zlacker

>>i13e+(OP)
I think it’s interesting to consider why touchscreen mobile phones won out over their predecessors.

Touchscreens are the worst interface option. The feedback of feeling a button, or anything made for the job, is better than typing without specific feedback onto a screen. It won out because of you don’t have a set interface it’s bettor.

Now for search, chat gpt will likely always be less reliable than a list of results you can vet yourself for content and source.

That said… I don’t think people care about truth that much these days so one response that feels correct is could be good enough for most. Terrifying times we live in folks.

>>marric+k9
>Touchscreens are worse than buttons

Hottest take I've seen on here for a while. In a very specific, very limited set of mobile usecases, buttons may offer a more pleasant experience than a touchscreen. I only say that because there's usually exceptions to any statement, but I can't actually think of one.

Touchscreens didn't win because of some lazy sheep-like consumerism, they won because the product is superior. If Chat-GPT style models defeat traditional search engines it will be because the product is better, not because people are content with a response that "seems correct."

>>humani+Ra
Trivial counterpoint: you cannot correctly operate a touchscreen without looking at it (unless the display is essentially one single interactive object). Trying to use a music player via a touch screen while running or cycling is basically impossible unless you look the screen. For this purpose, buttons are better.

When defining "better" you need to indicate the metrics you are using. LM's might be "better" when measured with certain metrics, but like most things are worse when measured using other metrics.