70 IQ CEOs realized AI could write code faster than humans, so they got rid of humans to have AI vomit out thousands of lines of randomly generated code and now we keep having outages in random things.
Anonymous A (OP) double-posted this , 6 minutes later[^][v]#1,473,716
Now the new thing is they’re saying that we should expect more outages because the AI is getting so good at fixing code there will be more failures due to software patches.
Anonymous B joined in and replied with this , 2 hours later, 2 hours after the original post[^][v]#1,473,747
Upper management is often dense. Because at big companies they’re often insiders, time servers, and jump on a trend they read in the NYT so their subordinates rush to anoint a half-baked vision.
It’s interesting to see someone, say the managing partner of a law firm or a traditional CEO, then interact with a company founder. All are intelligent but the founder/entrepreneur is more informed.
Anonymous D double-posted this , 7 minutes later, 6 hours after the original post[^][v]#1,473,785
Don’t get me wrong, LLMs are useful. They generate text. I don’t think we should be anthropomorphizing cloud software and calling it "super intelligence."
Anonymous D double-posted this , 1 minute later, 6 hours after the original post[^][v]#1,473,805
You know when I was born, computers could add, multiply, divide, subtract faster than any human alive? And they could do that ten years before I was born, and twenty years before I was born, and thirty years before I was born, and forty years before I was born?
Anonymous D triple-posted this , 1 minute later, 6 hours after the original post[^][v]#1,473,806
When it comes down to it right, LLMs train on data, then they output similar data. You train it on more data, it can do more stuff. I’m not going to start calling it super intelligence because that’s dumb.
Anonymous D double-posted this , 1 minute later, 6 hours after the original post[^][v]#1,473,812
LLMs in my view are not intelligent. AI companies really want us to think that LLMs are intelligent, so they’ll try and do certain things to generate hype, but then later when the details start coming out, and I learn what actually happened, I just don’t believe in the ideology surrounding this technology. I don’t believe in AGI or self improvement or super intelligence or even artificial intelligence actually.
Anonymous C replied with this , 13 seconds later, 6 hours after the original post[^][v]#1,473,813
@1,473,811 (D)
If it's smarter than a normal person and a trained computer security professional, and we can call them "intelligent" than it fits to call this "superintelligent".
Anonymous C replied with this , 8 seconds later, 6 hours after the original post[^][v]#1,473,815
@1,473,812 (D)
It's already proven it's more intelligent than a normal person, it can have more coherent and advanced spoken conversations than most people.
Anonymous C double-posted this , 49 seconds later, 6 hours after the original post[^][v]#1,473,821
@1,473,818 (D)
I never tried to learn differential equations, but a math Ph.D. would probably be considered smarter if we both talked about our qualifications.
Anonymous D replied with this , 1 minute later, 6 hours after the original post[^][v]#1,473,835
@1,473,832 (C)
So I want you to think about something here: AI improves the more data it’s trained on. The frontier models are several terabytes in size. If AI is intelligent why would it be that we need to train AI on more data to make it seem smarter?
Anonymous D replied with this , 1 minute later, 7 hours after the original post[^][v]#1,473,840
The way that OpenAI solved Navier Stokes for example, is they had 10,000 AI agents running in parallel for 88 hours generating math proofs and they spent $6 million on this (for a $1 million math problem btw but that’s beside the point it’s just a little ironic).
The thing is that programming languages have been invented to express mathematical proofs, and these proofs can be verified automatically by computers. So the way AI solves these proofs is by guessing over and over again, and every time the AI guesses at a proof, a deterministic algorithm will attempt to verify that proof, until a solution is found. That’s a technique called brute forcing and it does not require or imply intelligence or understanding of anything.
Anonymous D double-posted this , 2 minutes later, 7 hours after the original post[^][v]#1,473,841
Now OpenAI to solve Navier Stokes, the AI agents wrote 2.7 million messges. That means out of more than a couple million attempts at a solution, just one of those solutions was correct.
Anonymous C replied with this , 1 minute later, 7 hours after the original post[^][v]#1,473,842
@1,473,840 (D)
That's what humans do too, they come up with some new idea and then ask if it makes sense, over and over.
You could protest, but you'd never have any proof of something that humans could do that robots can't. It's all a feeling, it's your faith. This isn't a scientificly derrived idea.
Spare me the multiple posts of vagueness and insults, if you have a real test you can share it.
Anonymous D replied with this , 1 minute later, 7 hours after the original post[^][v]#1,473,843
@previous (C)
No it’s not. When you solved a math problem in school, you didn’t hand your teacher 2.7 million answers and then get an A because one of them was right.
By the way my point was it’s not unique to LLMs because deterministic software written by humans can also solve math problems step by step and has been able to do this for years.
Anonymous C double-posted this , 1 minute later, 7 hours after the original post[^][v]#1,473,852
@1,473,850 (D)
I already said once I'm not talking about LLMs specifically. Most people using these use the visual model, speech, and even agentic stuff for beaurocratic tasks.
