HN user
Wow it's almost like modern datacenters use way more power due to GPU consumption. Emit way more pollution. Use way more land. Make way more noise. Produce slop instead of functioning products. Do yall use your brain before chiming in and misinforming everyone?
Lol the entire pro-data US center is ran by elon musk and the other cronies that bribed the Trump admin to let them violate the clean air act. They use gullible losers that will swallow their propaganda whole and promise them "AGI."
Yes, the AI Datacenters are much worse for the environment. xAI is using methane-turbines to power theirs, and they are allowed to pollute your communities because they bribed the Trump admin.
https://www.theguardian.com/technology/2026/jun/16/elon-musk... https://www.canarymedia.com/articles/data-centers/elon-musk-... https://www.facebook.com/watch/?v=1014655584252361
nice try clammy sammy.
We're gonna continue to push for scientific based approaches to education not some lunatic's idea of an educational experiment that waste's children's education. We tried this with Bill Gates already.
I mean that's this entire thread in a nutshell. Armchair experts emboldened by their $100 ChatGPT University Diploma.
No, the rationality of humans is not defined by how gullible they are towards LLMs. Are yall getting your psychology degree from ChatGPT university, my god.
All AI is flawed and prone to "hallucinations"(doing exactly what it was designed to do) that's why Microsoft considers it an "Entertainment" product.
"this is akin to givin someone a textbook on an obscure subject that has certain factual errors." My brother all LLMs give factual errors so, no this is not a problem with the study. In your fake experiment you are hypothesizing a 100% factual LLM which does not exist.
"This study tested none of those" So the study is bunk because it didn't test your favorite LLM flaws?
What capabilities? Regurgitating a collage of training data? Again it's pretty easy to understand why this happens. The "magic" you are perceiving is language itself not these models. Language is a constantly evolving tool and we constantly imbue it with our collective knowledge.
Yes. When someone is teaching you language as a baby you have eyes and ears and 100 million nerve endings and constant "training" to understand language. You cannot understand the word "semantics" as a human without this crucial "training" step. My god you people will pull out the most bs metaphors to justify being in awe(and debt) to a software program.
a bone box.. with eyes and 100 million nerve endings to experience the world...
We know exactly how LLMs work. you don't understand how they work and so you think they are free of physical limits like scaling. There are several rudimentary articles on hackernews for you to peruse to understand weights and scaling. IMO the "magic" derived from LLMs comes from one of humanity's actual greatest inventions and constantly evolving tools: language.
The last part is a bigger issue for me. I can use my usb to charge most of my tech but when it comes to the exact cable that my interface needs to use I now have like 30 cables I have to try before I know which one was the correct one.
They have no preferences, LLMs spit out what's in the training data or what the filters tell it to do.
That's exactly what giant train corporations thought. "We own all the railways, we've squashed the competition"
and they STILL went out of business because they over-estimated the demand for their shitty rails they built to the middle of nowhere. Same with "AI."
doesn't matter. that just means they've incentivized all competitors to enter the market and let's be honest none of their tools are that novel.
google might have tons of integration. But if it invested too heavily into AI then it will also suffer when increased competition causes returns to fall:
Step 2. Rising competition causes returns to fall below cost of capital.
Thanks and apologies for the grief. I have replied at least to the parent comment from this thread to correct myself. I mentioned it I think two previous times and I didn't realize you couldn't reply to old comments. I'll send an email
I was incorrect there was no AMA, the soldier was in the comments of a thread about israel
My mistake you are correct. I will edit my previous comments
Also, the mods were heavily in the thread and flagged most of my comments calling the soldier a war criminal. I wonder how many ex-IDF have worked at ycombinator companies.
It was on the front page(I want to say the January after Oct 7), and has since been scrubbed along with the many comments I made on it.
Here is me referencing it about 11 months ago and I have older comments that reference it. https://news.ycombinator.com/item?id=44864419#44866331
I guess HN likes to scrub things that make them look bad instead of confronting the reality that they helped launder and "humanize" genocide.
It is theft.
"intellectual property does not exist" -> goes on to define various legal instruments for protecting intellectual property.
"you can't steal something if the original owner has it the whole time" Is this fifth grade? So if I go over to your house and trash it, it doesn't matter the house was under your possession the whole time.
No the rebalance is to sue these companies into the dirt.
There was a prominent AMA for an israeli soldier while stationed in Gaza oppressing brown minorities and blowing up their children. The world sees what Israel has done it's not an "online" thing there are real humanitarian orgs and the UN accusing Israel of genocide.
"anybody that doesn't agree with me is a cave-dwelling conservative"
yeah keep chatting with the sycophantic psychosis-inducing LLM chatbot, I think you'll have a better time.
"i'm 14 and this is deep"
LLMs are mass theft of intellectual property. I didn't need them to "drive" this computer for the last ten years.
Investors want to see their investment grow. What is the conspiracy? HN also hosted a likely war criminal to talk about his experiencing killing children in gaza. You think these people have morals or ideals?
having a large test suite does not equal to testing every potential edge scenario that has never broken before in production.
So the measure of a model is how well they can recreate something they easily have thousands of examples of in their training data. There's probably a better base RTS on github somewhere for free.