Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Grok3 Launch [video]

x.com

Post reply on HN
Print view View original

611–620 of 1001 posts

Re: Grok3 Launch [video]

#611

Post by panphora »

djaychela wrote: post #219
phtrivier wrote: post #169

Off topic, but just in case: is there a good reference on how people actually use LLMs on a daily basis ? All my attempts so far have been pretty underwhelming: * when I use chatbots as search engines, I'm very quickly disappointed by obvious hallucinations * I ended up disabling github copilot because it was just "auto-complete on steroids" at best, and "auto-complete on mushrooms" at worst * I rarely have use cases…

My wife has found ChatGPT extremely useful when dealing with her mother - who has bipolar and is obsessed with other people's health. I've got a terminal cancer diagnosis and handling my mother-in-law has been extremely difficult - nearly to the point of no longer having any communication with her. ChatGPT has a single conversation with all the back story and has put some useful points across when discussing how diff…

You might find this optimal conversation path finder app of interest: https://x.com/eddybuild/status/1889908182501433669

Re: Grok3 Launch [video]

#612

Post by esafak »

poulpy123 wrote: post #345
phtrivier wrote: post #169

Off topic, but just in case: is there a good reference on how people actually use LLMs on a daily basis ? All my attempts so far have been pretty underwhelming: * when I use chatbots as search engines, I'm very quickly disappointed by obvious hallucinations * I ended up disabling github copilot because it was just "auto-complete on steroids" at best, and "auto-complete on mushrooms" at worst * I rarely have use cases…

LLM are good at one thing, and totally by chance it is the thing they have been designed to be: be a word probability generator. If you can constrain your usage around that, they are great to use. But the people who think they can reason or know some kind of truth are delusional

Explain how o3 won a gold medal at this year's International Olympiad in Informatics, or provide your benchmark for reasoning.

Re: Grok3 Launch [video]

#613

Post by diggan »

KerrAvon wrote: post #597
InTheArena wrote: post #590

Earlier quoted context omitted.

Elon's unchecked power at building a model? Or at politics? I always worry whenever I see people telling me how to feel - rage in this case. We are in a political system that is oriented more around getting people to feel rage and hatred as opposed to consensus and deliberation. Elon is the face of that, but it's a much longer and larger problem. Throw in the complete dismisal that anyone not scared of this is ignora…

Has it occurred to you that the people who feel rage fundamentally understand the situation, and you may be undereducated in this area? What do you think are the root causes of that “senseless partisan drive”? I’d suggest starting with Rick Perlman’s book “Nixonland” if you’re interested.

> Has it occurred to you that the people who feel rage fundamentally understand the situation, and you may be undereducated in this area?

Regardless of how justified the rage is or not, being very emotional about things usually have one of two effects on people A) people gets taken aback by someone's strong emotions or B) people get inspired/taking with the emotion, even feeling that emotion stronger themselves. Sometimes also C) they don't really care either way.

What probably isn't helpful, is calling someone is "undereducated" when they're clearly saying that they're person (A), and just because they may or may not agree with you (although parent didn't even clearly say they disagree, just that they're "taken aback" a bit).

Some people are calm regardless of what's going on around them, even if the world would be on fire, they'd try to describe what's going on around them with careful words and consideration. It isn't wrong or right, just like the people who feel rage and very emotional aren't wrong or right, it's just a showcase how we're different.

But we should aim to at least understand each other, not by trying to provoke, instigate or look down on others, but by asking questions to clarify and to better understand.

Re: Grok3 Launch [video]

#614

Post by throwaway657656 »

smeeger wrote: post #542
_heimdall wrote: post #530

Earlier quoted context omitted.

I'm not sure how LLMs/AI couldn't consolidate power. By design, it will move power from the individual to those running the AI systems.

bingo. and everyone is fine with it as long as the consolidation is benefiting their own tribe.

Definitely not everyone. For many of us, checks and balances are a feature not a bug.

Re: Grok3 Launch [video]

#615

Post by yodsanklai »

phtrivier wrote: post #169

Off topic, but just in case: is there a good reference on how people actually use LLMs on a daily basis ? All my attempts so far have been pretty underwhelming: * when I use chatbots as search engines, I'm very quickly disappointed by obvious hallucinations * I ended up disabling github copilot because it was just "auto-complete on steroids" at best, and "auto-complete on mushrooms" at worst * I rarely have use cases…

I used ChatGPT all the time for

1. Small coding tasks ("I want to do XYZ in Rust"), it has replaced stack overflow. Very convenient when writing code in a language I'm not super familiar with. 2. Help with English (traduction, proofreading...) 3. Learning something, like tech, I like interacting with it by asking questions, it's more engaging than just reading content.

