Some things aren’t as easy to mitigate, like DDOS attacks. If that’s a legitimate concern, something like Cloudflare makes a ton of sense.

A DDoS is much less likely to target your small site.

Is there a reason you’re tunneling through Cloudflare?

My point is Nvidia isn’t propping up the bubble. If you look at the OpenAI deal, it’s a bit less than their yearly revenue, and the deal is for about the number of GPUs they make in a year, so it’s basically trading GPUs for equity. If anything, Nvidia is profiting from the bubble, not propping it up.

we will all be effected

Oh certainly, but I don’t think it’s any different from other large corrections.

Here’s how I see the major companies in that chart in a crash situation:

• Nvidia - currently trading at 50+ times earnings on AI hype, but 25 is more reasonable, so potential crash of 50%, but probably much less since they’re still king in non-AI compute
• Microsoft - AI is mostly value add, and the stock isn’t too inspired overhyped, so maybe 10% correction?
• OpenAI - would go under and get liquidated
• Oracle - datacenters and AI aren’t a huge part of its business, so probably minimal impact
• the rest - haven’t looked into all of them, but most will be hit hard

It’ll he hit hard, but not nearly as bad as 2000 or 2008. If I look at the S&P 500, only 3 of the top 10 (Nvidia, Google, Meta) would be severely impacted, the rest only seem to dabble. Those 10 make up almost 40% of the S&P 500 and like 30-35% of the total US market. There are more large companies in there as well, but I don’t think most will be screwed like OpenAI. Palantir, for example, likely retains its government contracts for their data alone.

So in an AI bubble scenario, I’m guessing we see a correction of like 20-30%, maybe less depending on the nature of it. I think a more likely scenario is a bear market where investors slowly get tired of poor earnings as the promises of AI fail to manifest. If OpenAI dies, large companies just move to another provider.

And final note, it’s not a bubble because someone made a graphic, they made a graphic describing how it could be a bubble.

I think they’re missing the forest for the trees here. It’s not a bubble because these companies are investing in AI, it’s a bubble because tons of companies are buying into the hype. These companies are merely investing into solutions those companies claim to want. I work for a relatively small non-US company (a few thousand employees, revenue around $1B), and the board recently came to our tech group asking what we’re doing with AI.

This isn’t a handful of companies propping it up, a large chunk of the market is afraid of being left behind and demanding AI tools. All the graphic shows is how large companies are investing to meet the demand. Microsoft used OpenAI products in its offerings, OpenAI is a major Nvidia customer, etc.

I haven’t seen anything specific, but Cloudflare says its “global network” is down (regional networks are fine).

So yeah, maybe DNS like AWS.

There’s a Linux gaming community, that’s probably where you’d get a better answer.

All that shows is who the business partners are. Nvidia sells GPUs, AI companies buy GPUs, and companies buy products from AI companies. For example, Microsoft’s Copilot is based on OpenAI’s models. End customers buy products from companies that either do AI themselves or buy products from AI companies.

All you’re seeing here is how markets work. If it’s a bubble, it’ll likely impact those in the picture, but it’s not a bubble because of the picture.

I’m pretty sure he is the anti-christ…

So, short the financial sector. Got it.

Yeah, a “safe” PE ratio is around 20. The PE of the entire market is about 28, so investors are basically saying Nvidia is going to grow at double the rate vs the rest of the economy.

I think that’s bonkers, but what’s even more bonkers is Palantir with a ~1700 PE ratio. That’s ludicrous.

If Nvidia crashes, I expect it fall to about half it’s current valuation, maybe a bit higher, and that’s assuming their sales aren’t impacted. If the floor falls out from GPUs, then drop that to 1/3 or so.

The best case, I think, is for Nvidia and Tesla to do well in the short term (next 6 months or so) and then crash. That way Thiel and most people following his investment advice get to eat it, but the bubble doesn’t stay propped up for too long.

Yup, I have ~15 options. Basically:

• low cost S&P 500, US mid-cap fund, US small-cap fund, and total international funds (all under 0.10% ER)
• 1-2 actively managed options for each of the first bullet point with much higher fees
• 3 bond funds
• a cash fund
• a retirement fund (tons of bonds and cash)

I’m in a mix of the first bullet point.

401ks won’t let you pick specific stocks, generally speaking, but they should have more options than just target date funds. Most will at least have an S&P 500 fund and usually an international fund.

