Live data from Hacker News

Grok3 Launch [video]

x.com

601–610 of 1001 posts

Re: Grok3 Launch [video]

#601

I don't understand how and why Grok would be related to "understanding the nature of the universe", as Musk puts it. Please correct me if I'm wrong, but they basically just burned more cash than any human should have to buy Nvidia GPUs and make them predict natural language, right? So, they are somewhat on-par with all the other companies that did the same. This is not innovation, this is baseless hype over a mediocr…

There are two answers to this. Answer 1: Some people think that LLMs are a path to the singularity, a self-improving intelligent program that will vastly exceed human intelligence and will be able to increase its knowledge exponentially, quickly answering all answerable scientific questions. Answer 2: LLM companies need to keep the hype train rolling. I didn't watch the whole clip; I jumped around a bit, but I notice…

he may have deleted it but… around 2020 or so there was starting to be a lot of hype about llms. elon musk responded to a “doomer” on twitter saying that he “didnt see the potential for that” referring to LLMs achieving AGI. it was a 100% dismissal of everything he is saying now. at that point elon musk had already been saying publicly for years “AI is more dangerous than nukes.” but he also had voluntarily walked away from openAI which he would never do if he thought there was any chance of AGI. i just want to known the truth… is this really just advanced search and some jobs will be lost because they ended up being nothing more than search tasks (ie coding boilerplate) or are we really on the cusp of AGI (and therefore in a great deal of danger)? its impossible to say whether or not elon musk really believes what he is saying… there are public figures on both sides providing conflicting explanations.

as i watched the grok3 stream i became very angry. so very tired of being jerked around and not knowing whether or not i should be planning for the future or investing in the world as it is now… its really a form of psychological torture

Re: Grok3 Launch [video]

#602
post #541

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.

I do not share your fear & anxiety. What concrete danger do you imagine will happen as a result of xAI? Try to be as concrete as possible. Also, dang, is there anything we can do to keep the comments on this submission tech-focused? Perhaps the Elon-bashing political digression can be split into its own thread?

The problem, as I see it, is that the results from AI systems will either be used to make decisions, even if those results are flawed. Or worse, those flawed results will be used to justify decisions that negatively impact peoples' lives.

This isn't something specific to xAI, but it turns out that the person who controls xAI also holds an unusually strong influence over the highest level government officials. These officials can use xAI as an excuse to implement harmful policy, "because the computer said this is the best course of action"- not unlike people who end up driving on train tracks or into large bodies of water because their GPS told them to go that way.

Re: Grok3 Launch [video]

#603

Earlier quoted context omitted.

Naive question from a bystander , but since DeepSeek is open source and is on par with o1-pro (is it?), shouldn't we expect that anybody with the computer power is capable to compete with o1-pro?

It's not on par with o1, let alone o1-pro

It's on par/better/worse depending on the problem. o1 is significantly worse, for example, in Rust programming than Claude 3.5; at least for me.

Re: Grok3 Launch [video]

#604

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.

As a Historian (and a German historian in particular) - I've spent a reasonable amount of time educating myself on the nature of fascism and in particular the break down of democracies (Wiemar, France, and also the erosion of civil liberties during the great depression in the United States).

I have also been a delegate to both the RNC and the DNC at a state level.

This is not a appeal to authority, but rather a honest response to your request for my education level.

IMHO, The root cause of the "senseless partisan drive" is the fact that he founding fathers could not come up with a way to restrict parties (they called them "interests") and left them unchecked. This is a constant "sin" of the American political system, and is a key reason Slavery survived as long as it did, why separate but equal became the law of the land, why America shot itself in the foot several time with the Banks of America and why we are looking at the wrong side of history now.

The parties now act to destroy each other as their prime directive, rather then to better the country. I liken this to Wiemar Germany, where the increasing radicalization of both the Nazis and the Communists led to political instability and eventual violence that destroyed the government. That erosion of democratic norms, as well as the "other side must be destroyed for us to survive" messaging is the true threat, IMHO.

I would strongly suggest Richard Evan's three part history on Nazi history to understand Fascism. Don't worry, you can still hate and worry about Trump and think he is the next coming of Hitler afterwords - it will just be for better reasons.

Re: Grok3 Launch [video]

#605

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 use it for lots of stuff where I'm not an expert, or that are low stakes. I don't use it for the "core" of my job, but there are many things that are not "core" and still eat up a lot of time, in fact, most of my workday would be in this category. Some typical examples from my daily life as a university professor:

- Writing Python scripts to make charts out of Excel sheets, and then refine them. I could do it myself, but I would need to learn a library like Seaborn or similar which honestly is not especially intellectually stimulating, and then spend nontrivial amounts of time iterating on the actual code. With LLMs it's a breeze.

