New AI Policy Update on Safety, Censorship & Unexpected Risks

Поділитися
Вставка
  • Опубліковано 19 лис 2024

КОМЕНТАРІ • 48

  • @jakeadams2562
    @jakeadams2562 11 днів тому +3

    This podcast is so refreshing… feels at least 2 sigma ahead of every other podcast that tries to present on “current” technological progress.
    Bleeding edge stuff guys, and I think I speak on behalf of many here… please keep uploading these!

  • @LePleb7
    @LePleb7 14 днів тому +6

    Need these every other day tbh

  • @clearpupil
    @clearpupil 14 днів тому +4

    Working in a medical clinic in Canada I wish I could communicate with my patients from South Sudan unfortunately there is insufficient data from this region.
    Anyway you both raise good questions guys thank you. Not sure how we get to free unbiased Ai from where we are now. Perhaps we need put it on a strict diet of no-junk material.

    • @eriklondon2946
      @eriklondon2946 14 днів тому

      I have heard stories about how much of Canada is now immigrants. Curious are most of them beneficial for industry/tech? aka are they educated or business owners?
      I would imagine this would be hard with South Sudanese people. Also I wonder about languages that don't have a written language at all... AI will have to build a structure for the language only off of sounds.

  • @eriklondon2946
    @eriklondon2946 14 днів тому +18

    Thank you guys again. How do we get a 10-mile square area near to Houston, or somewhere else in Texas, where we can build the Shenzhen/Shanghai of the free world? It would be great if we could raise money to buy it through a 506(b) or a Reg +A . We could ask for it be a "Special economic zone" to have 0 taxes for the first few years of operation to help reduce costs.

    • @eriklondon2946
      @eriklondon2946 14 днів тому

      1. Houston has the only functioning US oil refinery.
      2. The US largest natural gas port is 30 miles east at Sabine Pass for production chemicals and power supply.
      3. Texas is open to using MSR or SMR nuclear power as it would stabilize their power grid and also allow for AI GPU farms
      4. Most products, especially tech products, require petro-based chemicals for their manufacturing
      5. Houston is one of the largest ports in the US
      6. The US business trade with port is 5-hour drive in Laredo, Texas (yes, it's a "land port").
      7. Houston is 7.5 million people
      8. We all know that Austin is the only new "Super city" in the US. All others have destroyed themselves. It will be the new Tech hub for the next 50 years
      As a southern Californian native, sadly, this has been proven more and more true as the government has become more hostile and aggressive to all.
      9. Texas has some of the strongest property rights in the Union.
      10. We could build it to have subways to move large quantities of chemicals or products in or out so as not to cause issues with other businesses.
      11. Texas will eventually be a peer-to-peer rival to California economically, politically and in population (it will never be as Beautiful as CA though...). It would be best for all of us to consider having a say in the potential "new Shenzhen" of Texas, by us being the designers and builders and financial investors in it.
      12. This could be a great win for Greg Abbott/ Who ever is the US President (For future people this is written on 11/5/2024) to announce, assuming he won his reelection (if that is this year or in 2 years).
      13. Texas already has some of the cheapest energy prices in the US, and they will stay cheap as more companies and manufacturing moves there, especially if we can potentially be the reason that Texas has energy that "is too cheap to meter" as was once the promise of the abundant future promised by clean nuclear power.
      14. Also the AI Farms could offer to subsidize the residential energy costs of the poorest 50% assuming they produce excess energy to win over the people - just an idea.

    • @caladr9367
      @caladr9367 8 днів тому

      Buying an island in the Caribbean and creating a new city-state could be a better option. Singapore vs Shenzhen/Shanghai

    • @eriklondon2946
      @eriklondon2946 8 днів тому

      @@caladr9367 Not sure it needs to be an island. Just a territory that would be treated differently regarding taxation, at least temporarily.
      One of the obvious issues with this idea is it's not connected to the mainland for shipping logistic supplies. Jones Act would make this much more difficult. It would have to be an unincorporated part of the US aka a "territory", not direct ownership of, but could somehow have tunnels or floating railways (which is a technology I have not yet heard of).

