xAI

[paxys]

> We are actively recruiting experienced engineers and researchers to join our team as members of our technical staff in the Bay Area.

It's funny that after all the talk of remote work, tech globalization and the SF/Silicon Valley exodus in the last few years (including much of it from Musk himself), we end up with pretty much all mainstream AI companies – OpenAI, Anthropic, Google, Nvidia, Scale AI, Midjourney, and now this one – primarily hiring in-office in the Bay Area.

The technological center of world continues to be a hundred square miles in Northern California, whether people like it or not.

[jimlongton]

> The technological center of world continues to be a hundred square miles in Northern California

If your definition excludes the likes of the entire semiconductor industry for example.

[duxup]

Especially for startups / small orgs I think it can be critical to work closely with the first or specific employees in person as much as possible.

[starshadowx2]

Midjourney is not entirely based in SF, it's a global remote team without an in-person office.

[esalman]

Musk tried but probably couldn't make it in Texas.

[paxys]

> The goal of xAI is to understand the true nature of the universe.

Who wants to bet it's going to end up launching a "me too" generative AI product.

Maybe Musk should first take these skills to Tesla's self driving division then worry about the rest of the universe.

[dragonwriter]

> Who wants to bet it's going to end up launching a "me too" generative AI product.

You mean after Musk explicitly, piblicly said that’s what he was going to do, and make it anti-woke as its main differentiating feature?

Pretty good bet.

[blaze33]

> The goal of xAI is to understand the true nature of the universe.

Indeed, I'm sure it's marketing speech but it tells me nothing about what xAI actually does. Yeah, something AI-related I guess, way to go...

I'm currently reading C. G. Jung ("L'homme a la découverte de son âme", French title, doesn't seem to be an English edition to this one though, roughly translates to "man discovering his soul") where among other things he talks about the nature of reality: "the psyche is the real entity in the supreme degree, since the only immediate one". And also that your typical Western man usually suffers from "the disease of knowing-a-priori-everything-better-than-anyone"...

Maybe Elon should read it and get a grasp of himself because I probably won't take lessons about the true nature of the universe from the guy who literally wants to do dick measuring contests on twitter.

[bearjaws]

If it follows the path of Twitter it will be a "free speech AI" that will censor anything negative about Elon.

[md_]

My money is on shipping the OpenAI API, wrapped with a prompt like,

"Below is a conversation between an intelligent AI assistant which deeply understands the true nature of the universe, and a human."

[systemvoltage]

I wonder if you would say the same with many nations doing their own space program? Why copy NASA and others when you can focus on poverty and other things than to worry about the rest of the Universe?

Some competition would be great IMO.

I wish Elon started a TSMC competitor though. A moat so high it affects geopolitics.

[411111111111111]

Their self driving will work right around the same timeframe spaceX actually succeeds a launch with a final budget lower then the most expensive NASA launches were.

It's honestly hilarious how he was lauded as tony stark in the flesh considering the only thing he excelled at was getting simps to gift him money

[mrbonner]

Don’t we already know the answer is 42? /s

[haswell]

It has been the collective goal of the many fields of science, throughout the history of humankind, to understand the true nature of the universe.

There's something that feels incredibly ostentatious about the opening sentence "The goal of xAI is to understand the true nature of the universe.". Not because I think it's a bad goal, or a goal that no company can claim, but because of the broader context around AI and Musk's antics.

I'm genuinely curious to see what comes out of this, but it also gives me the same queasy feeling in my gut that I got when looking at various crypto initiatives over the years.

I don't doubt that there are incredibly smart people getting involved. It's just a bit difficult to take it seriously. I hope to be proven wrong.

[jrm4]

Eh, I'm comfortable taking this further and dismissing it out of hand.

This looks like a scam. Not the typical regular person type of scam, but the kind specifically targeted to "wealthy Silicon Valley idiot."

[3cats-in-a-coat]

- Tesla was launched to transition the world to sustainable energy (not that electric cars do that up close, but never mind).

- SpaceX was launched to make us a multiplanetary species.

- Twitter was bought to fight the woke mind virus that is destroying the world.

- And now xAI will understand the Universe.

One of the clear rules of PR for Musk is that whatever he says it must involve at least one whole planet, ideally multiple planets, galaxies, or ideally, the whole Universe. What is delivered is another story entirely.

[burkaman]

That's obviously straight from Elon. He said he bought Twitter to help humanity and started SpaceX to protect the light of consciousness.

[stillsut]

> The goal of xAI is to understand the true nature of the universe.

People are misinterpreting this as the research goal. No, it's an implicit negation of an ethos - AI Ethics, Trust & Safety, ESG, Politically Correct - that have dominated decision making in product and hiring in tech for the past decade.

Whether or not you agree with him, it's hard to deny that Musk is the only one offering researchers an escape from the bowing to the orthodoxy while also offering an A+ pedigree institution.

[agnosticmantis]

> The goal of xAI is to understand the true nature of the universe.

Genuinely curious: Does anybody take this stuff seriously?

I just feel embarrassed to be working in tech (in Silicon Valley of all places), being associated with these charlatan clowns peddling nonsense like this and being praised for it.

[harisec]

Same here, this is ridiculous

[erlend_sh]

I count a dozen dudes. I don’t think we’re gonna figure out the true nature of the universe via that kind of monocultural uniformity.

[ToValueFunfetti]

I'm all for diversity for diversity's sake, but I find this to be a very strange position. What difference between men and women are you proposing that allows one or both to have unique insights into the nature of the universe?

