Hi, this is Wayne again with a topic “Twitter Inc no longer exists”.
Let’S do it do what the one the one yep we have to talk about Twitter we’re just going straight down the list. Let’S do it Elon Musk pivots to AI uh uh. You know he was talking about the everything app in the past. So if you’re doing an everything app, then you have to uh chase the flavor of the month all the time, and that means right now, chasing AI uh musk has recently hired Talent from Google subsidiary deepmind and purchased around 10 000 gpus in an apparent pivot towards Development of a chat, GPT rival, um at the or some other AI based thing, I’d, maybe say open AI, not chat, gbt. At the same time, Twitter Inc has been folded into x-corp. That’S an interesting one. One of two corporate entities created in March, alongside x, dot AI, it’s unclear precisely what form this new AI project will take.
I genuinely don’t know anyone that is confident in what the heck they would be doing, but it is likely it likely has something to do with Elon Musk proposed everything app as we mentioned earlier, which he calls X according to musk. He wants to make the X app analogous to WeChat in terms of functionality. Musk has announced that Twitter x-corp will now be Auto replying to press inquiries with poop emojis, making it difficult to get straight.
A straightforward comment on the future direction of the company very cool Elon also announced Twitter subscriptions creators can charge followers for exclusive long-form tweets and hours-long video or shorter up two hours. Long video, basically upgraded super follows with Twitter forfeiting their cut of the revenue for the first year, also known as a marketing budget. Tell me you’re desperate without telling me you’re desperate yeah.
Can I can I pitch a theory here yeah. I feel like a lot of the other coverage and conversation that I’ve seen around this has been. You know what game of 4D chess is he playing uh, you know what will be, what will be the vision for the everything app or whatever, whatever else it is? Um, I actually think it is way simpler than that.
Here’S, my here’s, my cynical um! The simplest answer is probably the correct answer. Theory. I think he wants to unload Twitter, so it makes sense to fold it into something else.
So now it’s X, Corp, okay, which, like is cool or something I think that by by tying x-corp to this everything app slash, it’s working on AI kind of you know: BS Silicon Valley, startup kind of nebulous, that is a cool Factor brand, there’s more than 2 000, AI startups in the last week, if I remember correctly yeah, so it makes sense yep so so so creating value creating a perceived value by throwing his very loud noise. Making skills at a really big buzzword could give him an opportunity to dump Twitter before anyone realizes that they’re not actually doing anything other than buying gpus, which has also been in the headlines, because nvidia’s been talking about how cryptocurrency stupid AI is our future. Everyone needs to buy gpus for AI we’re in video. Let’S Go, there are other processors for yeah.
I mean Nvidia is very dominant in the space right now, but they are not going to be the only one forever yeah um. So I I actually think it’s just the simplest and you know why would he? Why would he avoid press inquiries well, because press asks uncomfortable questions, man that interview I haven’t watched that I’ve watched some clips from it. I’Ve been told I should watch it, but it’s pretty rough um, because the Press asks uncomfortable questions and doesn’t just ask them but expects a real, well thought out answer they’re, not just going to accept whatever stupid thing. You say so by stonewalling. Any serious inquiry about. What’S going on, he gives himself the only speakers, gavel or speakers staff, or you know, whatever thing the speaker holds in this particular weird timeline: we’re on um allowing himself to build this Narrative of what x-corp slash Twitter, slash, xai actually is under this shroud of secrecy, Where actually it’s just doing weird little proofs of concept with Engineers it borrows from Tesla or whatever, and then hopefully he can find someone stupid enough to put a bunch of investment money into this so that he can dump his share somehow and just get the crap Away from this dumpster fire, I think he’s just. I think this is just a loss, minimizing strategy and I think it’s it’s it. Maybe it’s just because I’m a very cynical person at this point, but it seems pretty it seems it’s pretty transparent, somewhat go in line with a move that I’ve seen from a couple little startups recently, which is why that where they, they find an engineer that they Feel like they can label as a superstar like llm, AI machine learning, engineer they pay them big Megabucks to come on board.
They buy a ton of gpus, they slap a label on their project and they just try to flip it. Like literally like that week, yep there’s some there’s some flips that have been happening very very fast, there’s also big money in this type of engineering right now, because you know Twitter buying 10 000 gpus. These are expensive. Like a100 big super chip, very big money, like 45 000 on eBay, graphics cards right now, 45, 000 or 4, 500, 45, 000. uh – is that just scalping? That’S going on right now, like is this okay, yeah yeah? How much is a hopper? Let me see if I can find the h100 GPU: let’s go: 42 000 for a Tesla, h100. H100.
Sorry, not a 100 yeah! That’S the last gen you’re thinking of my bad uh, but yeah. I don’t think these were worth that much before uh yeah, probably not um, because the problem is a lot of these companies that you know some of them might just be trying to cash in on, like the.com, 2.0 hype kind of thing. Some of them also have very good ideas and they need these gpus to power those ideas and they can’t get them. They got investor money, they’re going to get them somehow right, apparently the MSRP for this one or well.
There is no MSRP for something like this yeah. It’S typically sold as integrated, but apparently the price for these is typically around 35 000., okay by seven grand um yeah and there aren’t even many available yeah sure. So it’s it’s been uh. It’S been pretty intense, there’s! So right, what I was saying was: there’s actually a lot of money right now in being an engineer that can work on this stuff, especially if you can be super efficient with it, because if you can reduce the cost of gpus and the cost of continually processing On those gpus that you need to have um, that’s very beneficial for companies right now, the open AI has been talking about how they don’t actually believe um. A lot of these models need to be trained on bigger data sets potentially cleaner, potentially better at parsing them. Potentially, better systems powering the the actual like chat modules stuff, like that uh they’ve, said that they’re not actually working on gpt5 right now, um and they think it is more beneficial and they will step forward faster if they work on like supporting tools for what they Already have which honestly kind of makes sense to me, because when you’re talking about it, like uh the previous uh topic, where we mentioned that it will like get distracted or confused with complex goals? Well, okay, it might be more productive to work on keeping it on task by teaching it how to um. You know break apart problems and compartmentalize parts of it like you do with like, like humans, do to be completely honest. If you get this massive complicated task, instead of trying to tackle the entire thing as one big unit, you often break it down into smaller things so that you can easily more easily tackle it and keep yourself focused well working on things like that might actually be More beneficial and get you to higher perceived performance and power faster right than just training, a bigger model.
That’S going to have the same problem, um, so yeah, .