In my experience, it's more helpful in some areas than others.
This year I used GPT-4 to write a significant amount of Terraform that was necessary to migrate an application onto AWS.
Writing Terraform, in my opinion, is problem that's broad but shallow. GPT-4 needed to do little beyond summarize documentation, but it was able to do so competently, and that was hugely valuable to me.
Conversely: In my free time, I've attempted to use it for a game-development side-project, and very little of its output has been useful at all.
FWIW, I don't have any pets (and haven't for some time), but I receive physical mail regarding dog food, dog toys, etc, all the time. There's a chance you've been fooled by randomness here.
That said, there's certainly no harm in treating your device as compromised. I do the same. "Better safe than sorry."
I agree. Also, I've had good luck asking GPT-4 to cite sources with its replies. It speeds up the process of fact-checking, and makes "hallucination" detection trivial. (Does the link 404?)
Obviously it's not perfect, but it's no worse than asking human coworkers, which are also sometimes wrong. (I'd prefer not to interrupt my human coworkers with questions with searchable answers anyway.)
I don't understand this way of looking at things. I sometimes learn something new by answering a question from someone who is looking at things from a different direction than I am, or discussing something I don't have a quick answer to. That helps both of us. If I don't even have to think about it, it helps the other person be immediately productive (and learn something) at the cost of taking me out of "the flow" as opposed to having them beat their head against it for who knows how long while I get a a couple more minutes of uninterrupted time. I can hop back into "the flow" pretty much immediately.
If it happened every five minutes that's a different story.
> I sometimes learn something new by answering a question from someone who is looking at things from a different direction than I am, or discussing something I don't have a quick answer to
This can happen on Slack or whatever too, but with much less friction. I'm not ignoring teammates for days - I just want to be able to take 10 minutes or whatever to reach a reasonable place to context-switch versus being forced to do so right now.
> I can hop back into "the flow" pretty much immediately.
I envy you, but I very much cannot do this. If I'm working on something of sufficient complexity, I'm going to lose at least 15 minutes every time I'm forced to make a substantive context-switch. It's a huge drain for me.
This is a great way to get management up your ass and get fired.
End of the day, it’s all about the culture of your workplace. No one here really understands that. I’ve been interrupted more during remote work at some companies than when in person at others. It’s entirely culture dependent.
Easy, just learn to push back or setup mechanisms like office hours to attend for folks whom are in need of support.
Like security, the human chain will always be the weakest link here. If you don't have backbone to stand up for your own time, that's on you - not the remote work inherently IMO.
I maintain one or two projects that are somewhat popular, and I'd love to monetize them (without compromising my FOSS ideals) so I could work on them more.
That said - perhaps I haven't looked hard enough, but it just doesn't seem like projects that accept donations make enough money (on the whole) to be worth the time spent setting up donations.
Sure, I've seen maybe two where a developer ends up being able to work on their project full-time (that's the dream), but most seem to make like $5/year or whatever.
You have to get used to mentioning it in every available avenue (readmes, docs, CLI greeting message, any web UIs, release notes, issue templates, etc), and also be happy with a really slow ramp-up time, but yes, it’s possible. Source: I’m Benjie on GitHub if you want to check out my sponsors profile.
