Not really — even if AGI doesn’t work and these models don’t get any better, there’s still enormous value to be mined just from harnessing the existing state of the art.
Or these guys pivot and go back to building CRUD apps. They’re either at the front of something revolutionary… or not… and they’ll go back to other lucrative big tech jobs.
Is there enormous value? AI is burning cash at an extraordinary rate on the promise that it will be an enormous value. But if it plateaus, then all the servers, GPUs, data centers, power and cooling and other infrastructure will have to be paid for out of revenue. Will customers be willing to pay the actual costs of running this stuff.
I don’t know if what they’ve built and are building in the future will justify the level of investment. I’m not an economist or a VC. It’s hard to fathom the huge sums being so casually thrown around.
All I can tell you is that for what I use AI for now in both my personal and professional life, I would pay a lot of money (way more than I already am) to keep just the current capabilities I already have access to today.
I’m a software engineer so for work I use it daily. It doesn’t “do my job” but it makes my job vastly more enjoyable. Need unit tests? Done. Want a prototype of an idea that you can refine? Here. Shell script? Boom. Somewhat complicated SQL query? Here ya go. Working with some framework you haven’t used before? Just having a conversation with AI about what I’m trying to do is so much better than sorting through often poorly written documentation. It’s like talking to another engineer who just recently worked on that same kind of problem… except for almost any problem you encounter. My productivity is higher. More than that, I find myself much more willing to take on bigger, harder problems because I know there’s powerful resources to answer just about any question I could have. It just makes me enjoy the job more.
In my personal life, I use it to cut through the noise that in recent year has begun to overwhelm the signal on the internet. Give me a salmon recipe. This used to be the sort of thing you’d put into Google and get great results. Now first result is some ad-stuffed website that is 90% fluff piece and a recipe hidden at the bottom. Just give me the fricken recipe! AI does that.
The other day I was trying to figure out whether a designer-made piece of furniture was authentic despite missing tags. Had a back and forth with ChatGPT, sharing photos, describing the build quality, telling it what the store owner had told me. Incredible depth of knowledge about an obscure piece of furniture.
I also use the image generation all the time. For instance, for the piece of furniture I talked about, I took a picture of my apartment, and the furniture, and asked it to put the furniture into my space, allowing me to visualize it before purchase.
It’s a frickin super power! I cannot even begin to understand how people are still skeptical about the transformative power of this stuff. It kind of feels like people are standing outside the library of Alexandria, debating whether it’s providing any value, when they haven’t even properly gone inside.
Yes, there are flaws. I’m sure there’s people reading this about to tell me it made them put glue on their salad or whatever. But what we have is already so deeply useful to me. Could I have done all of this through old fashioned search? Mastered Photoshop and put the furniture into my apartment on my own? Of course! But the immediacy here is the game changer.
I consider myself an engineer more than a “coder” - I solve problems. I don’t really care as much how. I do not enjoy code for its own sake, I enjoy a problem well-solved.
Software engineering turns out to be a way to solve problems and get paid very well.
I’d just as soon do some kind of other problem solving if it was interesting and paid as well.
If performance tuning a SQL query is your idea of a good time, don’t let me yuck your yum. I just need the data to solve whatever problem is in front of me that day.
I'm not trying to make a point, just curious -- what's stopping you from spending more money on AI? You could be using more API tokens, more Claude Code and whatever else.
I have a ChatGPT subscription, and work has one of those “all the models” kind of subscriptions. So I have access to pretty much most of the mainline models — don’t feel the need to pay more.
But if the business model collapsed and they had to raise prices, or work cheaped out and stopped paying for our access, then yeah, I’d step up and spend the money to keep it.
They've so far spent about what the world spent to build out almost all of the broadband internet, the fiber, cable, cellular, etc. If AI companies stop now, about 10 years after they got going, does their effort give us trillions of dollars being added to the economy each year from today forward, like we got for every year after the 10 years of internet build out between 1998 and 2008? I'm not seeing it. If they stop now, that's a trillion dollars in the dumper because no one can afford to operate the existing tech without a continual influx of investor cash that may never pay off.
Or these guys pivot and go back to building CRUD apps. They’re either at the front of something revolutionary… or not… and they’ll go back to other lucrative big tech jobs.