I used the fly.io GPUs as development machines.
For that, I generally launch a machine when I need it and scale it to 0 when I am finished. And this is what's really fantastic about fly.io - setting this up takes an hour... and the Dockerfile created in the process can also be used on any other machine.
Here's a project where I used this setup:
https://github.com/li-il-li/rl-enzyme-engineering
This is in stark contrast to all other options I tried (AWS, GCP, LambdaLabs). The fly.io config really felt like something worth being in every project of mine and I had a few occasions where I was able to tell people to sign up at fly.io and just run it right there (Btw. signing up for GPUs always included writing an email to them, which I think was a bit momentum-killing for some people).
In my experience, the only real minor flaw was the already mentioned embedding of the whole CUDA stack into your container, which creates containers that approach 8GB easily. This then lets you hit some fly.io limits as well as creating slow build times.
I've come across sobyte.net a few times now. Who's behind it? It's frequently updated with content that, at least to me, is very interesting. If it's just one person, I'd like to know how they approach it. Very impressive.
As I've previously commented (https://news.ycombinator.com/item?id=37236699), this website consists of content copied and translated from other Chinese websites, without permission (I've asked one author). The homepage is pumping out articles much faster than I would expect a person to be able to write them, and I found multiple articles with text stolen and translated from other people's writing, and code stolen verbatim. Flag and move on.
If this article is typical of their content, then they approach it by copy-pasting someone else's entire article into Google Translate and then publishing the output.
I own and use one of those nail clippers and I would say it is worth the money. The ergonomics are really exceptional, all those rounded areas have a purpose + it is fairly heavy, which combined give you great control while clipping.
Then it is just as sharp as on the first day (using it for more than 6 yeas), which sets it apart to my previous clippers which got blunt after a year or two.
The blade also seems to have just the right curvature so that you reach a really nice round shape with 3-5 clips.
Last I love that it looks a little like a Shinkansen.
This is in stark contrast to all other options I tried (AWS, GCP, LambdaLabs). The fly.io config really felt like something worth being in every project of mine and I had a few occasions where I was able to tell people to sign up at fly.io and just run it right there (Btw. signing up for GPUs always included writing an email to them, which I think was a bit momentum-killing for some people).
In my experience, the only real minor flaw was the already mentioned embedding of the whole CUDA stack into your container, which creates containers that approach 8GB easily. This then lets you hit some fly.io limits as well as creating slow build times.