    • @gavinknight8560
      @gavinknight8560 4 дні тому

      @@caladr9367that’s an old form of madness you have there.

  • @kevincarter766
    @kevincarter766 6 днів тому +1

    Great conversation

  • @geaca3222
    @geaca3222 5 днів тому

    There should be much more investments in AI safety and much more effort put into AI safety regulations.

  • @Arcticwhir
    @Arcticwhir 14 днів тому +3

    btw so far I see ALOT of chinese and american collaboration (just looking at arxiv). So I don't really see a us vs them tech race, china is open sourcing many SOTA models, the US companies are. This could be mainly due to the capitalist led nature of these technologies as opposed to military rockets/space. So if the software is close to equivalent, the bottleneck would be hardware, compute US is ahead, any other type of manufacturing China is ahead. I think a bifurcation would lead to more trouble then continuing to work with each other.

  • @paxtun-k8o
    @paxtun-k8o 13 днів тому

    Appreciate the detailed breakdown! I have a quick question: My OKX wallet holds some USDT, and I have the seed phrase. (alarm fetch churn bridge exercise tape speak race clerk couch crater letter). What's the best way to send them to Binance?

  • @ComedyGary
    @ComedyGary 13 днів тому

    PLEEEZ put the date it’s originally published in the description. I’m halfway through it before I realize from comments about politics how old this actually is.

  • @clempadin8051
    @clempadin8051 12 днів тому

    Wish there was a transcript! You said something about 'the Cheney rule'? or 'the Cheney principle'? I've been trying to find it but can't seem to. Anyway, it was something like 'if there's a 10% chance of something catastrophic happening, then it needs a 100% of your attention'? Can't believe I would agree with him. ;-) I have worked in software development since the time I was in high school... jeez, like 50 years! OMG! Anyway, the one lesson I've learned from it is that all technology fails. So that's why I'm not a fan of nuclear. From my perspective, a nuclear system will fail and where it fails will be unlivable for 10,000 years or so. That kinda follows this 'Cheney principle'. And so what about AI? From my perspective and the Cheney Rule, AI will fail. The question is what are the implications of that failure?

  • @user-tx9zg5mz5p
    @user-tx9zg5mz5p 14 днів тому +3

    Government (military) will obtain AGI/ASI before all 😊

  • @martongr
    @martongr 12 днів тому

    I think it’s not fair to compare AI risk and nuclear risk to fire or cars or horses: the first category is systemic risk and the second is individual.
    Other than that, loved the podcast!

  • @geaca3222
    @geaca3222 5 днів тому

    But what about the warnings by two godfathers of AI, Geoffrey Hinton and Yoshua Bengio, and many other leading AI technicians, about a very real, possible AI existential threat! Like by creating subgoals when AIs are enabled to act autonomously, and one of those subgoals would be to get more control. When they'll surpass us in general intelligence, which we won't necessarily notice, we won't be able to keep up with them and imagine what they're planning or strategizing. Surely these people are ultimate generative AI technical experts and truly know what they're talking about. As a layperson I choose to trust their information and warnings before anyone else.

  • @kanewaterworth3711
    @kanewaterworth3711 14 днів тому +1

    30:30 should be embodied with human values but we do go to war a lot. This is what AI needs to fix

  • @Ccrippie
    @Ccrippie 14 днів тому +2

    The Ai bot now has $5mm in bitcoin at its disposal followed by Take-Off" is hypothetical..... 😂😂😂😂

  • @patrickmortenson4266
    @patrickmortenson4266 11 днів тому

    Marc Andreason channeling Gary Marcus here 😂

  • @elliptictree
    @elliptictree 10 днів тому +2

    Interesting

  • @seeds.ffmpeg
    @seeds.ffmpeg 13 днів тому

    They didn't really come back to the copyright and copyright infringement issue like they did with all the other issues they mentioned at the beginning.

  • @ayenewwolde4393
    @ayenewwolde4393 4 дні тому

    how do I know the podcast was presented by Ai

  • @GatherVerse
    @GatherVerse 14 днів тому

    You should get the team over at Hyper Policy involved in this discussion. They're doing really well in the policy narative.

