I suppose you are right. If AWS doesn't support STS versions, these will be only applicable to Azure (I know nothing about GCP). It probably makes sense for AWS to stick to LTS versions (I would do the same). But isn't that a choice made by AWS (rather than Microsoft).
This looks really interesting. Getting type safety and editor support to configuration may change quite a bit of how things are done. I don't know if it will gain traction, but if it does, it may really help bringing some long awaited structure to all those YAML files. There appears to be examples specifically for Kubernetes (https://github.com/apple/pkl-k8s-examples).
Apart from the historical value, the most important part of this article now is the "Note of reflection" added 10 years after it's inception:
If your team is doing continuous delivery of software, I would suggest to adopt a much simpler workflow (like GitHub flow) instead of trying to shoehorn git-flow into your team.
I don't think this work flow is relevant any more even for teams that don't do CD, to be honest. It was a messy work flow to begin with and I haven't seen it applied successfully in practice.
GPT-4 was able to do this even though the training data for the version tested by the authors was entirely text-based. That is, there were no images in its training set. But GPT-4 apparently learned to reason about the shape of a unicorn’s body after training on a huge amount of written text.
It's as if they can in some way or other "see".
Maybe my explanation was complicated, but what I describe is not time consuming. It takes at most a few minutes to do all of the things I mentioned. The difficult part is discipline of keeping refactoring separate. Once that's done, the rest is trivial. And not all work include refactoring. That's even easier then.
I agree that this is a challenge. One needs to slice the domain such that it can be covered this way. But this also means more people. In my experience, moving from "activity oriented" teams to "business centric" teams require an increase of the headcount.
I think this may also be a problem with malicious clients. Currently the user enters the username and password via the client's login dialog. It's an effective way to collect credentials.
Old habits die hard. Thanks for pointing this out. I updated the post.
Impressive. I never got past the YouTube URL. But I refresh the captcha until it doesn't contain any numbers so that it doesn't mess with the "25". Still, it the YouTube video contains too many numbers, then it's difficult.
You can change the captcha by refreshing it using the button next to it
I think at that point the game requires two players cooperating. One searches for YouTube videos while the other is feeding Paul.
Because it is about a not-very-well-known feature of the language. Why would it matter that it's old? I don't think I have ever seen this in use in production code.