Why do all “ai can’t write good code” posts from experienced developers always attract at least several “LLMs are not the only form of AI and AI is a term that has been around since 1950s” lectures, mostly from old dudes?
Is this the first time they encountered an “ai can’t code” post? Is it their very own mansplaining moment?
Why do they act like everyone and their grandmothers don’t have the faintest idea about terms such as AI, ML, vectors, tensors etc?
This fascinating analysis by @iris_meredith helped me understand something I’ve been wondering since my late teens:
Why computer nerds can’t understand that most people don’t use computers just to have used a computer.
Anyone know how many ‘ktobers there are to follow to get the full October fomo experience?
If people with 20-30 years of experience in their field say “AI” can do their job at similar or better quality than them, believe them.
Then question, analyze and find ways to fix the fucking systemic failures that result in such incompetent people keeping getting paid and promoted for multiple decades while higher educated unemployment is at crisis levels.
Nor’easter weekend treat!
Thanks @cstross
As always with The Laundry, I’m having a hard time reading this with Bob’s (entirely copied from Nick Frost) accent.
Capitalism working as advertised:
Consumers hating ai and opting not to use products made with it or that are related to it.
Capitalism as it actually is:
Capital owners not giving a fuck and shoving it in every orifice and inventing new orifices to shove it in like the insufferable dictators they are.
I want to reiterate.
These things need massive infrastructure. If it’s gonna go the same route, it also needs physical capabilities to manufacture and build stuff.
Blowing shit up is always way faster than producing shit.
Rogue AI technology getting out of control is physically impossible.
How come none of these AI doom scenarios never mention the efficiency requirements for them to work?
If exponentially self improving machine intelligence is possible, it either has to get exponentially more efficient and run on the same hardware, or it needs a way to build and install increasingly powerful and sophisticated infrastructure for itself.
I mean, as long as we can turn things off or blow things up faster than it can be built, which is an inevitability, then AI can’t simply kill us all.
I have a challenge for AI proponents.
Give me one thing, just one friggin thing that would be impossible or even just worse if we shut all LLM based services down and went back right to the moment before ChatGPT was released to public with all infrastructure exactly as it was.
I mean, people pick up a brick and throw it for much less.
Imagine genuinely believing someone is building the doomsday machine and going on holier-than-thou arguments on the internet about it instead of picking up the bricks and heading to ai bro headquarters.
If anyone was capable of building an "Artificial super intelligence" that would bring about the end of civilization, they wouldn't keep talking about it. They'd build it in secret and then unleash it.
If someone is openly saying they're building the doomsday machine, they aren't building the doomsday machine. There isn't a reality that can plausibly happen.
And let me tell you, if any of the doomers actually believed it, it would be trivial for a couple hundred people to organize in short order and physically retaliate at this stage.
Nobody believes chatbots are gonna cause the apocalypse, they're just enjoying the drama.
Also, shitlibs who are angry at trans folks for threatening their vote..
If you think trans people are a potent enough force to mess up your life by going scorched earth, you better take them seriously.
That’s how the capitalist hellscape game is played. You made your bed by choosing fascists over socialists, you lie in it.
When they say things like “we need to take control of AI sovereignty for the EU”, you aren’t part of that “we”.
They couldn’t give two fucks, tbh.
In capitalism, politicians will listen to the shareholder class and maybe the C-suite managers.
Anyone else doesn’t count so when they make declarations about the future and growth and stuff, you aren’t part of the conversation.
Unless they’re forced to, that is. You have to scare them more than they’re scared of the aforementioned ghouls.
With unlimited money and investor buy-in, you really don’t have to be smart to make impressive-looking things.
And don’t be fooled by the phds they have. I’ve spent enough time in academia to find out that these certificates really don’t mean someone is smart. They just mean a person is hard-headed enough to take abuse on rails for five to ten years.
I sincerely believe, after this last week, that the “AI researchers” working at openai and anthropic are stupider than the LLMs they train.
Why do xtians find "I'm the alpha and the omega" to be profound?
If a politician or someone claiming to be a divine entity today went "I'm the first and last letter of the Greek alphabet", I bet nobody would even consider it.
Today I heard the phrase “masterpiece of shit” and I thing it’s glorious.
“Is it even worth it to make art anymore”, and a thousand variations…
The correct question is, “was it ever _more_ worth it than now?”
TL;DR: art is about defiance
Arsetechnica, anyone?
