Liam_on_Linux (![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
liam_on_linux) wrote2025-05-20 07:17 pm
liam_on_linux) wrote2025-05-20 07:17 pm
I am so sick and tired of "AI"
Spent much of today getting a Live AROS USB key working, which wasn't trivial... it needed a USB 3 key, which I had to go and buy specially.
But after that... I am so tired. I want to write about new stuff in software, but there feels to be no area not contaminated with "AI".
I get the depressing feeling that computing is just being eaten up by bloody "AI". Virtually every press release I've seen this week has been AI. Mozilla adopts new AI search engine. Red Hat releases RHEL 10 with built in AI chat bot to help clueless PFYs admin the thing. Windows bloody Notepad has AI built in. AI in Google Docs. AI boosters in my mentions telling me and my friends that AI is helping them read antique books or whatever.
I get the depressing feeling that computing is just being eaten up by bloody "AI". Virtually every press release I've seen this week has been AI. Mozilla adopts new AI search engine. Red Hat releases RHEL 10 with built in AI chat bot to help clueless PFYs admin the thing. Windows bloody Notepad has AI built in. AI in Google Docs. AI boosters in my mentions telling me and my friends that AI is helping them read antique books or whatever.
Is there anywhere outside of retrocomputing that doesn't have AI in it?
"AI" is not "AI". The liars and the shills redefined what people used to mean by "AI" as "AGI", artificial _general_ intelligence, so they could market their stupid plagiarism bots as AI.
AGI is not real. It doesn't exist. It will probably never exist. Hell, at the rate humanity is going, we won't be able to build new computers any more by 2050 and the survivors at the poles will be nostalgic for electricity.
AI is a scam. It's a hoax. It's fake news. There is no AI, and what is being sold as AI is such an incredibly poor fake that it is profoundly disheartening that so many people are so stupid to be deceived into thinking it is AI.
The blockchain is a scam. Everything to do with it is a scam.
Alternative medicine is a scam. All of it. There is no such thing. If it's called "alternative" that means it's been proved not to work.
All religions are scams. No exceptions.
People have made billions from selling scams for my whole lifetime.
Meanwhile, other scams, like plastics being recyclable -- they aren't, it's a lie -- or biofuels -- also a lie -- mean our civilisation is on the verge of collapse. We are killing the planetary ecosphere that keeps us alive with plastic and pollution from burning stuff. We have to stop burning everything, stop cutting down trees, and stop making all forms of single-use products. No more jet planes. No more private cars. No more foreign holidays. We can't afford it.
And with all that we are almost certainly still doomed.
But I kind of want to see my industry go first, if all it's got now is AI.
Snag is, I need a job. I have Ada to pay for.
no subject
Preach it, brother!
(Doing my bit corner: I just had to check the copy-edits on a novel this past week. Tor Production insists authors use Microsoft Word. But Word now includes an LLM. And my contract with Tor warrants that the novel is entirely my own work; use of an LLM would arguably put me in breach of contract. So that's my excuse for going against their express orders and using LibreOffice … which is finally mature enough to digest a change-tracked book, with comments, and not screw everything up.
And I was up-front about it and they're treating me as a semi-official guinea pig, so it might gain traction. As my editor is an editorial director there, and also uses LibreOffice because -- hello -- he doesn't like LLMs either, it's probably going to be all the rage by next year. And meanwhile, I'm getting close to cancelling my Office365 subscription next month because the only use I have for it is checking copy edits on one book a year reliably.
no subject
no subject
LibreOffice is still improving in quite interesting new directions.
I really think that just maybe it may be about to see a renaissance.
It's smaller, cleaner and faster than it's been in about 30 years.
One team has it running inside a web server, presenting a Google Docs-style UI inside a browser. So, trendy in-browser productivity, but all-FOSS. I don't care and don't really need that, but some do.
Another team has it running locally by by compiling it to WASM and embedding the result in a web page. I think WASM is a terrible idea in general and a one gigabyte applet in a web page is, er, suboptimal, but modern computers could handle that, and it can do impressive things.
Now it works they're trying to modularise it.
Those two companies just merged, incidentally.
Another team is working on implementing live collaborative editing, again Google Docs style.
Put the three together and a whole bunch more possibilities open up.
Now your browser can just handle any MS Office format, in browser. Word, Excel, PowerPoint, whatever, you can display it, or if you want edit it, or you can output to it. Run locally or remotely, as you prefer. Standalone or collaborative.
All this has been doable with proprietary stuff for years, and now it works on Macs and Chromebooks and Chrome and Firefox and Safari, and on your phone...
There have to be ways to use that, right?
As for Word -- I mean, as you know, professor, I'm a heretic and like Word -- but could you not just run an old version without the AI bollocks?
I still occasionally use Word 97. It's hilarious to me that just under 30Y ago it was a bloated mess, and now the same binary probably runs inside the L2 cache of one of my CPU cores.
