Side note: that site has well over three hundred vendors listed for cookies! I thought they were generating fake outs without end but I did eventually reach the end
Has anyone seen this implemented? Doesn't look like there is any code to try or links to a demo.
They mention a video but the paper didn't include details but a quick Google found this (which I vaguely recognise now as having been posted on HN before):
I'm the author of the paper. The prototype was never released by Google unfortunately. The implentation we used was a bit hacky, using a visual effects layer over the text editing layer. It was tricky to keep them in sync. A more robust implementation would try to avoid that. I'm happy to discuss specifics in you're interested.
First off thank you for designing this. Both iOS and Android have been focused on streamlining their user experience in the past few years but unfortunately it seems that text editing is just as annoying as before.
In theory how would a "robust implementation" be designed to avoid two layers?
I’m Indian. I absolutely loathe this part of Indian culture. My uncle was once eating a pizza like a naan (ripping pieces off straight from the box) and I called him out on it, and he said he wasn’t doing anything like that…while holding a piece of naan and the pizza looking like rats had a party with it.
It happens basically constantly, I have never heard my family admit to a mistake unless violently confronted by someone with authority.
Leads to all sorts of issues and societal breakdowns, like police beating people up before even trying to communicate.
I think it will depend on your route and the time of your commute. I see fairly distinct behaviour at different times on the tube & Elizabeth Line: come in or leave late and it's full of people who are much less considerate, go in with the majority and there's a bit more social pressure against being inconsiderate.
It's impressive but it's a shame that it's 2026 and despite remarkably lifelike speech, so many models fall on common issues like heteronyms ("the couple had a row because they couldn't agree where to row their boat"), realistic number handling and so on.
In popular areas with lots of overlapping photos, something akin to that was possible with PhotoSynth where it joined together matching photos projected into an inferred 3D space. Sadly the service was discontinued a while back.
Curious to understand what drove that the most?
Certainly for this article, Medium's annoying way it blocks zooming in/out on images when on mobile is limiting and frustrating! But I sense you'll have broader concerns...
The way Medium used my non-monetized account as a subscription/registration trampoline. I was basically always behind a registration/pay wall against my wishes.
This is great - the service and that you're extending it and considering a UI.
Personally I would go with option 2 as the colour from the covers beats the anaemic feel of 1 and it seems more original than the search with grid below of 3.
reply