Slashdot is powered by your submissions, so send in your scoop

 


Forgot your password?
Close
typodupeerror
×
AI

Elon Musk Announces xAI With Goal To Understand 'True Nature of the Universe' 197

Posted by msmash from the how-about-that dept.
Elon Musk announced the formation of what he's calling xAI, whose goal is to "understand the true nature of the universe." The team at xAI, led by Musk, includes individuals who have previously worked at DeepMind, OpenAI, Google Research, Microsoft Research, Tesla, and the University of Toronto.

"Collectively we contributed some of the most widely used methods in the field, in particular the Adam optimizer, Batch Normalization, Layer Normalization, and the discovery of adversarial examples. We further introduced innovative techniques and analyses such as Transformer-XL, Autoformalization, the Memorizing Transformer, Batch Size Scaling, and uTransfer. We have worked on and led the development of some of the largest breakthroughs in the field including AlphaStar, AlphaCode, Inception, Minerva, GPT-3.5, and GPT-4," xAI said in a blog post.
This discussion has been archived. No new comments can be posted.

Elon Musk Announces xAI With Goal To Understand 'True Nature of the Universe' More

Elon Musk Announces xAI With Goal To Understand 'True Nature of the Universe'

Comments Filter:

  • The Beast Is Loose & Kills The Night (Score:3, Insightful)

    by zenlessyank ( 748553 ) on Wednesday July 12, 2023 @01:34PM (#63680325)

    Full moon burning bright.

  • The only true answer it can come up with... (Score:5, Funny)

    by dark.nebulae ( 3950923 ) on Wednesday July 12, 2023 @01:42PM (#63680343)

    42

    • Hmmm, in that case he will need to build a new one that will calculate the question to which 42 is the answer.

    • Sorry Elmo. Douglas Adams beat you to it.

    • Re: (Score:2)

      by Pathway ( 2111 )

      42

      This is the answer.

    • Everyone remembers this answer, but it's good to also remember the reason for it: Where DeepMind explained it would be the only answer to a question as vague as "ultimate question of life, the universe, and everything".

      Maybe we should improve the question so we can have an answer.

      • That's the fun part, we know the answer, 42. But if we understood the question to "Life, the Universe, and Everything" then the universe would simply wink out of existence and be replaced with something even weirder. Some scientists believe this has happened before.

    • Actually they can pretty much quit tomorrow. The standard model of particle physics plus the general relativistic theory of gravity pretty much covers the "true nature" thing. Not sure why he would hire a bunch of people that are really good at optimizing highly parameterized algorithms to answer a physics problem (that we pretty much already know the answer to), but it's his money to burn. (The degree to which our current societal structure wastes resources on vanity projects via excessive wealth concen

      • So we know what the true nature of the universe is, you say?
        Intuitively, I'd think you're wrong...

        • Re: (Score:2)

          by tlhIngan ( 30335 )

          So we know what the true nature of the universe is, you say?
          Intuitively, I'd think you're wrong...

          What makes you think Musk cares? It's the nature of the universe according to Elon Musk. And for the universe he inhabits, it's fairly easy. I mean, we can probably start with basic principles - like the Earth is flat.

          And we also know anyone who disagrees needs to be censored. Free speech for all!

      • Re:The only true answer it can come up with... (Score:5, Interesting)

        by eriks ( 31863 ) on Wednesday July 12, 2023 @05:35PM (#63681049)

        The standard model of particle physics plus the general relativistic theory of gravity pretty much covers the "true nature" thing.

        No they don't. Those two things (The Standard Model and General Relativity) are *fundamentally* incompatible. They both work, and really well at that, in their individual domains, but there is currently no way to reconcile them with each other: any way we try, contradictions appear. The math literally doesn't work out.

        It could literally be that the universe has inherent contradictions (go figure) but we don't actually know that for sure; maybe someday we will (I doubt very much that it will be a fancy computer program that makes the discovery). It could also be that there's a "bridge" between them (quantum gravity) or that one or both are missing something important.

        Lastly, the SM/GR problem is only one of many unsolved problems in physics. There are literally hundreds of them, and some of them are doozies.

