for thank you for joining us my pleasure you have written a newsletter a couple months ago about the peak of AI we had reached Peak AI you also made a podcast episode about at the podcast you enumerated four reasons would you mind going into it a bit these things are too expensive they do not make money they burn cash and they require insane amounts of compute power on top of insane amounts of energy to run that compute power and the problem is with that is it’s not coming down they keep talking about these things being more efficient being better but better just means more they’re not trying to be more efficient they’re not trying to get more out of one GPU when you’re not trying to make it more efficient and it’s already deeply unprofitable which it is at some point you’re going to run out of money and now ai is putting a massive strain on global power gits as well and then there’s the other problem to make these models better to actually do more things to actually I don’t know realize any of the promises Sam mman has made at all you have to train them on data data that you’re running out of they’ve already eaten the entire internet they’ve trained these models on these massive data sets from the internet but there’s a researcher who claimed that to train GPT 5 they’re going to require four to five times the amount of available data to actually make it work which is insane and also does not exist what does Chad GPT do better exactly what is open ai’s actual goal to make AGI what is their product they sell why is this useful these are things that should roll off the tongue when it comes to the next big thing what I’m not saying is that generative AI things like llms like GPT are useless they’re not useless they can generate some boilerplate code that is useful but you can’t just trust it because the code’s probably got some weird stuff in it they can generate anod business copy but nothing really great the output is always mediocre people are using it to automate things it is probably a really useful Cloud compute solution right going through M massive data sets for example but the problem is it’s almost like they think that this kind of useful business thing warrants this insane investment this incredible amount of compute this remarkable amount of physical Hardware these crazy energy demands it doesn’t yeah a lot of people have been saying that it looks a lot like a bubble Wall Street is actually surprising ly itchy they don’t have the patience that they used to and the problem is that the people in the street don’t really know what they’re talking about these people aren’t Cloud computer experts they don’t really even understand what AI means and it’s terrifying when you realize that cuz you’re like wow the entire economy is based on The Vibes of a bunch of people with a lot of money that don’t understand what’s happening is that good it’s not good it’s very bad the long and short of it is they need these stocks to start growing gr Revenue in a way that allows them to kind of extrapolate and say okay this is growing enough so this company will be worth $1 trillion the problem is it’s barely making any money at all and when I say making money I mean producing Revenue if I mean making money is in profit absolutely none of these big tech companies doing AI stuff on making money right now the street needs to see Real Results soon they need to see them in the next few quarters because if it’s Microsoft Amazon all these companies they’ve been burying their AI earnings in different parts they don’t have to say this is actually losing US money the moment that comes out the street is going to start freaking out because guess what if the consensus of the stock market decides that AI is not actually the next hyper growth Market every single company that’s Bet On It is going to get trash and going to get trashed hard what would a bubble bursting look like I think that one of the major Cloud compute companies Amazon perhaps or Google or really Microsoft has a bad earnings season maybe next quarter maybe the quarter after that because the street will start saying you keep talking about AI as this big thing that’s so awesome where’s the money honey where is it where’s the cash where is the growth where is the thing we one and I don’t think Microsoft is going to have much of an answer at that point the stocks that are so thoroughly invested with AI will start to Rattle and fall apart and then we’ll see a crash of sorts I actually don’t know how crazy it will go I think it could be as minor as there is a big cut in tech stocks and everyone realizes they have to move away from AI it could be as bad as Microsoft takes a 35% haircut and Sam mman and open AI get absorbed into Microsoft as a means of hiding their dirty laundry same thing with anthropic and Amazon and Google they run the same risk but nowhere near as big as open Ai and I think that just the big public Tech wash out will be bad and it will affect the stock markets globally if I’m wrong it’s honest to God a good thing cuz if I’m wrong it means that everything isn’t going to get washed out do you think there’s any way out of this situation other than the bubble bursting I think if I really had to guess I think what’s more than likely is there’s going to be some sort of lawsuit that the resolution they’re in sets a precedent that makes doing business for these companies very very very difficult is there a situation where the bubble doesn’t [Music] burst it would take all of the large language models randomly finding a way to be like 150 times more efficient than they already are but this if that happened I don’t believe that there’s actually a product progression from there that makes any of this buildout of cloud comput these massive data centers necessary I think it’s just these markets are not built to be efficient they’re not built for stability for sustainability they’re built for growth and nothing else it’s why I’m so certain that this is a bubble there are a lot of tech companies thousands and thousands of thousands of them there are tons of companies you have not heard of that make 50 to100 million do a year in Revenue they have positive growth metrics they are profitable and there are tons more that are not and the problem is most of the ones that you hear about the most are not the profitable ones because everyone’s obsessed with growth these companies these big companies they’re not built for Innovation they’re not built for people to make cool new things they’re built to build abstractions of value onto the side of something that can create Revenue growth I was going to ask you to Define growth but it seems like the goal posts for growth have changed quite a bit it’s growth at all costs so you must be growing by double digits every quarter but what does growth mean growth in users growth in revenue or oh that’s the fun part the answer is yes oh sometimes it’s growth in users sometimes it’s growth in Revenue most of the time with public companies it’s growth in revenue and customers growth is everything if you’re not getting bigger you’re not getting good I think we’re at the end of tech hypergrowth era so the era of when every few years you had like Cloud computer you had mobile apps you had new hardware that allowed you to sell new software we’ve run out I it really is that simple we’ve run out of these giant multi-trillion dollar markets that Tech can build and sell into they had Cloud they had search they had ads they had apps they had smartphones they had tablets they had personal Computing and the various directions that’s gone and they haven’t had anything for quite some time and they’re running out of ideas but not just not really ideas so much as ideas that they can sell as this is the next big thing big Tech and small Tech is not built to build small sustainable businesses and by small I even me like hundred million companies that’s not sexy enough you can’t FL an $100 million company as the future it has to be 500 million or 500 billion or 1 trillion this is what the future looks like except I don’t believe it is and when this bubble pops the tech industry is going to have a massive Reckoning on its hands if you look at this movement and you’re saying I don’t I don’t get it you’re right you are right we are in a global gaslighting of honestly citizenry around around what AI can do and it’s disgraceful it turns my stomach because I would love it if these things could actually automate things that I needed to do I would love it if tech was actually automating away real people’s actual problems versus discussions about how they might replace workers at one point and I think it’s just deeply nihilistic how these companies are approaching people and how obvious that the it is that the utility isn’t there well Mr vitan thank you very much for your time my pleasure please come again my pleasure I will
Fique informado(a) de tudo que acontece, em MeioClick®