With ten finger typing, having the most-used keys on the home row is a significant advantage for speed and ergonomics. With swiping, having a sequence of characters close to each other makes it hard for the algorithm to predict the intended word. With tapping, it’s a disadvantage to have adjacent characters in a sequence on a small touchscreen because it increases the chance of fat-fingering them.

Maybe. I use Dvorak for real keyboards and QWERTY on my phone. I tried Dvorak on my phone and didn’t see any benefit.

Heliboard offers the option. The ideal layout for a small onscreen keyboard may be rather different from one for typing with all your fingers though.

We need root on our car computers.

The AI thing I’d really like is an on-device classifier that decides with reasonably high reliability whether I would want my phone to interrupt me with a given notification or not. I already don’t allow useless notifications, but a message from a friend might be a question about something urgent, or a cat picture.

What I don’t want is:

• Ways to make fake photographs
• Summaries of messages I could just skim the old fashioned way
• Easier access to LLM chatbots

It seems like those are the main AI features bundled on phones now, and I have no use for any of them.

it’s as simple as editing a single config file

It isn’t. Mastodon has a character limit hardcoded in two places. Critically, that’s not a limit on what it can receive and display, just on what local users can post. With Bluesky, it’s part of a schema that would be enforced on posts from elsewhere, if anybody was actually running a Bluesky-compatible appview in the wild.

There is a chance that I just don’t get microblogging. I’ve always felt that short character limits encourage people to make bad points that resonate emotionally but fall apart when thought through, and to yell at people they disagree with rather than being thoughtful.

I think the idea that forced brevity is an important component of microblogging is mistaken. Low friction to post, minimal formatting, and (optionally) collapsed long posts in feeds all encourage short posts without requiring them.

It might have served more of a purpose when Twitter launched because people weren’t in the habit of short text posts at the time, and because Twitter supported posting via SMS.

I don’t see value in a character limit other than whatever might be needed for technical reasons. Bluesky allows alt text for images to be 2000 characters, so clearly any technical limitations allow at least that much.

For those who prefer short text posts, hiding posts longer than a user-configurable setting behind a “see more” link would do.

Yet they still think it’s a good idea to limit text posts to 300 characters for reasons I cannot fathom.

And people wonder why I have no interest in getting a newer phone with an AI thing on it.

That’s good in theory, but a site behind Cloudflare won’t necessarily notice that a legitimate user got blocked. If you want them to care, you’ll have to find a way to contact them. For more impact, tell them which competitor you spent money with instead.

Total Webhosting Solutions

I’ve been with Porkbun since Gandi got acquired. No complaints.

That phone isn’t actually small. It has a small screen, but it’s 25mm thick and weighs 300g.

Great, but a web browser still does not need terms of service. There’s no ongoing relationship between the user and the creator of the browser, at least, there shouldn’t be unless the user signs up for additional optional services.

It’s great if Mozilla wants to offer some optional services users can opt in to, and those services probably need terms. I use Firefox Sync, though I’ve started to reconsider that given the recent fuss. The browser itself? I’ll move to a fork first, and stop recommending Firefox to others.

It already does, though not in the individualized manner he’s describing.

I don’t think that’s entirely a bad thing. Its current form, where priority one is keeping advertisers happy is a bad thing, but I’m going to guess everyone reading this has a machine learning algorithm of some sort keeping most of the spam out of their email.

BlueSky’s labelers are a step toward the individualized approach. I like them; one of the first things I did there is filter out what one labeler flags as AI-generated images.

Speaking with people is not a search. Looking in their bags is a search, however inviting them to show off their contraband is not.

Depending on how the requirement to accept the ToS is implemented, a config file might be able to disable it and any features that depend on it.

If the officers didn’t identity themselves it might be different

Why? Bragging about committing crimes to random strangers is a very stupid thing to do.