I think this is just doing things differently, not fixing them - and IMO not for the better either.
I've been playing with making a chatbot with llama.cpp and FLTK[0] and FLTK's default behavior is actually to add a newline in the multiline editor when pressing Enter even if a 'Return button' is in the form (Return buttons are buttons activated when you press Enter or Return though Return is also handled kinda differently). And i have a big Submit 'Return button' there.
And TBH it annoyed me a LOT that i have to move the mouse and press the button to submit or that Enter adds newlines instead of submitting so that i explicitly added code so that pressing Enter is not handled by the editor (letting the Return button submit the input) and pressing Shift+Enter is what adds the newline (Ctrl+Enter also works, this comes from FLTK's behavior, but i've been used to Shift+Enter myself).
Which is basically how pretty much every chat interface (be it AI chatbot or something like Discord or whatever) that i've used in recent times works. And TBH it makes sense to me that the simplest/easiest shortcut (Enter) is what does the most common thing (send text) in a chat interface whereas the more involved shortcut (Shift+Enter or Ctrl+Enter) is used for the exceptional/less common case. In such an interface, the multiline editing is there as an exception (for when you want to paste some stuff and even then often Ctrl+V by itself can be enough), but most interactions are going to be single line submissions (often wordwrapped to look like multiple lines but still a single line).
FWIW sometimes when making games it helps to have everything in one place, even though later you may want to split things up for later projects :-P so sometimes you get big balls of mud with lots of checks everywhere for the actions you want.
As an example, most of the gameplay logic for Post Apocalyptic Petra[0] (a 3D platformer/adventure/riddle solving game i made for an MSDOS game jam a few years ago) is in the entity class for the player character[1] - including things like collisions, etc :-P. It is a bit hacky, but it works.
Though if i ever made that sequel i wanted to where i want to have an extra character following you around, i'd need to move a bunch of things to other more generic places so that they can be shared by both the player character and the AI.
Some time ago i was wondering if the common "me at foobar dot com" you still see a lot of people do actually helps at all, especially now with LLMs, so i searched for some common "obfuscation" techniques and found this site (not the 2026 update, but the previous - it was a few months ago). Then i wrote a simple LLM query with a bunch of examples from the site[0] (the tool is just a frontend for a commandline program that uses llama.cpp and Mistral Small 3.1 in Q4_K_M quantization since it loads relatively fast and is fine for simple prompts). AFAICT it could reveal anything that wasn't relying on CSS tricks or JavaScript.
Like others mentioned, though, personally i haven't bothered by email harvesting for years now since spam filters seem to do a decent job. I have my email posted in plaintext here (which i bet is harvested very often) and in various other places and the occasional spam i get is eclipsed from "spam" from services i've actually signed up for (coughlinkedincough).
IMO a better approach would be individualized addresses.
Imagine someone visiting your blog who wants to e-mail you can burn some CPU cycles to "earn" an address that hasn't been given out to anybody else, e.g. user+TOKEN@example.com, where it is algorithmically-unlikely for them to be able to guess a different TOKEN that will work. Then if abuse occurs, you can just retire that one address. (In a non-interactive context, like a paper ad, you could just generate one yourself.)
Naturally, this would be best with an e-mail client that is aware of the scheme, and with a mail-service that has some API for generating new addresses, such as if you want to cold e-mail somebody and use a new from/return address.
Some years ago I had the fanciful idea of doing it with a phone-app, where it manages creating new addresses as-needed, disabling them, and keeping notes about who you gave them to.
Sounds like a similar approach to this service: https://addy.io/
I use it all the time in conjunction with Bitwarden to generate unique emails per site. You can have notes in each email, and they show up in a small banner on in the forwarded email. And each one is individually disable-able, so you can easily cut it off if you see spam from it.
I was really interested in this space and made my own homegrown tool for this. I used it for a while until I discovered Addy and switched over. IIRC there are similar services by Mozilla, Apple, and Proton.
I would expect that a llm based scraper is going to be better at parsing an email address from your instructions than some of the more inattentive people who's emails you might want to receive. So I think some of the dumber mitigation measures that still block the simple regex bots from this topic are probably a better bet now.
