Posts by KayArres@mastodon.nl
(DIR) Post #AtczQ9zTIUTxZlDkvI by KayArres@mastodon.nl
2025-04-30T20:34:24Z
0 likes, 0 repeats
@yura @evrys Oh and if you like webcomics, try Always Raining Here, which is completely finished so you can binge it! https://www.alwaysraininghere.com/index.php/arh/cover-2/
(DIR) Post #Av1g2zNPPylw38AxDU by KayArres@mastodon.nl
2025-06-11T12:50:58Z
0 likes, 0 repeats
@yura The kind of boy who reads Nagatoro in his spare time over here~
(DIR) Post #AwAaW9oRmaJFrM7Oee by KayArres@mastodon.nl
2025-07-15T21:20:14Z
0 likes, 0 repeats
@yura Hellsing
(DIR) Post #Ayj80k1wKykMVjG0jg by KayArres@mastodon.nl
2025-09-29T15:17:48Z
0 likes, 1 repeats
I exist within a social bubble that is mostly pretty negative on AI as a technology, with some exceptions. I do not feel very strongly about the tech itself, though I'd say what feelings I have regarding it are largely negative. Still, I feel that as a technology, it is simply not my field and that to have strong opinions on anything you know little about is a good way of making an idiot out of yourself.However, I do know finance, and the finances of AI kind of worry me.
(DIR) Post #Ayj80lNFLFNsg6EXaK by KayArres@mastodon.nl
2025-09-29T15:22:17Z
0 likes, 0 repeats
Nvidea is currently the most highly valued company on Earth, with a market cap of 4.44 trillion dollars as of time of writing. To put that into perspective, Apple was the first company to reach 1 trillion in evaluation back in 2018. Even accounting for inflation (which could be estimated to be about 3.7 % yearly since 2018 giving us a cumulative inflation of roughly 29%), Nvidea has almost quadrupled that number.
(DIR) Post #Ayj80mg4Uk2Kim35ZA by KayArres@mastodon.nl
2025-09-29T15:24:17Z
0 likes, 0 repeats
To be clear, I am not saying "number big so number fake, capitalism dumb". Nvidea is in an extremely enviable position: They are selling shovels during a gold rush. It does not fundamentally matter to Nvidea if the companies who buy their products, those products being graphics cards used for AI development, do well because Nvidea has already gotten their money. However, just accounting for that revenue does not justify Nvidea's price.
(DIR) Post #Ayj80nXxGP8dPtG7YO by KayArres@mastodon.nl
2025-09-29T15:27:01Z
0 likes, 0 repeats
In the stock market, there is something known as the "p/e ratio" or "price-to-earnings ratio". It's a really simple concept: It is the amount of money a company is valued at, compared to the amount of money the company makes in profit.
(DIR) Post #Ayj80oRxu9wQDbSqrA by KayArres@mastodon.nl
2025-09-29T15:27:37Z
0 likes, 0 repeats
When you buy a company that is not really growing very much, but doing well for itself, you can expect a p/e ratio of about 4-6. If a company is seen as having a big growth potential, you pay a premium for that, so the p/e ratio might be something like 10 or 12. Nvidea's p/e ratio is 50.
(DIR) Post #Ayj80pUq0xpjSnoeyO by KayArres@mastodon.nl
2025-09-29T15:30:17Z
1 likes, 0 repeats
This means that a lot of the value of Nvidea as a company, according to the market, is the future value of the company. The market is expecting Nvidea to keep getting bigger in order to eventually get their money back. Personally, I think it is hard to imagine a company ever being a good buy at a p/e ratio of 50+, but regardless, it means that Nvidea's value is much more reliant on the continued success of the AI industry than might seem at first glance. Future shovels are priced in.
(DIR) Post #Ayj80x85dQOH8ZEbcu by KayArres@mastodon.nl
2025-09-29T15:32:24Z
0 likes, 0 repeats
And to be clear, that could potentially happen. If you had bought Amazon at a p/e ratio of 50 in 2001, you would still have made tons of money by now (although the p/e ratio of Amazon at the time was much lower). But what I want to emphasize is that Nvidea is only worth this much if the AI thing is not a bubble. They need it not to be a bubble. And that creates a lot of perverse incentives.
(DIR) Post #Ayj814EjD8rNBBOVRg by KayArres@mastodon.nl
2025-09-29T15:34:43Z
0 likes, 0 repeats
Back in the Dot-Com bubble days, there was a thing known as "vendor financing". The basics of this is that in order to increase earnings (which is the money going into a company without subtracting the costs), which will make it look like your company is growing, companies would literally loan out money to their customers to buy their products at scale. This is not economically efficient, but it does make it look like you are selling a ton of products.
