I guess because I’ve been through several cycles of technology that was meant to make coding obsolete (RPA, no/low code, etc), I feel examples like this are why the OP has a point.
First of all “vibe coding” is not “building reliable professional-grade enterprise software”. So let’s just get that out of the way.
Secondly, “writing as much code as possible” isn’t the goal. The reason code is written is to automate some process or perform some analysis that is too complicated for humans. So one one expect a significant benefit of that sort of analytical processing being made available to a wider population of people.
Finally, most of the fears about AI seem to be a combination of plot points from science fiction stories and overblown Silicon Valley marketing hype. AI is not going to become self aware, take away everyone’s job, and then dissect everyone and place their organs in labelled jars (because AI like categorizing things).
Like any new technology, we do need to be concerned about the impact of AI, particularly in ways we haven’t thought of. Use of natural resources is a legitimate concern. Industry disruption is another. Data privacy. Weaponization of AI. Propoganda.
The fact is we are still in the “AOL.com” stage of AI. It’s a cool technology that shows some promise, but we haven’t really grasped the full extent of its capabilities.
Coding is just one example area I gave, based on my more tangential personal experience of the coding industry, and what I’ve heard, including from my son who is soon to graduate from college with a computer science degree, who spent a month this past Summer in the Seattle area, networking and talking to Tech Industry professionals about the current state of the industry.
As I said in my first post in this thread, I’ve seen the graphics and information industries change from using physical tools to using computer programs to create graphics, and then I switched from the print graphics field to web design and development. So I’ve been through cycles of technology as well.
The rise of AI is, to me, nothing like previous new technologies, that promised more efficient workflows but still required knowledge and skill to learn them. Skilled writers, marketers, designers, illustrators and yes, coders are becoming less and less necessary, their skills and talent being replaced by simple prompts saying ‘do (ABC) in the style of (X), using (Y) parameters’. And maybe a few tweaks to get it just right.
All of those concerns are real and happening now. Not science fiction or apocalyptic catastrophising. And AI is quickly getting better at the various things it can do.
You seem to be pretty sanguine and unconcerned about the future of AI regarding its impact on our continued ability to earn living wages while doing work that is actually rewarding. I hope your viewpoint is correct.
Look, computers could always do thinks better than humans can do. The same way industrial machines do things way better than people can do. That’s why we use them.
If I seem unconcerned maybe because I’ve spent most of my career building systems to automate various processes. While it can often be disruptive in the short term, long term it tends to free people up from work they found tedious and boring in the first place. Which then ideally frees them to do stuff they actually want to do.
I do wonder if people would be more open to the idea of AI if most of the stories about it weren’t basically all plagiarism LLMs being illegally (or legally but distastefully) fed “training materials” by companies that nobody really likes or understands and with CEOs that all are incredibly unlikable to everyone expect for the Chinese honeypots pretending to be interested in said CEOs.
Except, AI is not just replacing the tedious and boring work, it’s replacing the stuff people actually want to do. Which they also were reasonably well paid to do.
I think there’s some truth in that. For the most part, these are unlikeable CEOs running enormous and inexplicably wealthy companies that don’t really seem to be providing any real benefit to humanity. Their entire marketing narrative seems to be “AI is the most important thing in the history of everything and it’s going to destroy everything you know so give us a ton of money so we can build it first!”
I have a feeling that it’s not going to be much different than what we already see. You’ll have some companies that’ll be happy using AI slop in their business, while others are going to specifcally want human-created stuff. And others are going to want a mix. It’ll be a mix of cost and corporate image I suspect.
Not much different than crap made in China vs. stuff made in the US or EU. Or corporate graphics- we already see some lazy companies who obviously use the most basic tools like MS Word because they end up with janky, public domain fonts and stuff like that. Others have professionally done stuff. Or for that matter, with food. Some people are happy eating pasta that comes frozen in a bag with sauce nuggets, etc… Others are willing to boil their own pasta and slap some jarred sauce on it. Others will buy the cold-case fresh pasta (Rana, etc…) and some will go to restaurants where actual chefs prepare it from scratch daily.
What spooks me are the military applications. I can completely see battlefields being entirely lethal to humans, with a combination of advanced sensors feeding AI controlled drones and weaponry.
In my particular wheelhouse there are a lot of creative people who can malke a reasonable income working as graphic artists, illustrators, copywriters, etc. The best of them could eventually rise to be a creative director and make a lot more than just “reasonable.” They genuinely enjoy getting paid for doing what they like. Right now AI is able to do all the mundane stuff those people started their careers doing (and developing their skills along the way.) If you wipe out the entry-level jobs, they won’t come back and eventually AI will wipe out the entire “creative” industry.
I’m not a Luddite who thinks AI will be the end of civilization, but I think it will be WAY more disruptive than its fans are willing to admit.
It’s more usual to refer to “Luddism” than “Luddite-ism.” It is named after a fictitious “General Ludd.” As one contemporary poem had it,
Chant no more your old rhymes about bold Robin Hood
His feats I but little admire
I will sing the Achievements of General Ludd
Now the Hero of Nottinghamshire!
