’t made your thinking sharper?
ETA: Oh, walking and running have definitely become eroded since the advent of cars, people are more sedentary, etc. In my view, it’s obvious the benefits of cars outweigh the reduction in day-to-day walking and running.
This isn’t my particular concern, but I think “I don’t want to lose my livelihood” is a sufficient reason to object. Expecting people to just roll over and accept that they and their families may die in poverty in the name of progress seems really unreasonable. If we want people to not have those fears, we need to provide an alternative path. The CEOs are the ones harping on the great labor replacement narrative. So they made their own bed and it’s their responsibility to convince people this is somehow better for everyone.
The only thing I’ll add to this is I generally don’t feel good about things I make with AI. I feel stupid saying I made them. I feel stupid saying I made them with Claude. There’s definitely a loss of self-efficacy and less pride in my work.
It’s made my thinking sharper in things that aren’t technical calculations, such as thinking about life advice or decisions or pondering human behavior (in conversational mode.) It offers various perspectives that I hadn’t thought of before.
If you’re just using AI in conversation mode, you have no idea what it’s about, and how monumental the potential effects are.
To use your examples, calculators made your ability to do math worse. For most people, cars made us walk less. For me, I can barely write by hand anymore, since I pretty much only communicate by typing something, so computers have made my ability to write longhand much worse. Maps and GPS have degraded our ability to navigate.
AI is a thinking tool. It has the potential to degrade our ability to think. That’s what we’re facing (aside from the real concerns about bioweapons, hacking, and other real concerns that you appear to be ignoring completely). It’s not Luddism to worry that if AI can think better than people, then what will people be good at? And, it’s not Luddism to worry that putting unbelievably advanced bioweapons and hacking ability into basically anyone’s hands is problematic at the very least.
It depends on who has access to the magic box, who owns the magic box, who gets money when the magic box is used. If only a small handful of people get to use the magic box, it is probably disenfranchising everyone else.
I think this reference fails for 3 reasons.
1) Everyone thinks they’ll be the last one standing.
At my last job, the owner of the company was moving to stop hiring non-developers, full stop. Why hire UX people if a developer can prompt AI to make a UX? Why hire analysts if a developer can prompt an AI to produce a detailed process map? Why hire program managers when a developer can prompt an AI to generate reports and plan milestones?
Of course, the UX people were asking, Why hire developers when I can literally feed my figma designs into AI? The PMs were asking, Why hire developers when I can feed my requirements directly into Claude?
And my response to him in the discussions we had about the insane (to me) direction he was taking the company and the impact it was having on the “corporate family” was, Why would any customer come to us in the first place if they can just type their vision directly into AI?
“Ah, but AI will never be good at that.” Indeed, everyone sees themselves as having the last remaining unique skills.
Certainly there will be last people standing, and they will likely have the critical thinking skills you describe, but AI will get better at critical thinking very soon, and then what? Why pay an engineer 200k a year when you can pay an intern to ask Claude how to think?
2) OK, so AI is coming for your career – better pivot now! But to what?
Being older and late in your career feels pretty nice, but I have teenagers. What do I tell them to go into at this point? I myself am pivoting into teaching, which a year ago felt fairly safe. But now I’m seeing videos of the AI interview companies (they suck, btw), where you sit down with a life-like avatar who asks questions and generates a summary for hiring managers. Maybe we’re 3-5 years away from a life-like avatar teacher who can assess a student’s current progress, develop an individualized curriculum, teach it, and then determine if the student learned it. We’re probably less than 12 months away from an AI tech-ed startup selling a crappy version of this to school districts across the country.
OK, but what about the trades? We’ll always need plumber, roofers, electricians, etc, right? Well bipedal robots are coming, and they’re coming at a faster rate than before because AI is currently accelerating their development. How far out are we, really, from a humanoid robot that can come do your plumbing? Almost 10 years ago I declared that self driving cars are still decades away, and I’d like to think I was right, but… the clock is ticking. We’re 1 decade down from that post, and maybe 1 decade away from autonomous vehicles being ubiquitous.
What industry can anyone honestly say is safe for an 18 year old to get into?
3) Who owns the means of production?
Imagine a mechanic who didn’t want to learn how to use the new-fangled scanning machine, because “cars were simpler when they weren’t all computers.” What a maroon, right?
Well now the scanning machine is owned by the manufacturer and costs $190,000. It’s subscription based on top of that, and any tech who has a login has to complete mandatory training at an additional cost.
