https://cuelang.org/. I deal with a lot of k8s at work, and I’ve grown to hate YAML for complex configuration. The extra guardrails that Cue provides are hugely helpful for large projects.

If it’s a publicly-accessible repo, then immediately revoke the key and leave it. Force-pushing isn’t good enough because the old commit will still be tracked by Git until the garbage collector kicks in, and you don’t have control over the GC on GitHub (not sure about other providers).

If it’s an internal repo that’s only accessible by employees, then you probably should still revoke it, but you’ve got more leeway. Usually I’d create a ticket to revoke it when there’s time, unless this is particularly sensitive.

Probably because the individual engineers working on Takeout care about doing a good job, even though the higher-ups would prefer something half-assed. I work for a major tech company and I’ve been in that same situation before, e.g. when I was working on GDPR compliance. I read the GDPR and tried hard to comply with the spirit of the law, but it was abundantly clear everyone above me hadn’t read it and only cared about doing the bare minimum.

There’s no financial incentive for them to make is easy to leave Google. Takeout only exists to comply with regulations (e.g. digital markets act), and as usual, they’re doing the bare minimum to not get sued.

I haven’t followed the Sarah Silverman case, but I think it’s likely that’ll end in a settlement. NYTimes is less likely to settle, since they seem to be trying to set a precedent, and they’ve got the resources to do that.

Copyright law is full of ambiguities and gray areas, some intentional and some unintentional. The concept of “fair use” is an example of an intentional gray area, since the idea is that society as a whole will benefit from allowing people to skirt copyright law in certain circumstances, and lawmakers can’t possibly hope to enumerate every such circumstance. It then falls on courts to determine if a given circumstance falls under “fair use”. The problem is courts move very slowly when faced with a new circumstance that hasn’t been litigated before, and that’s what’s happening with AI companies training AI on copyrighted works. Once decisions have been made and stare decisis is established, then they’ll move faster. The NY Times vs OpenAI is the case to watch IMO, since that’s the biggest one challenging the idea that training AI is fair use.

Unfortunately, retrofitting CSP on an existing site can be nightmare, especially if you have external dependencies. At my job, we spent months trying to enable CSP on one our oldest sites, but ultimately gave up because one of our dependencies won’t work unless we added “unsafe-inline” everywhere, which kinda defeats the whole point of CSP.