While it's true that companies like Nvidia have great revenue/earnings numbers, Cisco had real revenue back in 1999 as well. It's not that nobody's buying the hardware, it's that the people buying the hardware might not actually have a business model that justifies that much hardware purchase, and as soon as they realize that they'll stop buying hardware and suddenly the revenue for the hardware suppliers goes poof.

I suspect that, like with the dotcom bubble, there is something important going on here. It's just that nobody knows what the big important business model for AI is going to be, so they're all madly investing tons of money to try to get there. Some of these AI companies are going to turn out to be the equivalent of Amazon, some of them are going to turn out to be WebVan, and we don't know which is which yet.

CSCO market cap topped out about $530B in Y2K, edging above GE (the most valuable public company before that). It's $317B today, so there's been substantial dilution of the stock in those 25 years, even with the buybacks.

And while it's true that the valuations of tech companies in general are much higher than in Y2K, their revenues are also much higher. Apple's is $400B, and that's not from the AI bubble.

The obvious slant here is to say that Spotify is just doing what Facebook and Youtube are doing -- algorithm selects content to generate "engagement" rather than actual enjoyment, in turn to drive more user time spent on the site, so that they can sell more ads. Except that Spotify doesn't make most of their money with ads, something like 80% of their revenue is from subscription fees so that doesn't really hold up. There's very little "network effect" in music services, about the only thing that keeps me using Spotify over other streaming options is that they've got a good idea of what music I like and can find new stuff that matches it (or, more commonly, old stuff that I just didn't discover when it was new), and that's precisely what the article is claiming Spotify DOESN'T do.

You know what happens when two people bring a new type of vehicle to the same place, right? They race them. I expect that's what happened here. :)

The problem with Dvorak is that it makes using vi much harder :)

This is yet another case of people confusing the Internet with the World Wide Web. The Internet is not information -- it is a network over which information is transferred. As such it is made up of networking equipment, routers, switches, cables, etc; the collective weight of which is surely many thousands of tons.

You've never driven through Nevada, have you?

Yeah, sorry, typed the wrong word (I would have put 'balanced' there in retrospect), but didn't realize until after I'd hit submit.

The stability question isn't about the tensile strength and other mechanical properties of the structure itself, it's about the orbital mechanics. A ringworld orbiting a single star is stable on a "peak" position, where any movement away from that position results in net gravitational forces that move it further away. In contrast, the stable lagrange points are positions where small amounts of movement away from that position result in graivtational forces moving it back towards that position -- a "valley" by comparison.

On the abstraction layer -- I suspect you've never had to work on a codebase that was written without abstraction and then later ported to different hardware platforms. Too much abstraction results in code that is bloated and slow, but too little abstraction results in code that is buggy as hell and impossible to maintain. Slow but correct beats fast but wrong every time in my book.

Because "free" means that you can do what you want with your body, your life, and your property. It does not mean you can do whatever you like with someone else's body, life, or property. Your freedom with an employer is the freedom to quit and go work somewhere else, not to decide what the job entails.

This isn't necessarily Amazon's fault. In an attempt to fight traffic congestion, a lot of cities have started limiting how many parking spaces companies are allowed to install at their offices, so as to force employees onto the public transit they wouldn't use voluntarily.

From what I've been reading, the current theory seems to be that one engine failed, the pilots made a mistake diagnosing it and shut down the wrong one (this happens distressingly often), then panicked and did a no-flaps, no-gear landing despite the backup options that existed to deploy either one. That meant a high landing speed, no brakes, and a plane that "floated" in ground effect longer than normal so it touched down halfway down the runway. Then it hit the antenna berm/wall. Note that some distance beyond the wall are things like trees and houses, so letting the plane slide further into those would not necessarily have produced a better outcome.

The investigation will answer these questions. If the above is true then while the wall would be a contributing factor, it would not be primary cause.

SSTOs still have some advantages over landable/resuable staged boosters. Most people think the idea of propulsive landing is unlikely to be considered safe enough for a crewed vehicle, whereas an SSTO spaceplane has a lot more backup/abort options since it glides to a runway. That SSTO also has a theoretical faster turnaround time as well, since it also takes off horizontally from a runway. No need to restack your rocket, you don't need an expensive Starship-style launch pad and water deluge system, etc. It also helps the sonic boom problem that has been in the news for Starship lately, since the transition from supersonic to subsonic happens at higher altitudes.

But yes, there's less interest in aerospikes and thus less funding now that SpaceX has changed the market.

And whose fault is that, anyway? Ironic, isn't it?