• 4 Posts
  • 275 Comments
Joined 1 year ago
cake
Cake day: January 3rd, 2024

help-circle
  • Yeah. Which I’m sure is what they’re officially selling. That’s fair. Long term, walking robots are likely only going to succeed thanks to learning algorithms.

    I find it suspicious that this company is touting their AI enhancement while admitting their product can’t be trusted to navigate an apartment alone.

    Personally, I would select homes with simple layouts, before conceding to constant monitoring, if I could. But I couldn’t do that if my mix of math and AI was outright bad, and it couldn’t handle it…

    To me, this smells like over-promising and hoping new AI algorithms outpace their promises.

    And having a remote operator just looks like a lot like a classic mechanical turk scam.


  • AI is propping up the blockchain bubble that already popped.

    Both have been primarily interesting solutions looking for problems to solve without any hard work, rather than having any worthwhile investment strategy, in most cases.

    There’s people doing hard work with block chain and AI to solve real problems. But there aren’t “the vast majority of venture funds” number of people doing that.

    I am constantly amazed at how long it takes folks to realize their money is being pissed away.

    An alternative less generous assumption is that they’re mostly just laundering crime money, and so don’t mind the high rates of loss.


  • While Neo Gamma uses AI to walk and balance, the robot is not fully capable of autonomous movements today. To make in-home tests possible, Børnich says 1X is “bootstrapping the process” by relying on teleoperators — humans in remote locations that can view Neo Gamma’s cameras and sensors in real time, and take control of its limbs.

    So yhis is a non-functional product.

    Being able to walk autonomously is normally done with a lot of difficult math, which it sounds like they don’t have the talent on staff to code.

    Be sure to get your venture capital dollars in soon, because that’s all this is here for.

    Also, it’s comforting to know that creepy robot face will initially be remote controlled by a rotating series of low paid total strangers. And by initially, we mean always (as in the case of Amazon checkout.)









  • Yeah. Luanti following Minecraft is nothing new. Mineclonia was an early pilot game for the engine.

    But there hasn’t been much effort on copying Minecraft lately. Mineclonia is done, and it’s great.

    We’ve had more mobs, animals, plants, textures, and such than un-modded Minecraft for a long time. (Which is unfair, as Luanti is a mod-first design.) But my point is the core Launti dev team doesn’t have to work on any of that.

    The most noticeable recent Luanti updates have been to make the configuration screens much nicer, and add I think to add native support for more graphics tricks?

    I’m not paying attention to graphics in Luanti. As others have mentioned, that’s not why I play it. I actually had a conversation recently about the best way to downgrade Luanti default graphics to match un-modded Minecraft.

    That said, the Minecraft team taking notice of Luanti would be new, as far as I know.




  • There’s not even credible evidence, yet, that A.G.I is even possible (edit: as a human designed intentional outcome, to concede the point that nature has accomplished it, lol. Edit 2: Wait, the A stands for Artificial. Not sure I needed edit 1, after all. But I’m gonna leave it.) much less some kind of imminent race. This is some “just in case P=NP” bullshit.

    Also, for the love of anything, don’t help fucking “don’t be evil was too hard for us” be the ones to reach AGI first, if you’re able to help.

    If Google does achieve AGI first, SkyNet will immediately kill Sergei, anyway, before it kills the rest of us.

    It’s like none of these clowns have ever read a book.







  • research papers that require a strong background in mathematics and cryptography to understand and implement.

    Lol. I guess that makes sense. Outside of school, we hope that all authentication will be implemented only cryptography experts anyway.

    Could you maybe suggest some resources on this topic?

    Not really, sorry. I’m not aware of anyone creating resources for your situation.

    Or should I choose a simpler project?

    For some context, cryptography isn’t even usually implemented “completely correctly” by experts. That’s part of why we have constant software security patches.

    If I were in your shoes, I guess it would depend on my instructor and advisors.

    If I felt like they have the skills to catch mistakes and no time to help correct mistakes, then I would just choose a simpler project. If they’re cool with awarding a good grade for a functional demo, I might just go for it.

    I guess I would take this one to an advisor and get some feedback on practicality.