Fuck reddit
this post was submitted on 23 Apr 2026
759 points (94.3% liked)
Fuck AI
6809 readers
3208 users here now
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
founded 2 years ago
MODERATORS
The grifters want to become "too big to fail", so that a "pause" would cause finance market drama. So, first, block the sale of the shares -- that's needed to avoid the grifters offloading to bigger fools.
A lot of people think that these companies want to become too big to fail but I suspect something else is going on as well.
Step 1. Make/pioneer new tech or buy it.
Step 2. Get investors onboard with it so that you can get a lot of seed money.
Step 3. Invest in the infrastructure to support widespread use of the technology.
Step 4. Develope the technology even though it has known flaws.
Step 5. Let those known flaws stir the pot and cause uncertainty in the market.
Step 6. If the technology fails and the bubble pops, get government bailout.
Step 7. Use money from bailout to buy up the infrastructure and components at a ridiculously rock bottom rate using these companies own money.
Step 8. Use a combination of bailout money and insurance to make the investors whole (or as whole as their contract stipulates).
Now at step 9 you are left with a bunch of commodities that tech companies need (data centers, supply lines for components, better power infrastructure, a knowlegable tech work force etc). And you don't have to pay astronomical prices because the bottom fell out of the market.
Wouldn't doug judy's "name one bad nurse" scene have been a way better fir for this meme?
Two major flaws is none of those examples had the potential to be highly profitable. And all three of those had major ethical and morality concerns that the general public wouldn't approve.
AI is being packaged and sold as a toy, so while people do object to it, it's not the same scale of playing god or literal war crimes.
Two major flaws is none of those examples had the potential to be highly profitable.
Neither does so-called "AI". OpenAI, Anthropic and the rest are burning money like it's nobody's business.
On the corporate side of things, it's being sold as a money maker. Costs a fraction of what an experienced employee does and marketed as less prone to mistakes or bungles. Obviously not true but since when has a CEO ever turned down a promise too good to be true?
That grift can't work indefinetly, though. At some point the piper will have to be payed.
Microsoft is already making billions from copilot licences and it's the same for Claude and Gemini and the others, just because they're also spending fuck tons of money building data centres to expand their AI business doesn't mean they aren't making billions from it already.
Nope, the cost of inference is still way too high to actually be profitable.
Pause? Don't you mean undo?
Sora
They "paused" that because it was way too expensive to run and it's popularity cost them too much money and scared investors.
I do not see AI going away I see it being the tool they originally pitched before firing everyone as an assistant for coding and the such and that only when running locally. I think it will do a decent job in finding and patching 0 day issues in Linux kernels and the such but Hollywood and the tech sector after making a really tough few years for all their former employees will be having a very tough time themselves, and if the US government has any balls they will have to worry about being broken up as monopolies and prosecuted for their criminal behavior, but that seems to be wishful thinking even if a more responsible party get all the levers of power.
Not too much to research with the blinding lasers needed. Keep in mind I am not any where near an expert but this seems like a sufficient implementation. I am 100% certain someone smarter and or more experienced than me could potentially simplify this further.
A sufficiently high resolution video camera with fast integrated auto focus on a mechanized cradle, a computer with two RTX 4090's, and of course a class IV (say 100w) laser with dual rotating mirrors should be enough to permanently blind the bulk of a crowd.
I bet there is a free model on hugging face for identifying the pupils for eye tracking and there are facing tracking models to drastically cut down the area for acquiring targets. Really the hardest part could be the camera but if you have enough money, Hollywood has stuff that would work. Other problem would be running the models fast enough to be effective.
Casing could just be a metal box that rotates and has a quartz glass window coated with hydrophobic chemicals.
But you can buy #2 on eBay