Word 2003 works quite nicely under WINE, too...
Doubt all official data
You have waken up to the idea that not everything is as portrayed in the corporate media. Big corporations and the governments are joined at the hip, practically the same people are running them, moving back and forth between them. Make the next step, reject everything they say, on principle.
And then comes the interesting part: Try to find out the truth for yourself. Maybe the official story is true, maybe the "alternative" is true, or maybe it's a 20/80 split type of a situation.
In many cases, there is not enough or too contradictory data. So accept that "we don't know". I'd say, that is the usual situation. Good data is incredibly hard to collect. If there is an interest to bend it in one direction or another, it is bent. So important data is definitely bent by the time it reaches the public. How much? Who TF knows...
Re: Doubt all official data
I don't know who you are or what you are doing in my comments.
No, it is not a scam.
Go read this. It's long but you'll learn something.
https://medium.com/@samyoureyes/the-busy-workers-handbook-to-the-apocalypse-7790666afde7
Do not comment again until you've read the entire thing or I will ban you on the spot.
There's no turning back the tide
Re: There's no turning back the tide
And it might even be correct, right?!
Re: There's no turning back the tide
Actually, I think MS-DOS 5 had a quite good HELP command. It had to because MS 5 was a response to DR DOS 5 and it did.
But I have a feeling the HELP.EXE command appeared in DOS 4.
Rewind a few years to MS-DOS 3.3 and you'd be right, though.
Saying that, it took me literally years to discover that the Unix for Help was "man".
It helped. Not much, because man pages were mostly awful until GNU info tried to replace it, then people woke up and improved them, but it was a bit of a help.
no subject
I do my best to filter AI topics from hacker news, and it regularly removes about 1/3 of the articles in my feed, and still there's always a new term, technology/technique, TLA, model, that I have to add to my filter. I think ublock supports regex, but I'm not very good at that, so I just manually enter variations of capital letters that make sense, like openai, OpenAI, Openai, Open ai, etc... otherwise one gets through just because of that small change. It's like playing Wack-a-Mole. Now AI search everywhere, even if you block it they try to force it back.
I just saw Adobe is adding "AI" to their crap AND increases prices across the board even if it's not a feature you want/need, and of course they have vendor lockin. Seriously the FTC in the US should go after Adobe if anyone is a bloody monopoly on a service like that it's them.
And the damn energy wasted. And Zuckerberg pushing for what amounts to ELIZA2000. The "loneliness epidemic" sure as hell won't be solved by talking to a chat bot as a "Friend" or a "Therapist", and just how much mental illness do the people leading us have? I mean this in a serious way, like they are mentally unhealthy.
no subject
Good idea!
Probably won't fly for me as I need the research materials, but it sounds good...
no subject
Amongst the many reasons for loathing LLM / graphics plagiarism models are that their makers are doing a DDOS against the web.
I thought I was being quite good at blocking bots - OpenAI's was obsessed with one page on a server, visiting it every few seconds, so that was blocked a while ago - but adding a few more "AI" bot-specific rules to fail2ban a week ago has resulted in traffic to a forum site halving.
I changed default search engine from Google some months ago, and I have no regrets at all about doing so.
no subject
It is such a massive PITA, isn't it? My sympathies.
XKCD 810 is starting to look like wishful thinking. FFS.
I still use a de-crappified Google, and wrote about how recently. But there are gaps...
The AI you hate IS avoidable
"There is no AI, and what is being sold as AI is such an incredibly poor fake..."
Hi Liam,
you are right. But that AI is still avoidable, if we all restart from these thoughts I recently shared, on this very topic.
Re: The AI you hate IS avoidable
https://www.project-syndicate.org/commentary/ai-productivity-boom-forecasts-countered-by-theory-and-data-by-daron-acemoglu-2024-05
Otherwise...
I think you're being too moderate here.
I am perfectly willing to admit that I could be wrong on this. However, my impression is that the only sane, coldly rational conclusion to draw from current LLM-bot powered "generative AI" is that people are easily fooled by parlour tricks.
As a comparison, worldwide homeopathy sales were about $10 _billion_ last year.
Homeopathy absolutely, categorically, DOES NOT WORK. Not at all, ever, not even a little bit on some people.
Cf.
https://www.howdoeshomeopathywork.com/
It is pure undiluted 100% horseshit. It is a scam. All alternative medicine is a scam. If it worked, it would be medicine. The fact that it's not means it's been proved not to work.
Acupuncture, reiki, reflexology, aromatherapy, whatever: all fake. All pure total scam and nothing else, no exceptions.
There are lots of things that are believed in by millions that categorically are not true. Every religion in the world for a start.
"AI" is homeopathy for tech bros.
It doesn't work. Like homeopathy, the idea, the mechanism, the execution, the result, and the tests all show _it does not work_.
I want to see more of
I am not likely to ever be swayed by "well maybe it's not _all_ bad" arguments.
It _is_ all bad.