It seems like everything I know about the fundamental universe does come from a dozen or so dudes (for historical reasons rather than competency, of course), like Socrates, Plato, Aristotle, Bacon, Newton, Darwin, Einstein, Turing, and von Neumann. I fully believe that women could have solved the same problems, it just doesn't make sense to me to say that men can't, considering that they already have.

[caeril]

This is terrible. No Italian people, either.

Moreover, it's important to note that five of them are East Asian. That's a population over-representation of eight times. This is a completely unacceptable display of Asian Supremacy in action.

No dolphins or tardigrades, as well. How the hell is their model going to "understand the universe" without dolphins in charge of racking the H100s?

Without a rainbow coalition of employees, the embedding tokenizer for this company will look something like this:

embed_tok("give me a recipe for the best Italian Cannoli") -> [71822, 3, 915, 74, 162, NULL, NULL]

Shameful.

[bezout]

It’s basically all dudes that jumped from FAANG to FAANG. Impressive but I would expect to see biologists, neuroscientists, physicists, psychologists etc

[wyager]

What makes you say that? The history of physics (our most successful endeavor in this direction) is very monocultural (by the standard of granularity you are using).

[numbers_guy]

What do you want Musk to do? Visit every university in the US and Europe to poach off the top women in AI, so that he could have a mixed team? Would that actually be good for the world? Those poor uni AI departments worked really hard to find a woman to hire and then Musk comes along and poaches her off.

And then somehow that would be better, because?

[oh_sigh]

I see people from Britain, Germany, Hungary, Canada, China, and America.

[fassssst]

That was my first thought too. That’s a lot of dudes!

[atleta]

> The goal of xAI is to understand the true nature of the universe.

That's kind of what the whole natural sciences community is doing. Like that is one way to define what natural sciences are about. That's basically the biggest goal anyone could mention. Like your cliche "world peace" answer that no beauty contestant ever says any more.

[kklisura]

Huh. I thought _understanding universe_ would require research in quantum mechanics, string theory, astrophysics… you know - physics. Not _AI_.

[friend_and_foe]

I think understanding information, cognition, computation, whether an LLM has the emergent property of understanding, all give us some great insight into the nature of the universe that can be applied elsewhere as well.

[seydor]

If the AI understands the universe, humans dont have to

[cma]

AI may start contributing new math and science fairly soon.

[pms]

In AI research, the acronym XAI stands for eXplainable AI [1]. To understand anything about the universe, AI will need to (i) make accurate predictions about some parts of the universe and (ii) explain the laws and the models it used to make these predictions, just like natural sciences do [2]. I believe eXplainable AI can get us there one day. However, I'm surprised that, despite it's name, xAI doesn't mention eXplainable AI and the team doesn't seem to have experience in eXplainable AI, which makes me think it's just another deep-learning AI venture.

[1] https://en.wikipedia.org/wiki/Explainable_artificial_intelli...

[2] https://en.wikipedia.org/wiki/Scientific_method

[duxup]

I can’t tell what this company does.

Does anyone know?

[gooseus]

My guess would be PR and myth-building for Elon, et al

[slenk]

It will be vaporware, like 90% of what Elon offers.

[david927]

The goal of xAI is to understand the true nature of the universe.

[4gotunameagain]

> The goal of xAI is to understand the true nature of the universe.

That is.. a bit vague ? What is the purpose, and what does it have to do with AI ?

Why launch such a teaser without announcing the (actual) goal ?

They cannot really think that they will be able to understand "the true nature of the universe" ?

[chasing]

What Musk needs to do is fund smart people to do interesting stuff and then get out of the way. If that's what this is, great. If it's just another opportunity for him to do something high-profile and half-baked to massage his weirdly fragile ego: Ugh.

[paxys]

> What Musk needs to do is fund smart people to do interesting stuff and then get out of the way.

So, OpenAI? The problem is always that his ego gets in the way. He wants to be recognized as the messianic figure behind AI, but the ship is sailing ahead without him. Hence all the new efforts (Tesla AI robot, XAI) and grandiose promises ("understanding the nature of the universe").

[choppaface]

... not to be confused with X the moonshot factory? ( https://en.wikipedia.org/wiki/X_(company) )

[ngrilly]

> The goal of xAI is to understand the true nature of the universe.

Understanding the true nature of intelligence would be a good start, and also a better fit for a company with AI in its name.

[dwaltrip]

For those in the know, any comments about the team? Looks impressive to my uniformed eyes.

[buildbot]

Unless I am missing something, the current company is literally 100% men - not unexpected given Musk's alleged chauvinism: https://www.telegraph.co.uk/business/2022/12/09/elon-musk-se...

[skilled]

If only Elon hadn't splurged the 40+ billion on Twitter, he could've been stacked with H100's right about now.

[uoaei]

"The true nature of the universe" is either about cosmology or it's (wrongly) about ontology/metaphysics. Both seem like interesting pursuits. Fitting cosmological models to massive datasets from astonomical observations seems laudable and not entirely out of reach (one approach: space- and time-localized NeRF fitting on the entire visible universe accounting for relativistic effects). It would be strange to describe discovering ontologies as learning the "true nature" of things unless you were a hardcore idealist/Platonist, but that seems more appropriate given the starting points afforded by the current state of the field and the hype that surrounds LLMs.

[rcarmo]

I wonder if this can understand Twitter. I'm also curious as to how it can possibly "even out" his portfolio even as it tanks in value, and how he got those folk to join up at this stage...

[kstrauser]

Five years ago, I would have leapt at this. Today, there's no way I'd want to be anywhere near a Musk-based tech company.