This is one of the worst takes from LLM enthusiasts.

[zeank@mastodon.social]

@slyborg I didn’t say McDonald‘s because their key differentiating factor is that you know exactly what you get. No matter if you’re in Seattle, Melbourne, Tokio or Paris.

But then, it’s just a shit take anyway. Don’t overthink it.

[colincornaby@mastodon.social]

@gbargoud That microwave post still does numbers.

I’m sort of amazed how everything I wrote about this last year remains absolutely true today.

[nnethercote@mas.to]

@arroz yes, mechanistic vs probabilistic

[stuartcoyle@fosstodon.org]

@arroz Makes one wonder if these people ever read the Dragon Book.

[poleguy@mastodon.social]

@arroz @gudenau our company isn't pushing that hard. But the implicit threat looms... that the old folks who don't embrace more efficient tooling will go the way of John Henry when he lost his job to the steam drill.

A round of layoffs a while back that got rid of several 35+ year veterans makes the implicit threat credible.

[raktheundead@vmst.io]

@aspensmonster@tenforward.social @zzt @arroz @Orb2069: I have said even before the slop merchants came up that the quality of software "engineering" is such that if civil engineers were as willfully incompetent, we'd have a lot more Tay Bridge disasters.

[ariaflame@masto.ai]

@sigismundninja @nils_ballmann @arroz @binford2k Why not roll dice to decide what to do?

[orb2069@mastodon.online]

@raktheundead @zzt @arroz

https://quoteinvestigator.com/2019/09/19/woodpecker/

[theorangetheme@en.osm.town]

@zygmyd @angry_drunk @zzt @arroz Executives were never on your side.

[soulsource@mastodon.gamedev.place]

@noplasticshower @arroz The last time I read about such an attempt (using genetic algorithms) was an excellent showcase of how difficult it is to come up with a scoring function that prohibits weird and practically useless results: https://news.ycombinator.com/item?id=43152407

I'm not saying it's not possible, just that it might not be as easy as one initially thinks.

[noplasticshower@infosec.exchange]

@arroz @soulsource great pointer. Thank you.

It is always easier to build a more powerful machine and suffer the consequences when it misbehaves.

This is the sort of thing we can put to ML.

[veer66@mstdn.in.th]

@arroz I still use disassemblers.

[marczak@mastodon.social]

@arroz it’d be great if you could also let people know that Steve is really a great developer, with a long time focus on macOS, who knows what he’s doing. As far as AI, he’s just sharing his experiences, and is looking at things through an “everypersons” lens. You’re also totally right here, but some of the replies here are really unwarranted.

[gabriel@mastodon.samfira.com]

@arroz no. No no no no.

Use unreviewed LLM code at your own peril. The s**t it generates, especially if you don't understand the code it generates....you might as well throw away your project.