Other posters have made important points about the real history of the movement and the impact the new machines had on workers. So the “straight-up Luddite-ism and ‘normal growing pains’” of the OP blur some important matters. The boss’s “normal growing pains” are often generations of poverty and dislocation for workers.
Fun facts: at the height of the movement, the British government had more troops suppressing the Luddites than they had with Wellington fighting the French in Spain. The authorities offered the equivalent of 2 years’ wages for a skilled worker as a reward for diming Luddites. So, kinda serious.
Lord Byron’s first speech in the House of Lords was a defence of the Luddites, as was his poem, Song For The Luddites, 1816
As the Liberty lads o’er the sea
Bought their freedom, and cheaply, with blood,
So we, boys, we
Will die fighting, or live free,
And down with all kings but King Ludd!
When the web that we weave is complete,
And the shuttle exchanged for the sword,
We will fling the winding sheet
O’er the despot at our feet,
And dye it deep in the gore he has pour’d.
Though black as his heart its hue,
Since his veins are corrupted to mud,
Yet this is the dew
Which the tree shall renew
Of Liberty, planted by Ludd!
For example, I share the same fatigue as the next man at seeing AI-generated billboards and flyers…they have a very set style which means they clearly read as AI and make everything look so generic and lazy. Right now, something crudely put together by someone with day 1 photoshop skills is preferable, because at least it looks unique.
But this won’t last. Every time someone says AI sucks because X, X is solved pretty quickly.
I’d give it…3-4 months before AI made flyers look distinct enough from each other that it becomes almost impossible to know they are AI made (I’m guessing it’s already possible to make good flyers with the right prompts, but I mean the point at which simple prompts to free versions of the main engines deliver such results).
A lot of people talk about AI pejoratvely, referring to the results of AI-produced content as ‘AI slop’. As if it will always be that way. And yes, for now, it’s still usually possible to detect when an image or filmed scene is a product of AI, but that realism gap is closing quickly. And a lot of times ‘AI slop’ is ‘slop’ not because the AI is at fault, but because of the laziness or bad intentions of the human who directed the AI.
Exactly. 6 months ago AI could not do, or do nearly as well, as what it can now.
And it’s not only getting better at doing what we ask of it, it’s getting better at understanding what we’re asking of it. I’ve done some work with AI prompting-- the company I work for actually encourages us to use AI to improve our workflow efficiency. The company viewpoint (at least, for now) is ‘Humans + AI’. Often, I get something wrong in my first prompt attempt and the AI will ask ‘did you really mean (X)?’ and yeah, it noticed and called out my mistake.
“Humans doing the hard jobs on minimum wage while the robots write poetry and paint is not the future I wanted.” Twitter - KarlreMarks
I suspect the mundane result of AI is that it will fill a particular niche, and it will establish a new (somewhat lower) normal for creative remuneration and quality, but it will enlarge the pie of creative works and those who can partake of them. Thinking specifically about artistic endeavors (graphic design, visual art, writing, etc.) the computer and advanced printers, whether 2D or 3D, have made art easier to generate, lowering costs, and allowing more patrons and commissions. More people can be artists too, but there is more art generation capacity than demand, so wages fall. The same will be the case for AI. The thing to remember is that every AI-generated restaurant menu or flyer or web graphic is not a lost commission. Just like every pirated copy of a song or movie is not a lost sale. “They were never going to be your customer before” is a valid argument, but some might have been, and that does put some downward pressure on the industry, if not at a 1:1 ratio. It’s just not as interesting as “the AI robot overlords are going to wipe out all of Humanity.”
I wish I could share the cheerful, optimistic long view of the impact of AI that some of you have.
What I’m observing in the here and now is established skill workers becoming unemployed, or at least finding their skills and talents marginalized. And young people out of college having a hard time finding entry-level positions.
Yes, as I mentioned upthread, for some established skill workers AI is making their jobs easier and improving their workflow efficiency. At least, for now, until AI might replace them completely.
But for the highly skilled writers, illustrators, software architects, etc., who used to enjoy their work and take enormous satisfaction in a job well done? “Man, I really nailed the wording on that AI prompt I gave the bot-- it generated exactly what I was envisioning on the first try” I think just isn’t going to have the same sense of ‘a job well done’.
At least over the course of my career, what I’ve noticed is a probably 15 year trend toward companies choosing to hire “off the shelf” to fill positions ASAP, and with as little disruption and training needed. It’s been more important to fill a position with a qualified person than it’s been to actually promote someone in a lesser position and try them out/train them up, then hire someone for that lower level position and train them up.
I mean, the only place I’ve worked that had a significant number of entry-level professionals was when I worked for a consulting firm. And I’m pretty convinced that it was because they could lure them in with the promise of (relatively) high pay, lots of travel, and all the fun that goes with being a single consultant who travels a lot- i.e. lots of going out to bars after the consulting day is done, etc…
But in boring old corporate IT? It was far more likely that they’d choose to hire someone 2-3 years out of school with some nominal skill set and experience than take a chance on a kid out of college. Same thing at every level. The only two places I’ve ever worked where they had a steady upward pipeline were consulting and local government. Both of those routinely promoted/promote people into higher roles and leadership roles. But only the consultants hire entry level workers. The local government seems to hire mid-career people who want off the frantic hamster wheel of private sector corporate IT, not kids right out of college.
