Post #3733891
2026-06-21 09:52 UTC
Replies (2)
-
@kauer@aus.social 2026-07-11 05:35
@cford@toot.thoughtworks.com @matt@toot.cafe Lets try to compare apples with apples, hm? It's not "lives saved" by cars vs "dogs run over"; it's "lives saved" by cars vs "lives lost and destroyed". The number of people injured and killed by cars each years hugely outnumbers the people saved by them. As in, literally millions to one. About 1.2 million killed - EACH YEAR. The number of injuries is 20 to 40 times higher (reference: WHO statistics). It's true that most things have good and bad uses, though it's VERY hard to argue that about guns, unless you think that people's ability to engage in a fun sport outweighs the desperate cost in human lives that guns bring - and that's just civilian gun use. For generative AI, BILLIONS have been poured into them. One thousandth of the amount so far spent could have probably have solved a pretty large chunk of the world's accessibility problems. As someone else has commented, using generative AI to find accessibility solutions is to use the most fantastically damaging, expensive and inefficient tool imaginable. That some solutions pop out is a good thing, but it is at extraordinary cost to everybody, including those helped. The solutions are founded on wholesale intellectual theft and channel vast amounts of money into the pockets of the worst people on Earth. In other words, that theft is also from people with disabilities, that cost is also paid by people with disabilities. So yeah - it's easy for me, as a person without a disability, to say that this is utterly fucked. But it's utterly fucked.
-
@cford@toot.thoughtworks.com 2026-06-21 09:53
@matt@toot.cafe As it happens, I'm in favour of using LLMs for coding, but I don't find it easy to evaluate the balance of goods and harms.