I keep my agent workflows distinct; we have (for historic reasons) a lot of non-git controlled context that differs between branches, and moving all that around on checkouts is untenable. I use this tool:
That the product is useful does not mean the supplier of the product has a good business; and of course, vice versa. OpenAI has a terrible business at the moment, and the question is, do they have a plausible path to a good one?
Yes, it's good! And I have a lot of sympathy for what I perceive to be Stahl's goals, and indeed, his methods. But I think I need an intercessor or interpreter to his prose style. This is probably a me problem and not a him problem, to be maximally fair.
Björn's writing style is intentional. He tries to make it understandable for people that have deeper understanding of the subject, but to keep out the people that doesn't know enough. It's primarily to convey what Arcan is to the people that can do something with Arcan, but to avoid attention that will ultimately lead to flame wars.
Ståhl _hates_ attention, both from being a quite private individual to having problem with weirdos showing up at his apartment when some of his videos have gone viral here on HN. (Everyone's home addresses are public here in Sweden)
The obsequity loop is fucking maddening. I can't prompt it away in all circumstances. I would also argue that as annoying as some of us find it, it is a big part of the reason for the success of the chat modality of these tools.
Just give the data item a status field, don't fix it by medling with a designated (non)value. And while you are at it you can add some valid-from and valid-to fields for the item. That's how you do it proper.
Hi, I have a well-rounded 3 years of professional work experience in Python-Django, Postgresql and currently building projects in AWS. I'm curious what roles are open to work from Remote, Canada
Yep, I do this in a tmux session. But sometimes I'm on a call and forget and the ROBOT VOICE says "DATABASE DUMP IS COMPLETE" and I jump out of my skin.