- Working with cumbersome LaTeX formatting, e.g. transposing a table, removing a column from a table, etc.

- Getting the tone just right in a professional email written in English to someone I don't know much (I'm not a native speaker so this is not trivial).

- Finding resources on topics that are tangential to what I do. For example, yesterday I needed to come up with some statistics on English words for a presentation I'm preparing, and I needed a free corpus where I could search for an n-gram and get frequencies of next words. I don't usually work with that kind of resource, it was just a one-off need. I asked for corpora of that kind and got a useful answer instantly. The manual process would probably have implied going through several options only to find that I needed a license or that they didn't provide the specific statistics I needed.

- Brainstorming on titles for scientific papers, presentations, names of concepts that you introduce on a paper, variable names, etc.

- Shortening a sentence in a paper that makes me go over the page limit, or polishing the English in a paragraph.

- Summarizing a text if I'm kind of interested in knowing the gist but have no time to read it whole.

- Answering quick questions on basic things that I forget, e.g. the parameters to make a Linux folder into a tar.gz. Man is too verbose and it takes time to sort the wheat from the chaff, Google is full of SEOd garbage these days and sometimes you need to skim a lot to find the actual answer, LLMs are much faster.

- Writing bureaucratic boilerplate, the typical texts with no real value but that you have to write (e.g. gender perspective statement on a grant request).

- Coming up with exam questions. This is a rather repetitive activity and they're fantastic at it. At my place we also have two official languages and we need to have exam assignments on both languages, guess who does the translation now (respecting LaTeX formatting, which previous machine translation tools typically wouldn't do).

- As an example of a one-off thing, the other day I had to edit a Word document which was password-protected. I asked ChatGPT how to unlock it and it not only answered, but actually did it for me (after 3 tries, but still, much faster than the time it would have taken for me to find out how to do it and then actually do it).

These are just some examples where they contribute (greatly) to my productivity at work. In daily life, I also ask them lots of questions.

Re: Grok3 Launch [video]

#606

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…

Many people replied with anecdotes, but recently Anthropic published analysis of claude.ai production traffic. As far as I can tell this is the single best currently existing source of "how people actually use LLMs". Everything else you can't be sure whether it is representative. https://www.anthropic.com/news/the-anthropic-economic-index

Thanks, this is a gem ! However, I suspect the fact that "programming" is such a bit usage, is that AI is closely integrated into text editor, as an "autocomplete on steroid" usage.

As they state in the report, I don't think they can measure how many people just ignored the suggestion from claude right away ? Or delete more than half of the suggested code ?

Imagine if the real impact of AI is "suggesting things that people are discarding immediately ?"

Call the "Unamusing misuse of resource"... [1]

[1] https://docs.kde.org/trunk5/en/amor/amor/amor.pdf

Re: Grok3 Launch [video]

#607

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.

Given the two of you probably have different models of reality, perhaps you two can try and figure out which is correct by seeing which model gives better predictions?

So try to come up with some sort of future observation that can be made where you think the other person's model would give a different answer to yours about what you would be able to objectively observe.

What do you reckon?

Re: Grok3 Launch [video]

#608

What are your first impressions using it? (Not available in Europe currently). Is it a game-changer?

>Not available in Europe currently I hate how its the same story for every new AI technology. If someone can tell me who to vote for or where to protest to change this awful EU law, that would be great.

You could move to America and avoid the fake delays blaming the regulations tech companies don't like.

Re: Grok3 Launch [video]

#609

Earlier quoted context omitted.

Considering that OpenAI subscription is $200 per month, and "Premium Plus" subscription that includes this thing is only $40 per month, does that mean instantaneous "Elon factor" is now at least -$160 per month per user, or is it supposed to be added up to more than -$240 per month? How would the math change after factoring in that OpenAI isn't even covering entirety of opex with the sub anyway, and/or people finding…

No... sigmoid10 was comparing with o1 (not o1-pro), which is accessible for $20/mo, not $200/mo. So, the "Elon factor" in your math is +$20/user/month (2x) for barely any difference in performance (a hard sell), not -$160/user/month, and while we have no clear answer to whether either of them are making a profit at that price, it would be surprising if OpenAI Plus users were not profitable, given the reasonable rate…

“The impression overall I got here is that this is somewhere around o1-pro capability”.

“Grok 3 + Thinking feels somewhere around the state of the art territory of OpenAI's strongest models (o1-pro, $200/month)”.

Re: Grok3 Launch [video]

#610

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]

He also runs tesla, twitter, xai, boring company, that brain thing company, government agency and has like 10 kids. I'm scared.
Post reply on HN