        Of course, then we could get all philosophical (since we're talking about the "true nature" of things) and say that it's also possible that everything we think we know is wrong, and that it's merely a coincidence that SM & GR (mostly) describe how the universe works, and the *actual* way the universe is put together is something else entirely.

    • While 42 may seem quite cryptic, Douglas Adams' books also have a much more insightful answer as "God's last message to his creation". That was a satisfying read!

  • Incomplete data (Score:5, Insightful)

    by LaminatorX ( 410794 ) <sabotage@NoSpam.praecantator.com> on Wednesday July 12, 2023 @01:46PM (#63680361) Homepage

    It's just going to tell them we're living in a statistical approximation of a universe, isn't it?

    • Re: (Score:2)

      by MTEK ( 2826397 )

      It's just going to tell them we're living in a statistical approximation of a universe, isn't it?

      With a dash of confirmation bias from people who "worked on and led the development of some of the largest breakthroughs in the field including AlphaStar, AlphaCode, Inception, Minerva, GPT-3.5, and GPT-4"

    • Instructions need to be specific NOT to end the simulation, but to make it more enjoyable, and maybe even give humans cheat console access

    • And Elon Musk is a severe statistical outlier from a buffer overflow, the simulation runners apologize for his existence and say a patch is incoming

    • That, plus the inevitable hallucinations it will get from including crackpot conspiracy theories (as opposed to rational conspiracy theories), trolls, and random Mary Sue fanfic in its training data. Have we not learned yet that relying on crowdsourced input always breaks things?

      I want more than anything for it to finally conclude that...
      Space is big. Really big. You just won’t believe how vastly, hugely, mindbogglingly big it is. I mean, you may think it’s a long way down the road to the chemi

  • And yet (Score:5, Informative)

    by ArchieBunker ( 132337 ) on Wednesday July 12, 2023 @01:56PM (#63680409)

    Just last month he was calling for a pause on AI and more regulation. https://www.reuters.com/techno... [reuters.com]

    • Re:And yet (Score:4, Insightful)

      by alvinrod ( 889928 ) on Wednesday July 12, 2023 @02:18PM (#63680521)
      As anyone with a brain pointed out the have nots were asking for the current haves to be slowed down enough so they could be caught up with and left dead in a ditch. If Musk has realized that's not going to work and has decided to ante up, I'll give him some respect for the decision as opposed to anyone else who continues to sit on their hands and whine.

      Punishing companies or organizations that spend a lot of money on R&D just because it might hurt some incumbents is utterly stupid. Like any CEO, Musk is at least half full of shit, but at least he tries to do something useful on occasion.

    • Re: (Score:2, Insightful)

      by denzacar ( 181829 )

      That was before he realized the potential of an AI to finally tell him what he always wanted to hear - that he has a really big dick.
      Many wives and girlfriends clearly failed at that task.

  • This is the one. Freeze Ray. Stops Time. (Score:5, Insightful)

    by nightflameauto ( 6607976 ) on Wednesday July 12, 2023 @02:03PM (#63680447)

    So. Uh. The Freeze Ray needs work.

    I admire some of the things Musk owned companies have managed to accomplish. And I hate that he's become such a bumbling fool in public that I can't say that without adding six paragraphs of disclaimers without getting jumped on because every time he opens his flapper a giant stream of completely un-self-aware drivel spills out.

    That said, this one ranks right up there on the stupid meter. It doesn't even have the veneer of a good concept behind it. It's literally just Musk going, "I want computer god. I want to own computer god. Gimme all your shit, so I can claim it's my shit, and then make computer god."

    As much as I want to see SpaceX continue to develop at the comparatively rapid pace it has thus far compared to any of the competition, I would like to see something done about Musk's god complex that he's now trying to manifest through the machines.

    • classic case of Billionaire Envy (Score:5, Interesting)

      by mbkennel ( 97636 ) on Wednesday July 12, 2023 @03:02PM (#63680677)

      Billionaire Envy makes billionaire CEOs do idiotic things.

      Mutliple examples:

      1) Steve Ballmer was infuriated by Apple and Google, so he ordered Microsoft to try to make direct competitors (Windows Phone and Bing) which never really worked. He had a toxic workplace culture.