It’s a combination of risk aversion combined with a mindset that training up the inexperienced is a net negative. I don’t agree with it at all, but it seems to be the way the industry has been going for a while. And I kind of doubt that’s been limited to corporate IT either.
AI might make that worse, but it may make it better, in that there may be real value in hiring a kid out of college who has already internalized how to think along the lines of being productive with AI, rather than some thirty/forty-something who’s still learning how to think in the new age.
Here is an article from the current IEEE Spectrum by a person who used AI to generate papers which he then submitted to the open access journal (with permission of the editor.)
He used a standard LLM for one, and a system that uses AI to basically do the research for the second. Both got rejected, but that was because he did not change the AI generated papers in any way before submitting them. The first one got rejected because it used hallucinated references, easy to fix. The second got rejected because the references were real preprints, too many, and the AI made an obvious mistake in the chart.
The problem was not that the papers were trash.I wonder what would have happened if he had done minimal editing of the result.
Where I worked, a Fortune 100 company, there were dual hiring tracks, one for recent college grads and one for experienced people. It was easier to hire a recent grad. This was 10 years ago, before AI, I don’t know what the policy is now.
The companies not hiring new grads are being short sighted, but I suspect it is because you get much less pushback from not hiring than from firing.
I’ve been unfortunate enough to spend a lot of my career at places that are run on a shoestring for one reason or another. I got the distinct impression that they felt like reserving the capacity to hire and train up new grads was considered a luxury of sorts, when if they needed a mid-level database administrator, it was easier to just find someone looking with that exact skill set and have them start right away.
Where I work now that’s definitely what it feels like, and we’re feeling budget crunches because the state legislature passed some laws a while back (~2020-ish) that hamstring our ability to bring in money, especially in light of recent high inflation. So basically we’re already short-handed because of some hiring freezes put in place recently, and if we’re allowed to hire anyone, they’re wanting to get the most bang for the buck so to speak, and that’s not a recent graduate.
I’ll be short for fear of getting off topic, but bang for the buck depends on your industry. I worked in tech, and I hired people who had spent a bunch of years learning new things that we experienced people didn’t have time to learn. In a year or two I could see the average graduate knowing how to use AI better than the average experienced employee.
However, in competitive industries a new grad might come in at a higher salary than an experienced person, since you are in a bidding war for them. Raises for workers were often much lower than increases on new hire salaries.
That was true at least 30 years ago. Consulting firms like Accenture or Cognizant or the Big-4 are more or less jobs programs and outplacement services for recent undergrads and MBA. It’s not so much that they “take a chance” as they need a constant flow of cheap talent they can work to death at an inflated bill rate
Big non-IT corporations tend to not hire many recent grads for IT (and presumably other services) because that’s not their core business. They don’t have the training and infrastructure to train recent grads. They need someone senior who already knows their shit or the will outsource it to a vendor.
In fact, what I’ve observed is that much of these big corporations are just hierarchies of middle managers and executives with no actual skills. Most real actual work is outsourced to vendors, consultants, and contractors.
But for their core business, sure, they have entry level training paths. The big insurance company I work at hired a shit-ton of interns over the summer and they hire new associates. Most of these roles are in claims adjusting or underwriting (core insurance business functions).
I wonder if the AI future will just exacerbate this effect. Large corporations will continue to exist. But as most of the real work and analysis will be conducted by AI. So the humans who remain are just this weird leadership caste playing work through corporate theater. And I imagine those people will be funneled in through a higher education that has become even more unobtainable and unaffordable to the average person.
I think that uncertainty is the problem that stresses people out. What makes AI different from the tech boom is that the main selling point is how AI can do all these jobs, not all the new jobs it will create.
When I was starting my career in tech, there was huge demand for all these jobs as developers and project managers and integration experts and all these new skills and technologies you could learn. New companies were building completely new products and old companies were spending money to upgrade to new technologies and business models.
So what’s the equivalent with AI? Learning a new job as “ask the AI guy?” The only new jobs seem to be highly technical roles at a handful of big tech companies. Other than that, I mostly see a lot of bullshit “AI coach” jobs or people advising others to “be entrepreneurial” or “start their own business”.
But business doing what? More “coaches” and “consultants” or selling power bars in pyramid marketing schemes?
As you said, a lot of jobs require years of study and experience. And many of them only work in the context of a large corporation. Or at least initially. Like the company I work for is employing AI to help automate underwriting. So will underwriter be a job that disappears in a few years?
It’s disconcerting because there is no clear path to being successful and the hype around AI ranges from “it will be the greatest invention in human history” to “it will destroy society as we know it but give us a trillion dollars to build it anyway”.