The thing is, Microsoft got its position of dominance exactly because they did that - and that was because by doing this, the users' programs kept working. Remember that users outnumber developers by far and the last thing Microsoft wanted was for people to not upgrade Windows because they broke their previously working programs.
This was even more important at a time when Microsoft had actual competition in the OS space and people weren't able to just go online and download updates.
Nobody said it was by accident. They had deals with PC manufacturers for decades where they sold licenses based on the number of PCs sold by the manufacturer regardless of if the PC used MSDOS or another OS (like DRDOS), making all other options more expensive unless the client asked for them.
But the important thing is that the clients pretty much never asked for other OSes because all of their software worked on MSDOS - and later Windows. People bought computers to run their software, so if the software they wanted to run needed MSDOS or Windows, they'd buy the machines that ran that OS.
And by extension, if the software they wanted to run wouldn't run on the next version of MSDOS or Windows, they wouldn't have a reason to upgrade MSDOS or Windows. But from a user's perspective MSDOS/Windows was the best choice because everything supported it.
Microsoft didn't rely on just backwards compatibility (especially since the idea of "backwards compatibility" relies on something to be compatible with in the first place) but it was an incredibly important aspect of their strategy.
There was (and still is) VerInstallFile, however this was introduced in Windows 3.1 and it is possible installers wanted to also support Windows 3.0 (since there wasn't much of a time gap between the two many programs tried to support both) so they didn't use it.
TBH i think a more likely explanation is that they needed to somehow identify separate instances of that data structure and they thought to store some ID or something in it so that when they encountered it next they'd be able to do that without keeping copies of all the data in it and then comparing their data with the system's.
FWIW a two digit amount of MB is usually at least 16MB (though with low hundreds of MHz it was probably at least 32MB if not 64MB) and most such systems could easily do 1024x768 at 16bit, 24bit or 32bit color. At least my mid-90s PC could :-P (24bit color specifically, i had some slow Cirrus Logic adapter that stored the framebuffer in triplets of R,G,B, probably to save RAM but at the cost of performance).
Open source software also needs a stable ABI because:
a) i don't want to bother building it over and over (not everything is in my distro's repository, a ton of software has a stupid building process and not every new version is always better than the old versions)
b) a stable ABI implies a stable API and even if you have the source, it is a massive PITA to have to fix whatever stuff the program's dependencies broke to get it running, especially if you're not the developer who made it in the first place
c) as an extension to "b", a stable API also means more widely spread information/knowledge about it (people wont have to waste time learning how to do the same tasks in a slightly different way using a different API), thus much easier for people to contribute to software that use that API
There is also UMU Launcher[0] which is basically all that without the Steam integration/dependencies so you can run games from GOG and other stores (it is a command-line tool but launchers like Heroic can use it behind the scenes). I used to install dxvk, etc manually but in recent months i switched to it as it tends to work much more seamlessly for games (i did disable its autoupdates though).
I've been playing with making a chatbot with llama.cpp and FLTK[0] and FLTK's default behavior is actually to add a newline in the multiline editor when pressing Enter even if a 'Return button' is in the form (Return buttons are buttons activated when you press Enter or Return though Return is also handled kinda differently). And i have a big Submit 'Return button' there.
And TBH it annoyed me a LOT that i have to move the mouse and press the button to submit or that Enter adds newlines instead of submitting so that i explicitly added code so that pressing Enter is not handled by the editor (letting the Return button submit the input) and pressing Shift+Enter is what adds the newline (Ctrl+Enter also works, this comes from FLTK's behavior, but i've been used to Shift+Enter myself).
Which is basically how pretty much every chat interface (be it AI chatbot or something like Discord or whatever) that i've used in recent times works. And TBH it makes sense to me that the simplest/easiest shortcut (Enter) is what does the most common thing (send text) in a chat interface whereas the more involved shortcut (Shift+Enter or Ctrl+Enter) is used for the exceptional/less common case. In such an interface, the multiline editing is there as an exception (for when you want to paste some stuff and even then often Ctrl+V by itself can be enough), but most interactions are going to be single line submissions (often wordwrapped to look like multiple lines but still a single line).
[0] https://i.imgur.com/K3m9KAD.png
reply