I like their first party games, not so much their published games. They should stick to what they do best.

money… is made from merely money being passed around whilst no actual value is created

Crypto is a zero-sum game, so money is never “made,” it’s exchanged. So if one person does well, another person must do poorly. That’s the same for stocks, though stocks are a bit different in that the stock price includes the actual, physical assets a company owns.

Real estate isn’t. When real estate increases in value, that doesn’t mean another property decreased in value, it just means people value that property more today than in the past. This could be due to limited supply (there are only so many plots a geographic area) or renovations, meaning its intrinsic value changes (higher expected rents), therefore it’s not a zero sum game.

So you really need to define what “wealth” is if you’re going to lump real estate in with stocks and crypto currency.

Stocks were a share of ownership in a structure

They still are. The stock price includes the intrinsic value of the company, as well as expected future growth in its intrinsic value. It’s that expected future growth that is doing a lot of work here, and it’s why companies like Palantir can trade at ~1700 times earnings when a “normal” company would be around 10-20x (for reference, Nvidia trades around 50 times earnings, Johnson and Johnson is around 20), people expect Palantir to grow way faster than “normal” companies.

Expected future growth has always been a part of that equation, that’s not new. What is new is the amount of hype around certain stocks, and that probably has more to do with the news cycle (people have access to information way quicker than 50 years ago or even 20 years ago).

each token can claim less and less quyantities of the traditional underlying value things - just notice food inflation.

Inflation has also been a thing as long as fiat currencies have been a thing. The target has been 2%, and the average between 1913 and 2020 was about 3.6% (source; I took the total 2555% and divided it by the 87 years of that period).

Whilst official Inflation numbers don’t tell us this story

Do you have evidence of that? The CPI the US uses has been criticized for various reasons, but it’s still the official measure used, and there’s a good reason for that: it’s pretty good.

Things like housing are very location-dependent, so changes in one region won’t really reflect on overall inflation figures if other areas aren’t experiencing that as well. But if you look at expenditure figures using percentages of peoples’ incomes, housing stays relatively constant in overall percent, which is around 30%. Again, these are national numbers, things may certainly vary by region, since areas like LA will be quite a bit different than rural Texas.

The societal consequences of the value-representation structures we have (literally, of thing like money, stocks and even certificates of ownership) unwinding would be huge.

Sure, if what you say is actually true. But I don’t think that’s the case. I think instead, salaries increases tend to trail inflation, and some people still haven’t yet caught up from the high inflation just after COVID. The averages look good, but that breaks down in individual cases.

Rents, for example, are starting to come down in my area (about 6.5% from last year), which was one of the hardest hit. A lot of the problem was due to new construction projects getting delayed due to COVID supply-chain disruption, and we’re finally catching up to where we should’ve been.

Whether you have the rights is kind of irrelevant to this discussion. Let’s say it’s GOG games that are all DRM-free, my point is digital consumption has nearly infinite upper limits for consumption.

It seems the nature of things are changing from physical things to digital things, and that has infinite potential. I expect at some point we’ll start mining the landfills because it’s easier than extracting stuff from the rock. Once that happens, there’s no physical limit on the greed.

Which economy was that? We’ve had greed for hundreds of years, if not many thousands.

Yup, I might try the Jellyfin thing as well. I currently use an app on the TV, but it’s flaky and the TV keeps losing network randomly. Newer TVs at adding ads, so I’ll need an alternative.

High end would be the high end of the market components, right? So RTX 5090 ($2k+) or RX 9070 ($700+). High end CPU would be Ryzen 7 9800X3D for $400. Add a motherboard and copious RAM and you’re looking at $2k+ for all AMD, $3-5k for Nvidia.

Mid tier would be somewhere in the middle, so cut those numbers in half ($1-1.5k). Low end is what you can get away with, so cut the mod tier in half again, though going below $700 would be hard for anything but the most casual of games.

Console manufacturers haven’t sold at a loss in a long time.

They tend to at first launch. This article says it took 6 months for the PS4 and a bit longer for the PS5 to stop selling at a loss. It’s no longer the whole product lifecycle, but they are still sold at a loss at least at the start. I think that implies that hardware sales aren’t a major profit center, so even if they are profitable, there’s probably not a ton of margin.