@danirabbit OMG the whole correcting some ancillary and irrelevant point to frame an argument that is winnable while ignoring the whole point is such a thing with dudes.
I wrote a whole ass essay about that shit (funnily, I used pipes as a metaphor) because they would try to pull me into argument about the most random shit. I hate it so much.
@danirabbit sorry, maybe because English isn't my first language but could you speak out your point about missing the forest for the trees?
Like is the missing part related to trying to find a better analogy when the problem is gigantically enormous ethical implications and death of people
or something else?
@lids @danirabbit Lead also (no pun intended) leads to mental health related issues which I thought was well known to the point of being part of insults in the English language as early as the ‘90s and ‘00s. Way more apt a metaphor with ‘AI’ psychosis, stunting basic critical thinking, and other issues created
“Childhood lead exposure has devastating lifelong consequences, as even low-level exposure stunts intelligence and leads to delinquent behavior. However, these consequences may be more extensive than previously thought because childhood lead exposure may adversely affect normal-range personality traits. Personality influences nearly every aspect of human functioning, from well-being to career earnings to longevity, so effects of lead exposure on personality would have far-reaching societal consequences.”
@danirabbit They're only saying that because Cory Doctorow used the "asbestos" metaphor.
Had Doctorow used the "lead pipe" metaphor instead, they would tell you the lead pipe would be the better metaphor whenever you compared it to asbestos.
They don't want to think for themselves. They just want to signal that they belong to the in-group, those "in the know". And spread the word of someone with authority as gospel, like missionaries.
@thepwnicorn @danirabbit "aisbestos" is my current favourite. It certainly is. A wonder-product to cure all our problems now. It won't take as long as asbestos for it to be revealed as an awful idea though.
I'm sure it'll be harder for big ai to hide it than for big asbestos who spent years suppressing science, lobbying governments and denying victims' claims.
Pliny the younger wrote of the illness of asbestos mine workers and he's so old he's in history books. He even wrote about history before it had been invented. Probably because it was the present when he wrote. However we still put asbestos in everything we could in the twentieth century.
"Did you eat LLM chips as a kid?"
@luc0x61 @danirabbit my experience today was another example of why you should have prod and dev separated.
Asked to test entry. Test succesful, entry removed 😂
The other was implementing smtp, and it sent all test emails to logs succesfully. Yup
@danirabbit
✅ Causes brain damage when used to ingest.
✅ Adds a sweet taste to everything, even shit.
✅ Is the preferred weapon of thugs-of-the-age.
✅ Easy to solder together haphazardly.
Yep.
@danirabbit I'm a bit curious how ElementaryOS (well, you lot 😄) are coping with upstream vibecoding if and when it happens?
I have no bright ideas, to be honest - so consider this my copium asking more knowledgable people 😊.
@danirabbit Jeppp, same energy as what @pluralistic wrote a while ago: "AI is the asbestos we are shoveling into the walls of our society and our descendants will be digging it out for generations:"
@sstendahl is he not now using llms to write with?
not the same as generating code but the principle
It's so wonderful to see such faith in the quality of the average human-written code!
But every time I look at Coverity (an excellent static analyzer for finding difficult bugs) I see their "defect density over time" graph which has the "OSS average" set at 0.5 (presumably per 1000 LOC).
Whenever I added new code, I could not get it to stay at 0.0, whether it was my code, other humans or AI. I had to rely on Coverity to determine the errors and fix them (now also AI's job).
@danirabbit unfortunately, it's the few powerful, and possibly non-techie, CXOs who are so convinced that this is such a great idea. The rest of us have little choice but to go along with this hysteria just to keep our jobs.
To be honest, the economics doesn't make sense to me. The amounts being spent on this — I am struggling to see the returns. Only time can tell a few stories. But until then, I need to feed my family and send my kids to college.
@danirabbit For decades we’ve worried about replacing COBOL because the code was often opaque; the underlying business rules transformed into a spaghetti bowl of logic paths exacerbated by the REDEFINES clauses making the data just as awful.
The folks that considered leaving COBOL in place as an acceptable technology choice will buy into vibe-code slop and proclaim (still) “it ain’t broke so there’s no need to fix it.”
@danirabbit
History is so filled with things that people (to be fair, mostly capitalists) thought were really good ideas that turned out to be disastrous that my head spins.
Lead in pipes, paint, etc. is certainly one, but, of the top of my head,
Radium watch faces
Thalidomide
Indiscriminately spraying DDT (the pictures of kids running after the trucks are truly mind boggling)
I’m sure there’s plenty more, but, apparently, we need to learn the same lesson over and over and over
And they thought keeping COBOL developers around was a waste of resources. Wait 'til they see the job boom _this_ creates.
@danirabbit It will be quicker to just back out the years of computer updates and start from there, or maybe just start again.
Both options assume the company is still in business which may be doubtful.
@danirabbit
I have a friend that went back to work coding after a very wet visit at local pub. A few hours of nightly drunk coding took about a week for 3 persons to clean out. (And I finally got through with my idea to always use version control). This was 1992.
AI codes much faster so I think you are probably right.
BUT to be fair, AI might as well sober up and at some point and do a good job fixing the code for us.