Post

Prior tech booms, even the NFT Moment, did seem to have more of a collective sense that an opportunistic normal person could cash in on them (and some did). Now, it *feels* like only the employees of a very small number of companies and their investors are actually *profiting* from the AI boom.

63 reposts598 likes

Hence, the half-hearted effort to discuss a Basic Income structure. They know they are trying to destroy peoples' lives, our economy, & our society. But anyone who questions them or the hack-handed way they work are treated as "Luddites"

The argument seems to be more that “these tools will help normal people build stuff that they can make money off of,” but where’s the evidence that this is happening? People ARE using LLMs to code Stuff, often quite cool stuff - but is this making them money? Idk?

It won’t make money because, by its nature, the advantage of these things is to be able to build tools for idiosyncratic niches that normally couldn’t justify a software development team to build. A bunch little toy apps with an addressable market of a few dozen people isn’t gonna make anyone rich.

It’s very cool. It certainly adds a ton of value to the world that people can do this for themselves. In theory it adds a bunch of value to GDP creates a big consumer surplus. But it doesn’t translate to money in anyone’s pockets.

For every guy who has "made it" from this new paradigm, there's 10+ who lost their job. After I lost my job of 3 years in devops after 8 years in support, I became a helpdesk manager and when I hired for a *temp help desk entry* I got at least 10 apps from former devs with 15+ years in the field.

Feels like tech has long had the baked in assumption that people making stuff will = Money for at least some of them (with good timing and luck) - and it at least FEELS (again) like this is breaking down rapidly.

The disconnect I think is that the form is now amenable to being completed so much faster than the substance. The amount of “I vibed up this cool player comparison dashboard” “Ok what data are you using for the comps?” “Oh that’s just dummy data for the demo, not my job for the rest” I’ve seen…

It also FEELS like companies do not have the foggiest idea how to hire or fund people who have become genuinely good at using The Robot

Based on talking to friends in various industries, that's entirely true, but moreover, management seems willing to literally stake the existence of their company on a technology they don't understand, so they won't be "left out". Insisting this stuff be integrated with everything.

In the long wake of the Great Recession, the classic tiresome, pat answer to people who couldn’t find a decent job was “learn to code.” Now, do we have a similarily oversimplified yet sorta-plausible answer for people who no longer can find stable employment? I don’t think we do. That’s unsettling.

I see a lot of software getting better, but I haven’t seen anything I want to give *my* money to. Anything I want I can just build myself or find someone giving it away for free.

Yeah - I think what's happening is that LLMs genuinely make software engineers who know how to use them more productive, but this doesn't translate to higher pay, just more stuff to work on and greater profits for the corporation.

I’ve made (and little) money with AI (and I started before AI) but it’s a long road that requires a lot more than just building. AI can’t replace a lot of it (networking, sales, etc). If I had what I have 5 years ago I’d be a billionaire. So the diff is more in how (and how much) money is made.

IMO the opportunity has shifted from a national one with VC backing to a regional one finding small to medium businesses that have processes to improve.

I got my first ever job as a Agent Orchestration Engineer. Setting aside the money which has massively reduced the precarity of my unsafe living situation, this job has allowed me to work with kind, compassionate, intelligent, insightful, amazing people who make me a better person by working with me

It certainly makes my job easier as a teacher, all my bs metawork tasks, leveling texts to different lexile levels, creating multiple versions of tests, the loads of narrative materials I need for gamification It's like having a C level teaching assistant which was never going to happen otherwise

Somebody's out there making money selling shit with AI-generated "photos" of products that don't actually exist. At least judging by all the "this is what I ordered, this is what I got" stuff. But I don't think that's a positive development for humanity.

Even for huge corporations, are they actually saving more money than they spend on the tokens? They don't seem to be doing anything that would bring in billions in new revenue with AI.

It's been really nice for skilled tradesman in the construction industry. Their pay rates have increased $20 per hour in many cases. I'm being pedantic here, of course. These guys are "cashing in" in as much as they can buy a house. So at least there's that.

This requires more typing than I have got in me right now, but I asked to get off the AI insider train. There's pervasive anxiety in the AI biz, and they grind. The work is often deeply stupid, everyone tripping over each other to build things that don't matter a month later. Aware of my privilege.

Not disagreeing with your observation, but rather extending it much of the AI hype bubble itself. There's hyperwealth sloshing around for a few hundred people but every 2022 unemployed eng is now a AI expert and it's a knife fight.

Speaking of hyper wealth and investing, both venture capital and private equity seem profoundly limited in their imaginations about interesting things to invest in in this new world

Their suppliers are profiting too. It is fascinating to see the shift of where the money is in Si Valley from software to hardware. The problem is that this like every other boom, it will bust, and there will be a very large amount of excess compute.

The buildout of fiber for the commercialization of the Internet that went bust in the dot-com wipeout, left behind excess capacity that laid the seeds for the substitution of cable TV with streamed content. What will innovators do with excess GPU-centric compute?

IMHO it's telling that even S&P 500 companies have difficulty pointing shareholders to how, exactly, AI is "transforming" their businesses. To the extent they say anything, it's basically what you'd expect: we can use it to cut some costs (presuming getting more work out of fewer people).

apollo.comFrom Who Is Deploying AI to Who Can Prove the ROI | The Daily SparkData from the S&P 500 second quarter earnings season shows that AI adoption keeps broadening: 69% of companies now point to a live deployment, up from 64% last quarter, see the first chart below. ...

Also interesting to see someone like Cory Doctorow point out that the big AI layoffs tech companies did were probably more just "typical" layoffs, in that there was a huge overhiring that happened during Covid and now this needed to happen. (signed, person in video game industry)

definitely, was thinking about this earlier today after reading amodei's recent thing

David Kaye@davidakaye.bsky.social

but perhaps the thing that most undermines his claim is a failure to acknowledge the basic inequity of AI development concentrated in the hands of a small number of people who are becoming unaccountably wealthy & further detached from the reality of how their tools impact human life.

Consider a career in infosec. Companies all need it. Big AI companies *Definitely* need it. And more scams => more need for defense. And using AI tools for defense is every bit as viable as for offense, and clever application of same can yield good results. Have hoodie/techno, will travel.

There were (not-even-AI) related layoffs at scale at most major tech companies, and I really do think these layoffs were barely or not even targeted at all. I don't want to anecdata thing but I get regular pings for a linkedin profile I haven't updated in ten years, which I regarded as a sign.