I'd say nothing is game changing, but it's a nice productivity boost.

Re: Grok3 Launch [video]

#616

Post by jessfyi »

sigmoid10 wrote: post #385
CSMastermind wrote: post #49

Karpathy gave his initial impression: https://x.com/karpathy/status/1891720635363254772 The pull quote is: The impression overall I got here is that this is somewhere around (OpenAI) o1-pro capability

The impression seems to be warranted: Grok 3 has directly jumpted to the top of all leaderboard categories in Chatbot Arena: https://lmarena.ai/?leaderboard In math it shares the top spot with o1 and is just a few points behind (well within errors). In creative writing it is basically ex-aequo with the latest ChatGPT 4o and in coding it's actually significantly ahead of everyone else and represents a new SOTA.

lmarena/lmsys is beyond useless, looking at prior rankings of models vs formal benchmarks or testing for accuracy + correctness on batches of real world data. It's a bit like using a poll of Fox News to discern the opinions of every American; the audience voting is consistently found wanting. Not even getting into how easily a bad actor with means + motivation (in this "hypothetical" instance wanting to show that a certain model is capable of running the entire US government) can manipulate votes which has been brought up in the past (yes I'm aware of the lmsys publication on how they defend against attacks using cloudflare + recaptcha, there are ways around that.)

Re: Grok3 Launch [video]

#617

Post by hector126 »

htfy96 wrote: post #53
ilaksh wrote: post #19

If what they say is true, then you have to give them credit for catching up incredibly fast. And slightly pulling ahead. Not only with the models, but also products.

I have a close friend working in core research teams there. Based on our chats, the secret seems to be (1) massive compute power (2) ridiculous pay to attract top talents from established teams (3) extremelly hard work without big corp bureaucracy.

Anecdotal, but I've gotten three recruiting emails from them now for joining their iOS team. I got on a call and confirmed they were offering FAANG++ comp but with the expectation of in-office 50h+ (realistically more).

I don't have that dog in me anymore, but there are plenty of engineers who do and will happily work those hours for 500k USD.

Re: Grok3 Launch [video]

#618

Post by kurikuri »

niceice wrote: post #598
angusturner wrote: post #520

Credit to the engineers that built this, but it fills me with rage that Elon has this sort of unchecked power. How long before this starts getting deployed in safety critical applications or government decision making processes? With no oversight because Elon seems to have the power to dismiss the people responsible for investigating him. Anyone not scared by this concentration of power needs to pick up a book.

[flagged]

Ah yes, replacing the ‘elites’ with a single person, much less scary. And, to be clear, the ‘elites’ are still in power (because, in America, power tends to follow money). The only thing that seems to be happening is the rapid destruction of any system Elon deems ‘bad.’

We aren’t better off at the whims of this robber baron, and I don’t understand how you can think that.

Re: Grok3 Launch [video]

#619

Post by EGreg »

ijustlovemath wrote: post #573
angusturner wrote: post #520

Credit to the engineers that built this, but it fills me with rage that Elon has this sort of unchecked power. How long before this starts getting deployed in safety critical applications or government decision making processes? With no oversight because Elon seems to have the power to dismiss the people responsible for investigating him. Anyone not scared by this concentration of power needs to pick up a book.

What's remarkable to me about criticism like this is how quickly it's rebutted by people claiming "where did they say they would do this," as if these people don't make incredibly rushed and poorly planned decisions all the time. It's like an idea immune system that rejects any criticism or self reflection. It would be sociologically fascinating if it wasn't being combined with a dereliction of congressional power an…

I literally post for years the same thing about the need for open source alternatives to social platforms where 1 person controls the algorithm to prioritize what a billion people see. And the response is “meh”. No one even bothers to read past the first paragraph:

https://news.ycombinator.com/item?id=43036350

But if you really want to see the “immune system” shine, mention web3 and smart contracts, and watch the downvotes pour in. Any time one even mentions “decentralized byzantine fault tolerant” anything, an army rises up to repeat anodyne versions of “grift… no one needs it… banks are great…” etc.

https://news.ycombinator.com/item?id=43073421

But if you mention any concerns with AI, no matter who or what you cite, the same group goes the other way and always repeats “(insert problem here) has always been possible, there is nothing to see here, move on, AI is amazing, deregulate and let the industry develop faster”:

https://news.ycombinator.com/item?id=40900155

It’s groupthink at its most obvious, repeatable, always on, and I wonder how much of it is organic.

Post reply on HN