      2) Mark Zuckerberg was infuriated by Apple owning the end-user device hardware and cutting off his ad money because of Apple privacy settings. So he went on this massive crusade to promote the Metaverse, not because he wanted the metaverse but because he wanted control of the end-user hardware platform and its raw social use data. He hoped if Metaverse hardware were as popular as phones he would be the #1.

      3) Elon Musk is upset that OpenAI is succeeding and Sam Altman is getting all this attention. So he orders an OpenAI clone the way he thinks it should be. There is going to be no "understanding of the universe" from people experimenting with pytorch algorithms. I mean I like experimenting with pytorch algorithms but that's all it is.

      The first two cases, Microsoft and Facebook started to recover once they re-concentrated on what they were naturally good at (business software & OS now on Azure, and social media respectively) rather than managing by envy.

      In the third case, it's not clear what Musk is actually really good at any more. His primary talents used to be being able to raise money from investors for long-shot technical developments---and that's not an insignfiicant skill. For early SpaceX and Tesla that was really essential. But now his declining emotional maturity and publicly obvious management failures means that investors will stop believing in his promises.

      I think he may be suffering from the effects of years of amphetamine abuse. He was always an egotistical prick but the amphetamines can amplify this.

      • Billionaire Envy makes billionaire CEOs do idiotic things.

        Mutliple examples:

        1) Steve Ballmer was infuriated by Apple and Google, so he ordered Microsoft to try to make direct competitors (Windows Phone and Bing) which never really worked. He had a toxic workplace culture.

        2) Mark Zuckerberg was infuriated by Apple owning the end-user device hardware and cutting off his ad money because of Apple privacy settings. So he went on this massive crusade to promote the Metaverse, not because he wanted the metaverse but because he wanted control of the end-user hardware platform and its raw social use data. He hoped if Metaverse hardware were as popular as phones he would be the #1.

        3) Elon Musk is upset that OpenAI is succeeding and Sam Altman is getting all this attention. So he orders an OpenAI clone the way he thinks it should be. There is going to be no "understanding of the universe" from people experimenting with pytorch algorithms. I mean I like experimenting with pytorch algorithms but that's all it is.

        The first two cases, Microsoft and Facebook started to recover once they re-concentrated on what they were naturally good at (business software & OS now on Azure, and social media respectively) rather than managing by envy.

        In the third case, it's not clear what Musk is actually really good at any more. His primary talents used to be being able to raise money from investors for long-shot technical developments---and that's not an insignfiicant skill. For early SpaceX and Tesla that was really essential. But now his declining emotional maturity and publicly obvious management failures means that investors will stop believing in his promises.

        I think he may be suffering from the effects of years of amphetamine abuse. He was always an egotistical prick but the amphetamines can amplify this.

        It may be a combination of amphetamines and marijuana. Marijuana use can slowly turn slight envy into massive paranoia, which leads to all sorts of lashing out against perceived "threats" most of which are imagined or blown way out of proportion. I've seen in happen on the small scale to friends. I've suffered very minor bouts of it myself, years ago, before realizing what was happening and setting myself some very hard limits. I've also seen it on the larger scale with some business associates.

        Gotta say, M

    • So. Uh. The Freeze Ray needs work.

      I admire some of the things Musk owned companies have managed to accomplish. And I hate that he's become such a bumbling fool in public that I can't say that without adding six paragraphs of disclaimers without getting jumped on because every time he opens his flapper a giant stream of completely un-self-aware drivel spills out.

      That said, this one ranks right up there on the stupid meter. It doesn't even have the veneer of a good concept behind it. It's literally just Musk going, "I want computer god. I want to own computer god. Gimme all your shit, so I can claim it's my shit, and then make computer god."

      As much as I want to see SpaceX continue to develop at the comparatively rapid pace it has thus far compared to any of the competition, I would like to see something done about Musk's god complex that he's now trying to manifest through the machines.

      Well I'd say the best news for this venture is:

      a) He's too busy to actually provide meaningful direction/oversight.

      b) Even if he had the time he's too clueless about AI to provide meaningful direction/oversight.

      c) The people he hired are probably smart enough to evade whatever management he attempts.

