This may not be popular, but I think somebody needs to make the argument in favor of scraping the web unhindered.

1. Copyright law controls the distribution of copyrighted works. AI models are not distributing the works that they’ve accessed, any more than a seller of applesauce is distributing apples. You won’t find copies of those materials anywhere in the database, and whatever the AI coughs up is almost always mangled and mashed up beyond recognition. (And when something comes out that’s recognizable, that’s most often a concern of trademark law rather than copyright. Meaning, if your AI generates pictures of Batman and you use them, then DC may rightfully want a word with you. They won’t care how the model was trained; they’ll only care what you’ve done with it.)

2. Copyright law is supposed to economically incentivize the production of creative works. It’s hard to come up with how AI scraping undermines that principle. It’s not a direct replacement for any of the works that it was trained on. It’s not breaking the various business models that got the training material posted in the first place. One could argue that generative AI is a supercharger for creative work and lets more of it be produced faster. That’s what we’ve always wanted, isn’t it?

3. The way generative AI works is analogous to the way that human beings learn and create. We observe and consume creative works, mentally digest them, and then we mix-and-match our favorite parts to create something new. AI is a more automated version of the same general processes. So, why should it be acceptable for every human being to do this “by hand” (as it were) but not employing automation to the same effect?

Going beyond copyright, I understand there are practical concerns about massive scraping operations and the sheer amount of traffic and hits that they can produce, and real costs that they can incur. So, I'm not addressing that. Those are issues that'll have to be negotiated and worked out between various parties.

The bit about Model Y becoming the number one seller is interesting, but it may be an unfair comparison, symbolic win, etc. . . I'm surprised at how little reaction I'm seeing to the big news about the Ford and Tesla charging deal. That, to me, was the shocker. I had to double check that it wasn't somehow April 1 again.

I hoped that NACS might get some traction, but I never imagined it happening this way. I thought the other charging networks (Blink, ChargePoint, etc.) would jump on it first, and maybe some of the startup EV makers like Rivian or Lucid would go for it. I thought established car companies like Ford would fight NACS to the bitter end.

The C64 was dethroned by the Raspberry Pi, if I recall right.

The companies that I listed are a rounding error of profit on photography? What do you think Pentax is today? It's not like sticking with DSLR is going to drag Pentax down from a huge corporation to a small niche. They're already a small niche. The only question is whether they're going to be a viable niche or whether parent company Ricoh will, at some point, shrug and shut them down entirely.

Also. . . Fujifilm are making more money off Instax film today than they are from their digital cameras. Does that matter? I dunno. Fujifilm is a diversified company, as you noted, and I'm not sure that cameras or film are big contributors to the bottom line. In Japanese business, cameras still hold a kind of cultural prestige, so companies like Fujifilm and Ricoh will keep those divisions running as long as they don't become an outright financial drag. (Which is what finally happened over at Olympus, not too long ago.)

I think you got Eastman Kodak and Kodak Alaris mixed up. Kodak Alaris is the spin-off company that is still manufacturing photographic film and, seemingly, doing okay with it. Yes, it is indeed a tiny business compared with the pre-digital era. But again, we're comparing with Pentax here, another brand whose heyday is already decades behind them.

From TFA: "The move leaves Pentax and Ricoh as the biggest remaining names in the rapidly diminishing DSLR space."

Odd sentence. They make it sound like Pentax and Ricoh are two different companies both producing DSLRs. In fact, Pentax is a brand owned by Ricoh. Although Ricoh sell some cameras under their own name (such as the Theta and the very successful GR line), all the DSLR products are branded as Pentax. So, really, Pentax is the only player left in this space.

Reasons why I think Pentax may be viable for the long run: Their DSLRs have excellent pentaprism viewfinders that are particularly nice to use outdoors in bright sunlight, where EVFs can struggle. Pentax cameras are known for their weather sealing and physical ruggedness. They are good outdoorsy gear.

There is also an independent (not supported by Ricoh AFAIK) forum website, including an almost wiki-like product database, which is one of the best-designed anywhere. It's a highly useful resource for Pentax owners, and it puzzles me that other camera brands (even far bigger and more popular ones) don't seem to have anything to match it. This will be a bigger factor with DPReview shutting down. (Remember those DPReview jokes about "dying Pentax"?)

For those who think sticking with DSLR is the road to ruin. . . Kodak Alaris, Fujifilm and the new Polaroid company are all making film and making money today. Ricoh-Pentax may have learned lessons from watching what happened to (old) Polaroid, who abandoned instant film just when it was about to rebound and paid a terrible price for that decision.

I lurv my Thelio, and I also got a Lemur Pro that is generally working well. I'm pretty loyal to System76 at this point. However. . .

