It would be interesting to do a study of how well programmers comment their code and see how that is correlated with touch typing, or with typing speed in general. My assumption is that it's strongly correlated.

Though I have to say AI is changing things. Now you can write a good comment, and then hit tab to accept the suggested code, sometimes needing to make a few edits. It's crazy.

And it gets more confusing as some people put 65C02 processors in or a 65816 (which includes 65C02 instructions). Even in the Atari community, people will get it backwards.

The summary mentions added instructions, so I assume those are 65C02 instructions, most notably the STZ instruction (store zero), allowing you to write zero to memory without altering one of the real registers. The downside is that there were undocumented instructions that people had figured out on the original 6502, and some people used them. Apparently it was done in some cases to obfuscate code to make removing copy protection more difficult. A few of the instructions were actually useful in some cases. So adding in new instructions from the 65C02 or 65816 reduces compatibility. The only exception would be the original opcodes that locked up the CPU.

As a side note, the undocumented instructions were caused by not having entries in the instruction decoding table for them, so they got a mix of entries for other instructions, enabling various parts of the CPU in ways that behaved strangely.

Much of the development of new code on retro computers is done using tools on modern computers. Last year I did some coding for my Atari using Linux.

I wrote a simple AUTORUN.SYS program to launch a BASIC program, but with a few tweaks like enabling BASIC if it was disabled and displaying an error if it's a model without BASIC. And I managed to get it in 124 bytes so that it would fit on a single 128-byte sector, so it can be a drop-in replacement for any other version without those features. That was fun, and I wrote and assembled it under Linux.

I also wrote a FUSE file system to mount Atari disk images as native file systems under Linux (and apparently MacOS and with some hacking Windows). That was fun, and more about analyzing the various Atari DOS formats. No Atari hardware was used in that project.

I was going to push back and say that those are things that Apple's own products also don't make use of (other than the iPhone itself), so excluding those from third party use isn't giving Apple an advantage in selling additional products, which is what I think the whole complaint is about. If Apple is using restricted APIs to sell additional stuff that works with iPhone, but competitors can't, I can see why that would be an issue the courts would be concerned about.

So the "I was going to" was before I realized that Apple does have just such a product: The Apple Watch. I haven't used one, but I expect it wants to work as an extension of the phone, so it would need to do many things that normally are only allowed to be done by the phone itself, including access to all notifications and sharing WiFi login data so it can also connect.

So, yeah, that's an issue.

The key word in the summary is "bankruptcy." Assuming the ISP is in bankruptcy, this settlement is an easy thing for them. They record companies are just another creditor, so all that really matters is that the bankruptcy court agrees that the settlement is fair. The real losers are the other creditors.

If it's such a security risk to open those internal APIs, just eliminate them entirely. Make all the Apple applications work without them as well, and then everybody is on a level playing field without terrible security risks.

If Apple really needs extra access to make add-on products work nicely, then the EU really does have a point. (And I assume that's the case.)

While generally this is a great idea, on unintended consequence is that laptop makers will limit the number of USB-C ports if adding support for all the features is a challenge.

I call those Wookiee Brands, as the brand names appear to be random Wookiee names that aren't necessarily pronounceable by humans. If Amazon's search did a good job of showing me the cheapest options first, then I wouldn't mind them, as seeing all the Wookiee Brands for a product make it clear what it is, and I can check the reviews on one that has been around longer, knowing they're all the same. But automatically combining them into a single listing would be much better, along with a note "Product may contain different branding."

What I hate is how all the streaming services fill up their interface with shows that you have to pay extra for. No, I don't want to add another streaming service, so I don't want to see Hulu content in Disney+. Prime is probably the worst, as they're trying to get you to subscribe to numerous other services, as well as rent or buy anything they don't include.

I think most of my SSH keys are 4Kb RSA. If they need a million qubits for 2Kb, do they have to square that for 4K?

Is there a good post-quantum recommendation for ssh keys?

The point was due to reports of new AI datacenters that are using generators because they're not able to otherwise get enough power. So if they need to use generators on a regular basis, that's not an emergency. So, yeah, probably only when the power goes out.

Maybe we should consider requiring new renewable power generation before allowing new datacenters to connect to the grid (and ban generators for non-emergency use). There would be a market for renewable power installation credits, so new datacenters would drive the greening of the grid instead of increasing carbon output.

This doesn't solve the water use for cooling issue, though.

I'm not ready to turn over major code development to AI, but I would be perfectly happy to have it scan the code base for suspicious code. Just the other day I had a small bit of code in a language I'm just learning, and I pasted it into an AI and asked it to explain it. The explanation was good, and it pointed out that what the code actually did did not match what the comment said; that's a big red flag. It was also able to fix the code to match the comment. Since I'm just learning the language, it saved me a good bit of time (but probably slowed down my learning of the language).

But it would be great to have it look over the whole code base and report issues that may need attention.

This is about cheaper batteries, not about range. We already have cars with over 400 miles of range. They're just expensive. All that matters here is the lower cost per kWh. If they're saying a 30% cheaper battery, then that means automakers have options to reduce prices or extend range (or increase profits for a few years until other companies get similar technology). And is 30% the real number? I'm not clear on that, and even if that's the claim, it's hard to be sure they're not cherry-picking price points on existing batteries to make their tech look good.

So perhaps this means your standard 300+ mile range car can move to 400+, but more likely it means a price drop, though perhaps with pricier longer range versions. Except for trucks where towing kills the range.

Speaking of trucks, the other thing to look at is weight per kWh. How much weight does this drop from a 500-mile semi? That's significant, as it means more cargo.

That's also the class that has legal requirements for a minimum passing distance. In many states, you have to have at least a 3' separation. Massachusetts didn't have a specified distance until recently, but set it at 4'. That's for pedestrians, bicycles, scooters, electric unicycles, horses, and pretty much anything else without a license plate. It's a legal category, so it makes sense to use the term when discussing regulations, safety criteria, and such.