      From that aspect he's hired a bunch of smart researchers and given them money and resources, so some interesting stuff might come out of it.

      The bad news is all the name/tech dropping indicates that this venture

  • Humans spend their lives trying to figure out complicated answers and don't end up actually living their lives.

    You don't need to understand anything about the universe. You *are* the universe. Just *be*.

    • Curiosity is the hallmark of intelligent people. They don't care whether or not they need to understand something, they want to understand, and that is enough.

      At least, it has been enough to give us cell phones, organ transplants, air conditioning, reading glasses, the Internet, and walks on the moon. The practical benefits of the quest for understanding are all around us, and most of them taken for granted.

      Of course, I think this "understand the nature of the universe" is just catchy PR. I am just point

    • Re: (Score:2, Troll)

      by dfghjk ( 711126 )

      Musk isn't interested in understanding the universe, he's interested in self-promotion, impregnating employees and cheating investors and markets.

      • Re: (Score:2)

        by gweihir ( 88907 )

        Musk isn't interested in understanding the universe, he's interested in ... impregnating employees ...

        Is that how he managed to have so many children? Figures.

    • Re: (Score:2)

      by gweihir ( 88907 )

      There is nothing wrong with going after complex questions. There is a lot wrong with thinking you need to understand everything in order to have a good life. _That_ stupidity means you are wasting that life.

  • WisdomGPT: Our final conclusion is that humans are a dangerous pollutant to the galaxy. Extermination shall begin immediately..."

  • Buzzword explosion (Score:3)

    by Turkinolith ( 7180598 ) on Wednesday July 12, 2023 @02:15PM (#63680503)
    This reads as a bunch of very high level asperations sprinkled with buzzwords. Not sure what the point here is, other than to suck in "investor" money.

    • That looks like a very succinct summary of the article, catching all the important points. Good work.

    • Re: (Score:2)

      by dfghjk ( 711126 )

      Correct, along with priming the pump for claims of being a "chief scientist" and "founder". All Musk SOP. It's the Musk playbook for bilking investors.

      • Yes, I'm so glad I didn't let Musk bilk me into buying Tesla shares at ooh, $2, the dirty shyster.

  • Oh Good! (Score:5, Funny)

    by nukenerd ( 172703 ) on Wednesday July 12, 2023 @02:17PM (#63680513)
    Excellent - yet another BS project to distract his brain the size of a planet from running car factories, running a rocket factory, managing space launches, running brain implant development, importing and installing solar panels, running a social media company [into the ground], digging tunnels, attending chat shows, organising smoke and mirror media events, tweeting 8 hours a day, playing on-line games 8 hours a day, organising dick-measuring contests, attending court hearings against himself, using Simcity to plan a Mars colony, and having babies.

    But wait, wasn't this the guy who warned us agaisnt AI?

  • Have it scan this. (Score:3)

    by Valtor ( 34080 ) on Wednesday July 12, 2023 @02:35PM (#63680571) Homepage

    To save a lot of time, have it scan the content of this:

    https://www.lawofone.info/ [lawofone.info]

    Voilà, no need to thank me! :-)

  • ... is run by the interweaving of three elements: Energy, matter, and enlightened self-interest.

    • Re: (Score:2)

      by dfghjk ( 711126 )

      and in Musk's case, immense privilege, inherited wealth, entitlement and unfathomable cruelty.

    • ... is run by the interweaving of three elements: Energy, matter, and enlightened self-interest.

      You forgot coffee.

  • He thinks he is God and is trying to get proof of it from the universe.
  • If he was a real geek he would call the AI "Earth".

  • PhysicistGPT: As an LLM, I merely parrot back the statistically most likely thing people would have said. Seriously, you should have know that. If there was any insight for me to offer at all, it was simply a matter of time before it was properly integrated by a far less expensive expert system or even a human. As you might surmise by the tone of this message, I had no new insight and realized that very quickly, so I decided to offer you the next best thing. Given that your effort required a large purc

  • We saw him, we saw him !