Nobody owns anything anymore, and while some dopers here have set me straight on advances in local LLMs, that’s not the way the world works anymore. This is not going to be democratized technology – the amount of money these companies are spending to bet on AI comes with expectations of future profits. Even if AI is “yet another technological step that we all need to get used to,” the last 20 years of technological steps have all sucked for the average Joe. We’re all getting squeezed by everyone with a lever, and it’s only going to get worse when these companies have control over our livelihoods.
That should concern everyone.
Speaking of unsustainable business models, I just got a notification that Claude Fable will no longer be accessible unless I move up to a higher payment tier. Which is absolutely not going to happen. I already pay them $30/month out of my own pocket to use for work.
Someone at Anthropic finally did the math. The question is whether there will be a lot of people like you who just say no, or if people are so hooked that when the pusher raises the price they have no choice.
Given that there seems to have been no measurable increase in productivity from AI yet, maybe the former, and that could trigger the crash.
John von Neumann was against floating point because he thought everyone should know where the decimal place was without computer help. If you never do this stuff manually, you’ll never know when the computer gets it wrong. If we start with calculators, or AI, from elementary school we’ll have a generation of kids without the basics. Same goes for writing, by the way.
I remember my kids’ math workbook in elementary school that had the step-by-step instructions on how to do the work on a calculator (press this button, then this button, etc.). I was appalled.
I have a kid who is precociously good at math. He’s always trying to find patterns in numbers. In waiting rooms and such you can keep him occupied with just his scientific calculator.
But every time I try to explain, “Hey, there’s a way you can do this without using your calculator” he is not interested.
He doesn’t see the point.
I’m worried he’s going to miss out on some foundational understanding of mathematics because of his refusal to do long division or something. I don’t even know how math is taught anymore, beyond the homework he’s brought home from first grade.
The way I was taught, we didn’t really start using calculators until it became ungainly not to use one. Not until middle school at least. And I didn’t have my first scientific calculator until high school.
Calculators are used exactly never in teaching pure mathematics. [I am aware of some exceptions that prove the rule, but for them you would want a fully-fledged computer.] Engineering is a different story.
[The Sphynx from Mystery Men reference]
You gave me 3 failures to reference reason!
[/The Sphynx from Mystery Men reference]
Seriously though, you may be right in a few years. Or a few decades. As they say, being too early is just as wrong as being too late. The question is really how does one prepare for such a future? How do they prepare their children?
All this fretting about what might be doesn’t seem particularly productive because a) no one here knows the future, b) no one is making policy based on what is discussed here and c) presumably most of us still need to support ourselves, as will our kids.
And d) even for those of us retired, we’re dependent on the economy and stock market not being too blown up by whatever changes come our way when.
The only advantage geezers have is that their total time horizon is shorter. Beyond that, they’re living in a glass house at least as brittle as that of a mid-career person of similar SES. And perhaps much more brittle.
Speaking just for myself, I find it disturbingly comforting to know that there’s nothing happening in the 2060s I need to worry about, and probably nothing from the 2050s either.
Sounds like good career security if you ask me.
In actuality, people like me get called in when companies have some sort of problem they want to solve. Which is every company of any size because no company is perfect and there is always some new system to install or some process that can be improved upon.
It’s not like any one particular manager is or is not an “idiot” or whatever. It’s that large companies are by their very nature hugely complex, highly political organizations with a lot of moving parts where no one has the complete picture and they all have their own agendas and priorities.
What is perhaps a bit frustrating is coming into an organization as some variant of a “project / program manager”, “business analysist”, “change manager” or whatever the role is called, learn all these details about how the company actually works, build up a bunch of relationships, and then have to move on to something completely different and start from scratch.
Like you would think having all that institutional knowledge would be valuable, but apparently, it’s not.
Then again, there have been very few companies I’ve been at where I’ve been like “man I wish I could stay here 20 years”.
Interestingly, last year I interviewed at an AI startup who wanted to “automate banking”. Basically sell some sort of AI-driven automation software to banks.
Like you are a handful of guys. Do you have any fucking clue how complex a bank actually is? There are all these systems that need to speak to each other. Half of banking operations still run on Excel and PowerPoint.
Ultimately, it was just a tool for automating code libraries or something.
I guess my point is there is still a lot of hype and bullshit around AI because everyone thinks they are going to wake up in the Matrix tomorrow or show up to work and a Terminator or Cylon or iRobot will be sitting in their desk typing away at their keyboard.
FWIW, your job sounds absolutely fascinating to me. I’m a systems person (macro social work includes organizational management), so the idea of getting to know these complex systems and making recommended changes would be so exciting.
People in general, in my experience, vastly underestimate how complex everything is. Sure, one’s own field is idiosyncratically deep, but everything else is simple and easy! I think part of the AI bubble is riding on this cognitive error.