Comment by klunger | Hacker News Reader

Comment by klunger | Hacker News Reader

> xAI will continue on this steep trajectory of progress over the coming months, with multiple exciting technology updates and products soon to be announced.

There is a lot of potential for using AI in drug discovery and development, biotech more broadly and chemistry/material science. Pharma is investing heavily in this right now. If useful, the output here could potentially also support Neuralink and even SpaceX.

Coupled with the line about the "true nature of the universe", I guessed this was really about entering that space.

But when you look at the careers page [https://x.ai/careers#open-roles], they are only hiring AI engineers. No biochemists or MDs, material scientists or any other natural science domains. So, if natural science discovery is actually on the road map, either: - it is in the long term future - they have no idea what they are doing

More likely, they are not going for natural science and this is basically just a play to compete with openAI. And, in that case, I don't understand how they convinced investors to put 6 billion dollars into it.

The “true nature of the universe” bit is that Elon believes that competing LLMs are too neutered because they disallow certain terms etc. (his words are much more politically charged and I do not agree with his take on this and many other things)

Therefore he believes that Grok can be an LLM trained on the voices of the people using his alleged free speech platform:X.

Is the assumption here is that we can somehow understand the nature of the universe if we stop censoring the common man and have an unmuzzled LLM that talks like him instead of the Bay Area AI elites? My uneducated guess is well learn more about the true nature of the common clay^w man.

For context, it should be noted that his platform disallows certain terms too, but sometimes worse (in a way).

For example: saying “cis” or “cisgender” flags your post as abusive and limits visibility. Saying the 6-letter (or 3-letter) f-slur does not.

Elon’s vision of free speech is a world where you can say anything you want as long as it isn’t mean towards Elon or Alt-right ideology. Which is actually pretty hilarious to think about in the context of a training dataset for a generative model…it’s literally gonna be a bullshit generator.

Are you not able to find mean things about Elon or alt-right people on Twitter? How hard did you look?

I agree this appears to be Musk's opinion of LLMs in particular.

However, as Musk has already got AI in his cars and was interested in the topic well before LLMs (founding investor in OpenAI when they were doing reinforcement learning), I'd be extremely disappointed if he had forgotten all of that in the current LLM-gold-rush.

(That's not a "no"; he's disappointed before).

Because OpenAI is the poster child and that kind of AI is already shoved in all kinds of products by Microsoft.

AIs like AlphaFold are hardly in the news compared to OpenAI and its competitors.

The thing about “figuring out the true nature of the universe” is that you *have to do experiments*. It’s non-negotiable. There’s no amount of really hard thinking or parameters or GPU’s that will let you know the secrets of the universe. It’s astonishing to me that both the AI-maximalists and the AI-doomers are both seemingly unaware of this basic, fundamental fact of science.

Szegedy and some others were working on science-related (math and natural science) projects at Google prior to leaving. This is probably just piggy backing on their prior work without any commitment going forward.

Of course they're not going to make any fundamental contributions to natural science or mathematics (or likely even LLM training/understanding).

Reply on news.ycombinator.com