dspillett

No worries, I'm happy to have very slightly reduced the amount of confusion in the world (especially if it might have been caused by my previous message not being as clear as I thought it was at the time of writing!).

The implication is that while they are waiting for the async loop to complete for one site, the bot can be scraping others that have already had the challenge completed (usually the one challenge stays in force for some time afterwards) or that have no such challenge.

Memory hard challenges will reduce the effectiveness of this, but not remove it completely.

Though that these methods are working at all implies that most (almost all?) bots are not bothering to use such async methods as there are sufficient resources not protected by them that this complication is not yet worth implementing.

Equating bot prevention measures with “please give us permission for 1,000s of companies to stalk you around the Internet, no? Here are some dark patterns to try make you accidentally agree” is rather unfair IMO. They are an inconvenience but for the protection of the sites rather than for the betterment-of-adtech-via-conning-you-the-viewer-with-malicious-compliance.

> Bots can use async or virtual threads and not even have to hold on to a stack while they are waiting for this bullshit

They could, but given these methods seem to work currently most (almost all?) are not. They will start doing so if the amount of resources protected in this way increases to the point where it is worth it, or course, that is why memory-hard processes are being worked on which will make that route less practical.

> cookie banners, and of the that crap should be an automatic F grade on WCAG and should make web sites immediately liable for ADA lawsuits

I can't disagree there, though that would be quite an inefficient process and requires a bunch of money up front from who-ever champions the legal challenge, so I doubt it would actually be at all practical.

I think many don't see the point of robots.txt as it will just be ignored by most bots anyway, so instead concentrate on more complicated measures.

> Today, AI can generate lightweight, standalone, accessible HTML with just the CSS it needs, so a Minimal CSS Framework for Semantic HTML matters less than it used to.

If your answer is "use AI", then why not use AI to not need Pico. This could be maintained for those of us avoiding AI, of course, but we are fewer and further between with each passing day, and likely rolled our own instead of using someone else's framework anyway, so he probably didn't see the point and wanted to spend the time elsewhere.

> brand recognition that shouldn't be throw away

How useful do you think this brand recognition would be in his life? Enough for it to be worth maintaining a project he had presumably moved away from using for his own work & play?

If it is so valuable then maybe someone will make a deal to take over maintainance. Though TBH if this were mine I'd probably prefer to keep the project mine-but-archived than let someone else do what they will with the "brand" thereby diluting what value it might have to me personally. It can just be forked instead, or people can move into one of the already existing forks.

I think "pretty average" is the thing to strive for in a README. The right balance of detailed but concise, not overly festooned with unnecessary verbiage like a stereotypical recipe page, a simple "what is it", "key features", "how to get/install", "more detail on features / example use" structure.

Of course it gets harder to find that balance as the size and scope of a project expands, you have to become judicious with cutting things out and excising them to extended documentation instead of everything being in the README. For a thing this size everything is close to the right balance if you word+arrange it right, for larger projects the README should be a summary of what, how to get, and where the other documentation is.

If this one is AI written (there are signs, but those are things humans have long done which is why AIs copy them) it has been human edited rather than just thrown up as-is.

> as someone who struggles to write readmes

If you are struggling a lot with that, perhaps you are overthinking it or trying to shoehorn in too much detail?

Nah, it's SciFi. They'll find a way of writing them back in. Won't they?

Though I've never seen a v6 address represented that way, it would make then noticeably longer, so I don't know if anything will accept such a string as a valid address. Though IPv4-mapped addresses are sometimes show as a mix (::ffff:10.5.2.1 for 10.5.2.1) so maybe I'm wrong there.

Likewise I've never seen IPv4 addresses written in hex in the wild, but the relevant specs say it is valid. `ping 0xc0a82bc2` works just as well as `ping 192.168.43.194` but I doubt many forms accepting IPv4 addresses will validate that and allow it.

Hence my having a /64 (/48 in fact, should I ever enable more than the current /64) shows that 64-bits is enough for global addressing. If it wasn't then giving me a /64 would be part of a path to creating an address exhaustion problem just like the one with IPv4 (which was exacerbated somewhat by handing out /8 ranges unnecessarily in the early days).

If it is only free in order to inconvenience competitors in a way that will eventually reduce choice, either giving you undue control of a market or being done with a view to charge later when people have no choice as the competition is broken, or both - that is when free becomes an anti-consumer matter.

It is an issue that you sometimes need to think long term to notice. If you wait for it to be an immediate problem, it is often too late to easily do anything about it.

Your food example is exactly what people feared would happen with IE, and what actually did happen for a long time despite the anti-trust case: client side developments stalled because IE was such a large part of the market and it did not implement anything new because there was no reason for them to. There was not sufficient competition to make it inconvenient for the incumbent to stagnate, so the whole area stagnated somewhat until Chrome came along and stirred things up a bit. MS was in control due to an effective monopoly, the consumer suffered.

IIRC something like that was considered when IPv6 was first being designed, be deemed insufficient for the long-term future. 64-bits is thought to be enough to stave off address starvation though, which is why IPv6 address ranges are usually assigned as /64s (my homer link has a /48 range should I wish to use it all!), so I can see your reasoning on that level.

Thanks. I might need it!

> sooner or later they will start tracking if we fired up Claude for the day

We track who uses what AI and how much. For billing/expense tracking purposes, as we've gone from “use it as much as you can” to “use it as much as you need to but don't waste it”, but there will be a table somewhere with my name in the bottom section next to a bit fat 0.

> has been pretty pushy when it comes to GenAI use tho … Flat out saying no comes with …

You'll have to deal with that or get out, eventually, how soon depends on your management. AI is not going away, even when/if the costs rise when all that investment capital starts being clawed back. If it is any help in choosing an exit strategy: under UK employment law (if you are elsewhere you'll have to verify how things work in your jurisdiction) they can't force you to use AI or punish/sack you for not as it can be argued to be a material change of role (as it is described as revolutionary, or in less breathy terms “like becoming a manager”, etc.) not just a different tool. But if you refuse then the same argument can be used to make you redundant (“we need more AI people and fewer traditional devs, traditional devs are being declared redundant, would you like to switch roles instead?” would be perfectly legal). As much as the payout for my long years in the same place would be nice, personally I don't think I can sit it out long enough for that to happen.

That is a function of the training data/method, like many of AI's problems. This doesn't stop it being a problem of course, but it shouldn't be a surprise.

Big bugs, like heartbleed, over the years have had breathy websites going over how serious the problem is because the problem was that serious, so the models associate that sort of wording with bug reports. Less serious bugs don't get the same excited treatment, repeated in many many places, so more measured descriptions figure in the training data relatively less.

I'm not qualified for anything else that isn't also remote, full of AI and getting more so, or both, so I may quite literally end up waiting tables, I have friends who can help me skip the queue there with an “he's a good bloke” recommendation. While I was in hospital recently (gallbladder trouble and liver infection) talking to one of the porters who was wheeling to & from the CT scanner, they apparently need more bodies ATM, I could certainly do that (other than the temporary issue I was in there for I'm fit & healthy) and might actually find some enjoyment in it. Both involve dealing with real people who are in the room with me. Anything I do once I leave here will be result in a massive pay cut, but if the change saves my sanity then that will have to be acceptable.