We are talking about $400k for a movie. That's an insane amount of money, only very few movies can recoup that much money. The cost of AI video generators is likely to go up, once the companies start trying to move their business model from "burning investor money" to "not making to much of a loss".

In fact such systems have existed before fiberoptic cables... however there is a problem they have called "weather". While microwave links, which BTW can also reach 10 Gigs if you plan them right, will fade a bit in rain, laser-based systems typically fail completely and much faster.

For decades it never managed to get out of its niche, and there seems to be no progress on the horizon to change that.

I mean that was one of the big things with Chicago when they wanted to re-rejuvenate Windows with Windows 95.

Particularly if you have a system that supports multi-agent teams, you can just spin up a bunch of them, create a harness that makes them talk to each other in real-time and tell them what you want to do. It's not hard to do that.

For companies like Amazon that probably makes sense. They want to prolong the bubble of "Frontier model" companies. For the rest of the world, price hikes will eventually make them creative. My prediction is that we might see something a bit more clever than current coding agents that can deliver good, if not even great, results with comparatively tiny models. The smaller the model the easier it is to train, so maybe... future models will be so small you can train them for your project, on your own computers, for a fraction of the price of current solutions.
Nobody knows, but at our company, the next Anthropic price hike will mean very substantial costs for the company I'm working at. The kind of cost that gets you a substantial number of programmers.

Yes, but you can easily achieve that by just slowing it down. It's CMOS after all. Considering you get speed anyhow by having lots of parallel chips, that's perfectly doable.

It's not that hard to build such an SSD. Yes, cooling can become a problem, but there are established ways to deal with that. Essentially it just boils down to comparatively simple packaging. You just put more chips onto boards and cram more boards into your case. Putting chips onto boards is something that scales very well.

The hard thing is to do that cheaply.

I remember having exactly one use for my ZIP drive, to get some files from my PC to a PC with a CD writer. I did that exactly once.

Retrospectively it seemed like such a short time solution. It was rare, so you couldn't exchange discs, and after a few years there was the Internet, a fast and simple communications medium where you could just operate an FTP-Server at home and move files that way.

Exactly I have formated larger than 32 Gigabyte devices with FAT from systems like Linux before. It just works. No idea where the actual limit is.

I mean who downloads software from some random website?

This either is completely fabricated or something _way_ more mundane blown up a chain of bad reporting.

There is no plausible way you can detect any physical signal from reasonably long distances... and separate it from noise that looks identical.
Also it uses the word "quantum" which impresses some people, even though it means very little. My computer uses quantum effects, sure... but so does a rotting fish.

Linux does typically require more RAM than Windows. I mean Linux is a full operating system including network stack and multitasking, while Windows essentially is a shell for DOS... with some added benefits like having a GUI-Toolkit that supports multiple applications running at once due to cooperative multitasking.

So while your typical Windows system was sold with 4 Megs of RAM, a Linux system typically isn't usable with less than 8 Megs of RAM.

Of course this is now largely irrelevant as RAM is fairly cheap. You can literally have a machine with multiple Gigabytes of RAM, way more than himem.sys would support. (the limit is at 1 Gig, BTW)

It seems to me that many people confuse "Apple Computer" with "Apple" since they have a similar logo, one with a rainbow apple, one with a plain apple, and worked in tangential areas, one sold computers, the other one sold computerized consumer products.

I mean "Apple Computer" had an excellent reputation as treating their customers as equal, providing them with all necessary documentation for fixing and modifying their computers in any way. In contrast "Apple" today doesn't even let you run non-approved software without going through their hoops.

You might as well say that we are at 58 years of Apple, since Apple Records was founded in 1968. They too have an apple as their logo.

Both Intel and AMD need each other. They both sell the CPUs for virtually the only open computing architecture there is the "IBM-PC"-derived "PC". That architecture is what's running all serious workloads as it is currently the only standardized architecture. For ARM you need to port every operating system to every SoC individually. on "PC" you just pop in your USB-stick and can run the same OS image, regardless who made it. There are only minor incompatibilities which can be resolved by checking the PCI bus for what hardware you have there.

This is an important market. If the number of companies making those CPUs went down to 1, there would surely be governmental interventions. Either in making sure this company works, or by moving away from them. That's why Intel needs AMD and AMD needs Intel. If one of them falls, the other one will get big problems.
Yes, such interventions are not that likely, but just think about Microsoft, they sold BASIC interpreters and PC operating systems for a while and had a fairly big market share. Even though that was nowhere near an important monopoly, there were serious considerations of splitting up the company.

Giving a project top much money is a fairly easy way to kill it. Look at Mozilla, for example.

For years it swam in insane amounts of money. Since they cannot simply put that much money onto a bank-account, they reasonably did all kinds of non-browser related things with it. Since software development kinda doesn't need much classical management, the managers focused on those areas. A Mozilla manager didn't need to care for the browser, as that is managed by technical people.
However once that source of money has dried up, you still have the management, which will, obviously, first cut funding for the browser part... eliminating the path to more sustainable ways of funding.

If instead Firefox was a true "Free Software" project, this would have gone differently. Whatever money they needed would have come in by lots of small donations. They would have had limited developer resources which may have helped keeping web standards less complex. Maybe we would even have had a better web where things like Passkeys wouldn't rely on Javascript.

It combines 2 things, very good portraits of the people designing computers, and very good "bird's eye views" of how computers work. Particularly the later is great as the author managed to distill the core of each concept into something even a lay-person can swallow and understand. It's didactical simplification to the point and not beyond it. So what he says is stays correct, only ignoring things that are not important.