(DIR) Post #Ayj81BXm2jFnqHMJbU by KayArres@mastodon.nl
2025-09-29T15:37:04Z
0 likes, 0 repeats
By pumping your earnings, you attract outside capital/investors. Then, with that money, you can keep the cycle going because you cannot just loan out money and then get it back. Other companies have overhead, salaries to pay, etc. So it's not a good way to be profitable, but it's a good way to seem like you're growing, which gives you access to more money, which you can use to pump your growth more, etc.This does not make money. The investors are losing money, and thus the money dries up.
(DIR) Post #Ayj81IZ5wDTBcP1y8O by KayArres@mastodon.nl
2025-09-29T15:38:14Z
1 likes, 0 repeats
When the investors are fed up with not seeing profits (earnings are high, profits are not), the bubble pops. Vendor-financing did not save dot-com companies, but it slowed their demise and worsened the eventual crisis.The reason I bring this up is because Nvidea is spending *a lot* of money financing AI companies.
(DIR) Post #Ayj81PSGK1ACzEt6xM by KayArres@mastodon.nl
2025-09-29T15:40:58Z
0 likes, 0 repeats
Now they are not directly financing purchases of their cards. What they are instead doing is minting new unicorn start-ups every week, sending them hundreds of millions in investments. Those companies then turn around to companies like Oracle, who build them data centers. And Oracle, of course, buys its cards that it builds the data centers with from Nvidea. There is an image I am reminded of.
(DIR) Post #Ayj81WOGXH7sUs4WCO by KayArres@mastodon.nl
2025-09-29T15:42:34Z
0 likes, 0 repeats
True believers will tell you one of two things: 1. AI companies will become massively profitable because the product they are creating is going to find a big market and all the investment will be worth it2. This is all just wheel-spinning until AGI is developed, and it is worth it to throw tons of money at development because being first is all that matters.Let's adress my thoughts on both.
(DIR) Post #Ayj81d6RZw1tJcmstM by KayArres@mastodon.nl
2025-09-29T15:45:17Z
0 likes, 0 repeats
For point number one, I think there is some market for AI. It is clear that there are people who do have use cases for it. From programmers to customer support, there are already use cases. However, the amount of investment in these companies is vast in relation to the actual market the product has found. AI is making money, but it is not making close to enough money. I am not saying AI is worthless, but it sure seems overvalued, and simply saying the market will be found seems...optimistic.
(DIR) Post #Ayj81kGGxn3DW8RKgS by KayArres@mastodon.nl
2025-09-29T15:46:31Z
1 likes, 0 repeats
As for point two....like, maybe? I am not an AI expert, so maybe spicy autocorrect is indeed the step in the tech tree right before superintelligence. Some smart people sure seem to think so. But a lot of smart people also seem not as convinced, and boy are we making a big bet on this.
(DIR) Post #Ayj81rRsF3URo8vCEq by KayArres@mastodon.nl
2025-09-29T15:48:29Z
1 likes, 1 repeats
So why does this matter to you? Most of the people seeing this post are probably pretty skeptical of AI. Even if you're not, that doesn't mean you're deeply invested in it.Except, a lot of people are deeply invested in AI regardless of their beliefs. If you have a 401K or non-American equivelant, it is likely buying stocks in the S&P 500 based on the size of the companies. That means most of the value of your pension is in Nvidea and other AI companies that crowd the top end of the market.
(DIR) Post #Ayj81yyOGYex91BlOS by KayArres@mastodon.nl
2025-09-29T15:50:10Z
0 likes, 0 repeats
And if this stuff doesn't work out, if they cannot fake it till they make it, that money is gone. You "took the risk" by investing in them, and when it doesn't turn a profit then the investors, the pension funds, they take the hit. I genuinely do hope I'm wrong. I *love* the idea of us finding some awesome technology that superboosts our productivity and brings more wealth into the system. If that does happen, because of those pension funds buying it up, we all get a slice of that.
(DIR) Post #Ayj826S4SbYoL68480 by KayArres@mastodon.nl
2025-09-29T15:51:31Z
1 likes, 1 repeats
But I do worry we're witnessing one of the biggest market manias of our time, and that we are spending way too much money to optimize what may turn out to be a relatively niche technology. And I do mean relatively, because AI would have to be astoundingly non-niche in order to justify these valuations. We'll see how well this post ages in 5-10 years.