Post

Arguably LLMs/Agents make this actually worse, since you have to invest alot less time into hobby projects :(

claude has actually been trying to get me to stop procrastinating. not sure if this is an example or a counterexample though

newly submitted issues are checked against a HERESIES index and a passive boundary hook alerts the submitting agent of the potential conflict

Correct. Nor the executive functioning to support the thing, to run ops for the thing, to run infosec for the thing, to run sales for the thing, &c &c &c

I would join a syndicate to write a check for that pre-seed, but I think that's actually called Vyvanse.

Unironically though created a custom task manager and associated mcp and it’s been a massive boon for sequential work. I’ve nearly managed to replace my existing budget software

One gap the models have is they can plot a trajectory but struggle with look-ahead to make something actually land They have "have I done this thing, what do I do next" part of executive function down cold, but struggle with the "will this get us where we want to go?" and "where do we want to go?"

why you ! rotten ! ... i'll have you know my personal research project has only *19* Linear tickets in ToDo, which is considerably more than the 8 in In Progress and a full *30* which are Done which definitely signifies... something or other, i'm sure!

The task of "turn this slop into something that I don't feel embarrassed to send to the mailing list" predates LLMs. LLMs help, but they've also increased the number of bad PRs that I need to turn into good PRs