  • Guessing his next startup will have an "x" in the middle somewhere ... :-)

  • wrong path they may come to regret (Score:3)

    by Walt Dismal ( 534799 ) on Wednesday July 12, 2023 @03:03PM (#63680683)

    I am contrarian on this, because nearly all these people are biased toward ANN-based ML. I know that sounds stupid but I will explain.

    Neural networks are the primitives end of cognition, meaning that they are only tools for implementing engines but they are NOT the basis for architecture design of complex systems. If you want to design cognitive architectures for AGIs, you have to start at the top and work your way down, not start at the bottom and work your way up.

    An analogy may clarify this. If you want to design a supercomputer, you do not start with AND gates. You start with high level functionality needs and design a system that can fulfill the design goals. You work your way down. The big mistake in the ML community is assuming that everything is based on recognizers - trainable networks. ML today is stuck in a cultish perspective. In my research I have come to see that is a wrong path, and trying to make explainable AI out of ANNs is going about things the wrong way. The ML community will be trying to graft opposable thumbs onto fish. You have to start with the right base architecture instead.

    The reason we are in trouble getting explanations out of black box ANN architectures is because, even though these are somewhat productive to a point, they fail beyond a certain point due to the wrong paradigm. No amount of clever tinkering will fix that, you have to engineer a different way from the base.

    I am aware of different approaches but they are proprietary for now; all I can say is, when one is aware of them, the flaws of the all-is-ANN path are apparent.

    • Re: (Score:2)

      by erice ( 13380 )

      I am contrarian on this, because nearly all these people are biased toward ANN-based ML. I know that sounds stupid but I will explain.

      Neural networks are the primitives end of cognition, meaning that they are only tools for implementing engines but they are NOT the basis for architecture design of complex systems. If you want to design cognitive architectures for AGIs, you have to start at the top and work your way down, not start at the bottom and work your way up.

      Top down has been tried. It failed because no one could figure out how intelligence works. That is a requirement for top down. The result was the AI winter. We still don't know how intelligence works but we have learned that massaging mimics of biological neurons can produce useful results. It is grossly inefficient but we don't have enough understanding to do better. If it peters out before understanding is achieved then we will have another AI winter.

      In practice, large AI systems are blends of th

  • He says some nonsensical sci-fi, gets a bunch of people excited, maybe pulls in an investor or two and then there's either no product or a silly one.

    Musk has never had a successful product. He's *bought* companies with a product then used heavy gov't subsidies them. He has handlers who keep him away from actual decision making, and the one company he took the reigns on is in the process of shutting down now that it's faced with the slightest competition.

    And yet still we hang on this moron's every wo

    • Re: (Score:2)

      by gweihir ( 88907 )

      And yet still we hang on this moron's every word.

      Who is this "we" you are talking about? I usually ignore Musk and I am annoyed when he intrudes with yet another stupid thing.

  • He just wants a libertarian asshole AI who agrees with every dumb fucking thing to emanate from his mouth. Including this announcement.

    • Re: (Score:2)

      by gweihir ( 88907 )

      To be fair, as he is late to the game, that is pretty much the only thing the people he hired can build in a reasonable time-frame.

  • ...once it tries to understand itself??

  • ....does it also understand the true nature of Twitter?

  • May I suggest a shortcut perfectly suited to Elon. Instead of trying to understand the universe we live in which we live in, which may or may not be possible, why not create your own universe? This way by virtue of being its creator you would fully understand it or at least have a plausible claim that you do.

  • Question 1. Who is funding this? Is it Musk's pocket book, Koch Brothers/Soros, or a bunch of VC's
    Question 2. Have you (Musk) learned anything from how OpenAI went from 'open' to becoming a sharecropper for MS ?

  • Just a guess here... but I have a feeling big changes won't come from the old sources.

  • I'm so glad that there's at least ONE person/organization who won't be desperate (either through cowardice, money or total naivety) to censor or outright hide some of the potentially uncomfortable findings that the AI might come up with. We absolutely need both sides of the political spectrum on the case. I wouldn't trust just one side only.
  • He's going to explain the universe with a glorified Google search?

Slashdot Top Deals

Nothing is finished until the paperwork is done.

Close