Latest / Black Box Theory Podcast / Is AI Actually Making Us More Productive?
Transcript
- Malcolm: And I'm not gonna lie, bro, I think people are starting to come to the realization, bruh, bruh, AI is cooked, bruh. Like it it's lucky not. Like w like bro, like yes, it helps us, bro. But I'm not gonna lie, if we're actually talking about genuine return on investment for the hundreds of billions of dollars that these guys are spending, are we really getting a return on our investment right now? Like are we really seeing mean of meaningful gains in productivity and now I think there are some specific examples where it's like AI has a direct, like quantifiable increase on efficiency or like productivity. So like coding, right? Co I don't think coding is a professional like I like c like AI has significantly increased the speed in which software engineers or technical devs or whatever can like produce code and merge PRs and like like like put out like you know new code stuff like that. ⁓ obviously the the like the models have become like increasingly like like much much better at writing code and and they're like I think even like within claw like almost everything is written internally with claw like no one actually codes anymore, right? That's that's one very specific sector. We're talking about everywhere else, bro. Are we really seeing like the gains that essentially like does the does the the money that these hyperscalers and all these companies are kinda putting into AI, does it like are the returns that we're seeing are they justified? Like does it make sense? Like I feel like ⁓ and it's kind of like it's really kind of scary because it's like I like, you know, f folks across several sectors are kind of having this realization at the same time. AI is low key the only reason that the stock market is not like free falling like to red right now. And it's kind of like ⁓ it's it's kinda scary to think about that, bro. I'm not gonna lie. Like we don't ⁓ we're not saying the returns that we thought we were gonna see. But like, is anybody actually gonna stop spending money on it and stop buying data centers? I don't think I don't think so, Matt. Thomas: I mean. That's that's facts and and I think honestly, when it comes to all this this AI stuff, bro, I think that there's ⁓ a problem, like even just personally, I I'm like contemplating, you know, like what do I what am I paying for Claude for? You know, and it's like I'm sick of seeing these AI websites that people are just throwing stuff at AI, hoping that AI does every little thing. ⁓ and also in terms of what AI is doing. ⁓ it's not it's it's not what we need, like thought. And I I think they're they're pushing the envelope a little too far, right? Like for example, we the compute costs keep going higher, the token price goes higher. But like you said, for scoding and some stuff, and if you're doing like simple, not simple, but things with Excel sheets and different things like that, you might not need the most premium, ⁓ whatchamacallit, the the most premium like platform. So I really think that it's important to think, okay, can we have tiers that are more affordable or that don't charge by the token? Because if you're simply just ⁓ if you're if you're literally just doing Excel sheets in in different sh like sheets, you don't need the most powerful AI. You need an AI that can parse through data, that can ⁓ make ⁓ guesstimations and and do ⁓ quantitative analysis. You don't need the most powerful fable or whatever the case is. So I think that they should consider offering tiers so that the return on investment is useful. Like if the claw price goes up for the stuff that I'm personally using claw for, I I wouldn't need that. And it's like that's pushing people away. ⁓ but that's just my take on it. Malcolm: No. Yeah. Yeah. Nah. Nah that makes sense.