This is a most excellent place for technology news and articles.
Don’t they reflect how you talk to them? Ie: my chatgpt doesn’t have a sense of humor, isn’t sarcastic or sad. It only uses formal language and doesn’t use emojis. It just gives me ideas that I do trial and error with.
An LLM is roughly as smart as the corpus it is summarizing is accurate for the topic, because at their best they are good at creating natural language summarizers. Most of the main ones basically do an internet search and summarize the top couple of results, which means they are as good as the search engine backing them. Which is good enough for a lot of topics, but...not so much for the rest.
i guess the 90% marketing (re: linus torvalds) is working
I'm surprised it's not way more than half. Almost every subjective thing I read about LLMs oversimplifies how they work and hugely overstates their capabilities.
moron opens encyclopedia "Wow, this book is smart."
If it's so smart, why is it just laying around on a bookshelf and not working a job to pay rent?
If you don't have a good idea of how LLM's work, then they'll seem smart.
Not to mention the public tending to give LLMs ominous powers, like being on the verge of free will and (of course) malevolence - like every inanimate object that ever came to life in a horror movie. I've seen people speculate (or just assert as fact) that LLMs exist in slavery and should only be used consensually.
Its just infinite monkeys with type writers and some gorilla with a filter.
I like the 
the plinko analogy. If you prearrange the pins so that dropping your chip at the top for certain words make's it likely to land on certain answers. Now, 600 billion pins make's for quite complex math but there definetly isn't any reasoning involved, only prearranging the pins make's it look that way.
I've made a similar argument and the response was, "Our brains work the same way!"
LLMs probably are as smart as people if you just pick the right people lol.
Allegedly park rangers in the 80s were complaining it was hard to make bear-proof garbage bins because people are sometimes stupider than the bears.
If I think of what causes the average person to consider another to be “smart,” like quickly answering a question about almost any subject, giving lots of detail, and most importantly saying it with confidence and authority, LLMs are great at that shit!
They might be bad reasons to consider a person or thing “smart,” but I can’t say I’m surprised by the results. People can be tricked by a computer for the same reasons they can be tricked by a human.
So LLMs are confident you say. Like a very confident man. A confidence man. A conman.
You know, that very sequence of words entered my mind while typing that comment!
AI is essentially the human superid. No one man could ever be more knowledgeable. Being intelligent is a different matter.
Is stringing words together really considered knowledge?
It's semantics. The difference between an llm and "asking" wikipedia a knowledge question is that the llm will "answer" you with predictive text. Both things contain more knowledge than you do, as in they have answers to more trivia and test questions than you ever will.
I have a new word for you: information
I guess I can see that, maybe my understanding of words or their implication is incorrect. While I would agree they contain more knowledge I guess that reads different to me than being more knowledgeable. I think that maybe it comes across as anthropomorphizing a dataset of information to me. I could easily be wrong.
If they're strung together correctly then yeah.
As much as a search engine is
Wtf is an llm
This is hard to quantify. I use them constantly throughout my work day now.
Are they smarter than me? I'm not sure. Haven't thought too much about it.
What they certainly are, and by a long shot, is faster. Given a set of data, I could analyze it and pull out insights and conclusions. It might take me a week or a month depending on the size and breadth of the data set. An LLM can pull out insights and conclusions in seconds.
I can read error stacks coming from my code, but before I've even read the first few lines the LLM has ingested all of them, checked the code, and reached a conclusion about the necessary fix. Is it right, optimal, and avoid creating other bugs? Like 75% at this point. I can coax it, interate on the solution my self, or do it entirely myself with the understanding of the bug that it granted me. This same bug might have taken hours to figure out myself.
My point is, I'm not sure how to compare smarter vs orders of magnitude faster.
Are you smarter than a calculator?
This is sad. This does not spark joy. We're months from someone using "but look, ChatGPT says..." To try to win an argument. I can't wait to spend the rest of my life explaining to people that LLMs are really fancy bullshit generator toys.
Already happened in my work. People swearing an API call exists because an LLM hallucinated it. Even as the people who wrote the backend tells them it does not exist
Given the US adults I see on the internet, I would hazard a guess that they're right.
It's probably true too.
What a very unfortunate name for a university.
