One of my 2024 predictions was that Trump would push through the biggest tax increase in history, and that his anti-tax base would cheer it. (Deficit spending doesn't exist and tax increases aren't tax increases if a Republican is in office.)
I thought the scenario would be "we're going to abolish income tax and implement a national sales tax or VAT!" but then the abolishing of income tax part never happens and we just get income tax plus national sales tax plus VAT.
Instead he did it with tariffs. Don't know if it's the biggest tax increase in history but it's pretty sizable, and of course it's regressive.
People are angry about economic issues like low wages and unaffordable real estate but we've been beaten into learned helplessness on those things. Nothing can be done. Both parties are bad. The right promotes corrupt oligarchy and regressively transfers wealth upward, the left prevents new home construction (driving real estate costs) and regulates away everything but service industries and white collar jobs that AI is now replacing.
The worst AI hate I've seen comes from artists and creatives. I've found the AI hate in those communities to be white hot bordering on people talking about violence.
My early take, which I think is still valid, is that actual art is very unlikely to be replaced by AI. AI generated visual stuff looks bland, cliche, or has this weird "plastic" look. AI generated prose is boring. AI generated stories are hilarious barrages of tropes played exactly straight, cliche characters, basically just like paint by numbers bad TV writing or even worse. If real artists find ways to use AI, it won't be this way. It'll be as an assistant or they'll get in and hack it and make it do exactly what they want, much like artistic experimental photography.
But I'm only half right, and when I realized this I understood the hate.
AI is not replacing "true art," and it won't even if true artists end up finding ways to use it (like artistic photography etc.). But what it is replacing is what a lot of artists make their money doing: commercial art, making "content," ghostwriting, first-pass editing, graphic design, web design, that kind of thing.
That's not pour-your-soul-into-it capital-A Art, but it's what pays the bills. AI is absolutely decimating the market for that.
Yes, I strongly agree with these points, but particularly the second half. "Real art" is not at all under threat by AI, which has effectively no ability to even touch, let alone dilute the supposed magnitude and breadth of the human spirit in expression. The existence/proliferation of AI is not going to infringe upon the abilities or opportunities of a bus driver who comes home every day after work to work covertly on an entire house full of sprawling, elaborate tinfoil miniatures which aren't discovered until after his death.
The inconvenient reality of most "creative" work that AI has excelled at supplanting is that it already is, and has been, exactly what people accuse AI to be doing to the form: pure commercial churn, meant solely to occupy temporary space, occasionally brilliant, mostly by accident, but nearly always produced lazily, insincerely, and without pride by people who are not paid enough to care and would rather be working on anything else. AI accelerates human mediocrity not because it is inherently mediocre by nature but because most people, and the things they produce, are. And I very squarely place the bulk of all my own creative efforts and accomplishments within that category too.
I think "AI is replacing art" is a category error. If I read a book or look at a piece of art or listen to music, the whole point is to receive the thoughts, feelings, and lived experience of a human being.
But yes, AI is supplanting the mundane "art" that artists use as potboiler work, and that is the economic threat. The bitter horrible irony of art as a career is that the best art is often the hardest to sell, and it also takes the most time and energy to make. It's insanely hard to make a living doing just "pure art."
Much like self-driving cars being better than drunk, incompetent, or texting human drivers, AI software is generally better than the garbage churned out by code monkeys who don't care or cheap lowest bidder outsourcers.
It's nowhere near the quality of hand-crafted expert human code, but that requires a hand-crafting expert human engineer and takes a long time.
I predict that the latter will be reserved for the highest value, longest lived, or special (high performance, high security) parts of systems.
AI is awesome tech but it’s also to some extent mass piracy. The models are trained on huge amounts of material with dubious or non existent rights.
I have a hard time being concerned about “you pirated my piracy.”
I hold the view that many of these models should not be copyrightable. Anthropic and all the others talk about “safety” but you never hear them bring up attribution of the data that trained the model or compensation of anyone for it.
UI QA only works well if your model plausibly matches the average user behavior and/or real-world edge cases. These models are far from that, and they are much less random than you'd like them to be for fuzzing (mode collapse).
It doesn't need to be that kind of QA. Even just a basic "I want the AI to build the beginnings of a GUI app for me" will work much better if the AI can see the output of its work and iterate on it. Similar if you want the AI to fix a GUI bug—much better if you can show it the the bug and tell it how to test to see when it's gone.
Okay, fair, I haven't really paid attention to marketing.
> the LLM does not require computer use to see the GUI and
It can take screenshots without computer use, but it can't click around. I didn't have access to computer use until recently (I'm on an OS where Claude Code technically shouldn't run, I had to patch the binary), and when I got it working it made a big difference because of this.
I'm not sure I agree on your definition of legitimate...
If we want to say that any value generating - including crime - is inbounds, then LLMs are FANTASTIC scam machines. There are incredible uses for LLMs in a lot of stealing money spaces.
It’s true that things could be slow on those old machines, especially when it had to hit disk. It’s true that things were often single threaded and would stall.
But what’s also true is that this old code was orders of magnitude more efficient. Put that code on even a modest machine from today with an SSD, even a Raspberry Pi, and it would scream. Everything would be instantaneous.
Some of the reasons for this degradation are unavoidable, like high DPI displays and feature depth, but a lot of it is just bloat on top of bloat on top of bloat.
I’ve wondered if this belongs on the Fermi paradox pile. Many biospheres may be more massive planets that are so hard to get off that a space enterprise never starts.
Meanwhile lighter planets might have trouble holding onto atmospheres.
I thought the scenario would be "we're going to abolish income tax and implement a national sales tax or VAT!" but then the abolishing of income tax part never happens and we just get income tax plus national sales tax plus VAT.
Instead he did it with tariffs. Don't know if it's the biggest tax increase in history but it's pretty sizable, and of course it's regressive.
reply