Anonymous D replied with this , 24 seconds later, 7 hours after the original post[^][v]#1,473,853
@1,473,851 (C)
Not necessarily, I think animals possess intelligence. I don’t think that computers do. I think you’re anthropomorphizing inanimate objects because of all the marketing and feat mongering that AI companies have flooded the internet with.
Anonymous D replied with this , 17 seconds later, 7 hours after the original post[^][v]#1,473,860
You’re really convinced that LLMs are intelligent. I don’t feel like I’m trying to defend my belief that they’re not. I don’t feel any doubt you’re the one who keeps rejecting what I say when I give explanations for why LLMs aren’t actually intelligent even though they might appear that way at face value.
Anonymous D double-posted this , 1 minute later, 7 hours after the original post[^][v]#1,473,861
If LLMs are intelligent, what do you think the consequence of that is? If they are intelligent, why do you think that’s something that should be known and defended? Is there a moral consideration attached to your position on this?
Anonymous D replied with this , 31 seconds later, 7 hours after the original post[^][v]#1,473,864
@1,473,862 (C)
To be clear, I’m not claiming that a "machine" can’t be intelligent. I’m claiming that these machines in particular are not intelligent. What counts as a machine and what doesn’t count as a machine in a way is a bit arbitrary. For instance some people describe proteins as "molecular machines." That said, most people don’t mean human beings when they say "machine" we’re talking about classical computers that use silicon chips. I don’t believe that LLMs are intelligent.
Anonymous C replied with this , 13 seconds later, 7 hours after the original post[^][v]#1,473,866
@1,473,864 (D)
Computers as they are today can speak, listen, write, make art, code, and move limbs to navigate their environment. They have an education well beyond what a typical human ever gets.
Anonymous D double-posted this , 1 minute later, 7 hours after the original post[^][v]#1,473,869
If you believe that humans will be replaced by computers and you believe that I’m telling lies when I say that computers are not intelligent, then why do you think that the lies I tell to humans make any difference if humans will be replaced anyway?
Anonymous D replied with this , 30 seconds later, 7 hours after the original post[^][v]#1,473,872
If you want to know my actual take on this, I do not believe that computers are conscious so I am entirely concerned with the wellbeing of humans. I think that "AI" will have an impact on society, but what matters is not the wellbeing of the robots.
Anonymous C replied with this , 1 minute later, 7 hours after the original post[^][v]#1,473,881
@previous (D)
At first I just added a note that it's not scientific, and then that would always send you into a rage where you refused to ever admit it was true.
Anonymous C replied with this , 2 minutes later, 7 hours after the original post[^][v]#1,473,888
@1,473,886 (D)
If I refused to say it for multiple threads and I called some stupid for saying he wasn't, then you would get the impression that I thought he was.
If I said he's not a Disney princess the first time it came up, it wouldn't be a reasonable conclusions. Here: He's not a Disney princess.
Anonymous D replied with this , 49 seconds later, 7 hours after the original post[^][v]#1,473,904
@previous (C)
Yeah see, you’re rejecting the scientific consensus and all of modern genetic research in favor of one specific study. That’s not scientific.
Anonymous C replied with this , 23 seconds later, 7 hours after the original post[^][v]#1,473,905
@1,473,901 (D)
You're strawmanning something again, probably the whole "there are no clear dividing lines between discrete races" or some BS that I never said.
We've been through race more than anything else and you just repeat the same old lines. It would probably take 10 posts just to get you to shed some light on what claim you're pretend I made.
Anonymous D double-posted this , 58 seconds later, 7 hours after the original post[^][v]#1,473,910
For instance in this thread, you didn’t actually engage with anything I said, you just kept responding with "not scientific." You didn’t bother actually coming up with any kind of real counter argument to anything I said.
Anonymous D replied with this , 55 seconds later, 8 hours after the original post[^][v]#1,473,916
@1,473,914 (C)
Most people make money by working to earn a salary. You think AI will replace people but it will not replace people. It will only replace the economic system.
By most definitions of capitalism, that's still a capitalist system even if they can't find work. I don't know what you call this new state of affairs.
Anonymous D replied with this , 11 seconds later, 8 hours after the original post[^][v]#1,473,922
You should look into opinion polling on American’s views on AI and data centers. It has not led to massive job loss yet, but the public perception of AI is already negative. If AI does lead to massive wealth inequality and job loss, it will lead to a spike in populist economic ideas.
Anonymous D double-posted this , 7 minutes later, 8 hours after the original post[^][v]#1,473,923
For instance, look what happened in Mississippi and Texas. That’s not normal. Of course there are a lot of other factors that led to that, but if you tell everybody AI is going to take their jobs, people might start reacting to that.
Anonymous D double-posted this , 1 minute later, 9 hours after the original post[^][v]#1,473,965
Kind of a weirdo actually. You hate women, love AI and think it’s sentient, hate black people, and think rich people are always right. It’s like if you took the dorkiest traits and mashed it up into one psychopath.
Anonymous G triple-posted this , 1 minute later, 9 hours after the original post[^][v]#1,473,982
All I’m saying is you can’t make, "you people are just lazy you don’t want to work" and "you’re literally worthless AI will replace you lol" work at the same time. You can’t win an election on both those things together at the same time. It doesn’t work.