They do seem to take ages to accomplish anything with hardware R&D. I mean, I'm sure there are multiple reasons for that; I know that System76 isn't Apple. The gap between everything they want to do with hardware and what they have the actual resources to accomplish is understandable-but-frustrating.

When it comes to software though, I have no idea what they're even trying to do. I've given Pop!_OS a spin a couple of times (when I got each of these machines), and both times it quickly drove me up the wall with its missing features, ugly aesthetics and bonkers UI. I've got them both running Mint Cinnamon now.

Well, what I can vaguely remember is that those 8-bit processors had few registers to work with, but they were pretty good at stack operations, which is what Forth is built around. And since you, the programmer, were manipulating the stack deliberately and explicitly, you tended to think a lot about the best way to do it.

By default, C also allocates variables on the stack, though all the stack manipulation is concealed from the coder. Then as processors with more registers arrived (68000, etc.) the "register" keyword provided some opportunities to speed up some C code by specifying register variables.

The register keyword still exists, but I think it's rarely used now, since compilers have become smart enough to make those decisions for us.

Back in the Glory Days, Forth was reputed to be faster than C. I don't know how that would play out with today's hardware and compilers, though. I guess it's moot since Forth is now a footnote.

Charging statioins in rural areas are not for rural people who live in those areas. They're for the travelers cruising down the highway, so they can stop in that little town at the 7-11 (or similar) and spend 20 minutes charging, grabbing snacks and coffee, and relieve themselves, and then get back on the road and continue their journey.

Rural folk who own EVs charge them at home. We only need to use public chargers when we are away from home, cruising down the highway.

There have been rumors of a CD revival making the rounds, but is it really happening? From my observations, CD interest and activity seems to be up, but it's almost all oriented around used CDs. It's thrift shops, it's the used CD bin in the record store, it's eBay and Discogs. They've reported a steady upward trend in CDs sales on Discogs for a few years now. None of that shows up in the record industry's sales statistics.

From my own observations, the Loudness Wars are still raging. I mean, perhaps not as bad as the worst times of the early 2000s, but still. . . Everything is still mastered "hot" instead of mastered for dynamic range, and the difference between today's releases and those from the pre-war era remains very obvious.

If record companies ever start mastering CDs for sound quality again, I'll pop the champagne cork and start buying new discs. But, sadly, I just don't see any trend in that direction.

I don't think you can blame the lack of Steve Jobs for Apple messing up their gaming platform. For as long as I can remember (and I go back A Ways), Apple have either been apathetic toward gaming or, a few times, completely inept when they did try to get into it. This seems to be deeply embedded in Apple's culture, and I can only assume it must have come down from Steve Jobs.

NACS is not the same as the original Tesla Supercharger standard. Tesla Supercharger worked much more like CHAdeMO, and it used the automotive CAN bus for communication. My understanding is that NACS is much more like CCS routed through the Tesla physical connector. (We might almost think of it as CCS Type 3.) Older Teslas don't support CCS Type 1 adapters, and they don't technically support NACS. It won't matter when using one of Tesla's own pedestals, of course. I'm guessing that third-party NACS chargers, assuming they ever appear, are less likely to work with older Teslas that don't have the CCS protocol support built in. (But older cars can be retrofitted, supposedly.)

Saying Tesla "lost the battle" is sort of like saying Microsoft "lost the battle" for desktop OS. I mean, look at how many more companies are backing Linux!

Personally, I don't expect either CCS Type 1 or Tesla/NACS connectors to go away, as far into the future as I can usefully look. They'll have to co-exist, and we all better get used to that. Tesla seem to be paving the way for inter-operability with simple adapters and the "Magic Dock" device, as well as publishing the NACS specs for other charging networks to use.

I think that's also a good description of how quality was in the USA back in the "good old days" when the USA was the world's factory. Quality level was related to price. We did cheap, shoddy stuff, and we did premium-priced, premium quality stuff, and we did levels in between. It's fascinating now to look back on how this changed. As wages went up, and inflation and competition from overseas markets put pressure on US manufacturers, at first things got more and more shoddy. Then all the companies making the shoddy stuff packed up bags and moved overseas. The relatively much smaller portion of companies that made premium stuff either went broke (and their brands became zombies) or stayed here and kept producing premium stuff at even higher prices, which some of them still do today. But it took a long time for the "Made in USA" label to rebuild its reputation.

I wouldn't be surprised to see China go through a similar cycle, although there are also some complicating factors that make their situation different too. On the one hand, we have political (not only economic) factors pushing some production away from China. On the other hand, the electronics supply chain has become so deeply rooted in China now that it's difficult to get away from.