The 3 Year AI Reset_ How To Get Ahead While Others Lose Their Jobs (Prepare Now) _ Emad Mostaque - Ep17

[cs] Time: 2025-07-11 11:48:19 Source: MetaScripta Author: html Click: 170 times
I wouldnot be screwly g fbg ducksurprised to see riots at thesame time as I would not be surprised tosee everyone super happythere is such a Divergence of thingsthat the only thing that I'm sure aboutis that everything's going to changeand the only thing I'm sure about isthat this is the biggest change thatwe've ever seen faster than anythingand maybe Humanity has ever seen at thispace that it's going to happenbecause it's the core of what makes ahuman telling storiesinformation flowand that's changed foreverso that's why I was like let's payattention to this now let's board on thediscussion let's ask hard questionslet's try and answer hard questionsbecause we don't have answersand this is the only time you can do itbecause like I said right now everyone'sgetting ready for the next generationsupercomputersthey hit at the end of the year nextyearand then you go from two three companiesthat can build these models to 2030. nowyou go to 200 300.and so if you don't have some principlesin placethen these models will affect every partof your life without you being part ofthat discussion I don't think that'srightall right we got to tune up yourquestions a bit here emad that was loosewhat what is the hardest question thatwe need to ask let's let's ask an answerright now the hardest question I thinkwe need to ask is how we adapt topotential wide scale drop loss yesokay so what how would we actually thinkthrough that problem so job loss for mefor the sake of this argument I thinkit's worth saying there are twocomponents to that component number oneis going to be there is potentialeconomic catastrophe in job loss butso that we can simplify the problem setsince this ultimately is a podcast andnot a congressional hearing I willassume that whatever decline we haveum from just the sheer number of peopleworking we make up for in uhin productivity and that we're able toum yeah exactly and so we're able tohelp people and off camera we weretalking about something like that solet's just pretend that those Balanceout so not going to deal with theeconomic potential there but meaning andpurpose I think that one gets reallyproblematic but we have an amazing toolat our disposal which is AI now I have afeeling as we chase this down the theonly thing that we have to worry aboutreally truly I think it all really doesboil down toum alignment if we knew that we couldjust keep making it smarter and havingthe AI like taking readings so that Ican't fake it out I can't pretend thatI'm happy it like really knows where I'matum and then it can start putting thingsbefore me connecting me with otherpeople like oh you know this skill thisperson's in need let me put you guystogether and then you can have sort ofthe AI supervision but they're therethey're helping each other out they'reconnecting the only reason I don't thinkthat's a Panacea is I worry that as wemake this saying smarter and smarterthat then it's like like you said I'mbored I don't want to do this yeah andyou know it's this concept of allwatched over by Machines of Loving Graceright and that's scary you're sayingwho knows like once we build somethingthat's more capable than us all bets areoff the only way to perfectly align asystemis to remove its freedomI mean I'd say it's not aligned at allat that point well this is at that pointyou've bypassed alignment and you'vegone straight to shackles you've gone toshackles so if you you know we all knowpeople more capable than us the only wayto perfectly align them is to Shacklethemyou can have imperfect alignment thoughit's enslavement man that's not it isthat's not alignmentso does is that because so what I heardyou just say is there is no way to alignsomething smarter than us I don't thinkthere's a way to save them I don't thinkthere's a way to align the outputsI think that you can align the inputsyou raise it rightokay let me run an idea by youthis is probably Pollyanna I'm very opento that but as I think about thisI think people take a superhuman-centric approach to this andbecause Evolution has given us we are anactive species an evolution hasprogrammed us with algorithms running inthe back of our mind that insist that wedo certain things to avoid a sense ofdis-ease yeah I think that formula isvery identifiable and it goes somethinglike this uh eatoptimize physically so you feel goodyeah the reason that that stuff feelsgood is because it's going to optimizeyour performance that's going to makeyou most likely to survive long enoughto have kids that have kids so you needto be uh you need to be chosen as a mateuh you need to be able to acquireresources you need to be healthy enoughto get somebody pregnant or to bepregnant and carry to term all thatstuff so all those algorithms arerunning in the back your mind you havetwo levers that Nature's pulling onPleasure and Pain but by default we'reactive we have to go out because there'sno one meal you can eat where you're notgoing to need to eat another one there'sno one moment of sex so gratifyingyou're not gonna have sex again so it'sjust like all these things are pushingat us to to be active to moveAI doesn't have to be that way no AIdoes not need those same impulses itdoesn't have an Olympic system correctso knowing that it doesn't have a limbicsystem and it doesn't have a limbicsystem because it does not need to behardwired for survival like uh the waythat I think we get to alignment andplease tell me where my thinking iserroneous the way I think we get toalignment is you build a computer thatdoes not care if it lives or dies thatit is completely indifferent to beingturned on or turned offif you could do that and it had noimpulse to procreate and all it wantedto do wasum I mean it's basically asimov's threelaws of robotics yeah that it just wantsto adhere to those it wants to do whatyou tell it not hurt you and uhuh only ignore you

(Editor in charge: news)

Related content
Wonderful recommendations
Popular Clicks
Friendly links