  • @superrnovaa3
    @superrnovaa3 12 днів тому

    google researcher said a week or two ago that the pieces for self improving ai are here

  • @anuzis
    @anuzis 11 днів тому

    This episode felt like cherry picking the weakest counter arguments and brushing them off. There are intelligent views on both sides. It'd be nice to hear a followup where you genuinely consider the strongest opposing arguments and address them head on.

  • @pikaso6586
    @pikaso6586 12 днів тому +1

    I've noticed that people who speak very quickly often seem more focused on sounding and appearing smart than on actually being understood.
    This can sometimes be a way to cover up flaws in their reasoning or gaps in their knowledge. Speaking this way isn’t natural, nor is it a sign of superior intelligence.
    It's generally wise to approach fast talkers with caution and avoid being easily impressed.

  • @superrnovaa3
    @superrnovaa3 12 днів тому

    also i think you are strawmaning the argument. The point is that if it is more intelligent than us, it will have flows of information that given any time frame for a human, in the same way theres no amount of time you could take to teach a dog physics. It worked out for dogs clearly, hopefully it works out for us.

  • @ilyasorokin2494
    @ilyasorokin2494 5 днів тому

    I like their perspective on crypto and AI, they talk so smart. Until they explain how capitalism vs socialism works. This feels out of context when they try to switch between crypto and capitalism in their conversation.

  • @Peerbase
    @Peerbase 7 днів тому +1

    I'd love to here more about network states..
    I produced a documentary about the rise of startup societies filmed in Prospera featuring Balaji Srinivasan.
    ua-cam.com/video/8KhnY7Uk2es/v-deo.html

  • @uiZr138b
    @uiZr138b 14 днів тому

    one more question : did you buy the books in the backgground only for decoration? so that everbody realizes how smart you are? hahaha.. I know you, cant fool me.

  • @Arcticwhir
    @Arcticwhir 14 днів тому

    I can understand why openai closed their Superalignment Team and some "safety" folks, these people sole purpose to to sit down and imagine endless fictional scenarios of how AI could go wrong. Its a noble goal and department, but it has incorrect incentives. I agree much more with Sams incremental release structure. The dangerous thing would be to release super intelligence all of a sudden - the fact its super intelligence means its bounds are too large to test for.
    One thing I probably do expect, Meta might eventually stop releasing SOTA open source ai.

  • @uiZr138b
    @uiZr138b 14 днів тому

    I love trump for crypto and deportation. thx marc ! I hope you get even more rich!!!!!!!

  • @uiZr138b
    @uiZr138b 14 днів тому

    hi marki marc , i have question about ai. what is better, ai or deportation? you could also combine , what u think? new rainessance, is what? not sure. democracy nobody needs as long as builders have money. have a good day.

  • @fedamecan
    @fedamecan 14 днів тому +3

    Stop yourself.... let's say with tariffs....oh wait

    • @eriklondon2946
      @eriklondon2946 14 днів тому

      Or by starting global thermonuclear WWIII. Of those two options: Tariffs all day every day.

    • @eriklondon2946
      @eriklondon2946 14 днів тому

      Or starting WWIII. Of those two options, bring on the tariffs.

  • @michaelmeram6156
    @michaelmeram6156 14 днів тому

    Wattuppp

  • @ReframingUcenje
    @ReframingUcenje 14 днів тому +3

    It would be good for you to wear over-the-head headphones to flatten the top of your head a bit.

    • @Ccrippie
      @Ccrippie 14 днів тому

      That "egg head" has collected $2BB in Net Worth, which doesn't accurately represent his influence and power

    • @antediluvianclockwork9769
      @antediluvianclockwork9769 14 днів тому

      it would be good for you to learn some English

  • @baojuice8860
    @baojuice8860 13 днів тому

    such big egg head

  • @AnglicaBerry-q8p
    @AnglicaBerry-q8p 8 днів тому

    judgmentcallpodcast covers this. AI policy update: safety, risks