Title: agent experiments and dialectic disenchantment

2024-12-30 20:57:24 +0800

-----

2024-12-29 03:01:01 -0800

of course no the new agent and the manager agent did not work together. spent a lot of time, four hours?, going around and around the same circular series of errors. finally gave up. now it's morning. I'm going to just delete the directory where that new agent dwells and start over. I think I have to have the cursor agent aware of the manager agent if it's going to set up the second, organizer agent correctly, rather than start a new session where the cursor agent is not aware of the manager agent. worst case scenario I have to zap the manager agent too and start from the very very beginning.

Im just going to assume that I'm getting enough xp here to make all of this worth my timel

since I have the file server machine running, i"l zip up the useless agent's dir and save it on the server until I open a digital museum someday.

2024-12-29 04:23:35 -0800

looks like the html version of these rss posts are being delayed by one upload each time. meaning that this rss post will not have a corresonding html page until I upload the rss post after this.

2024-12-29 05:09:18 -0800

from today's farnam street "brain food" email

"Richard Feynman sums up dialectical thinking:

“We are trying to prove ourselves wrong as quickly as possible, because only in that way we can find progress. ""

which gives me more reason to hold the idea of 'progress' as a noun ("find progress" "make progress" "see progress") under suspicion. this way of talking about progress (as a noun) is so often witnessed in the context of marxism.

is science marxist? marxian even? I dont know. science is defintely materialist though. materialsim is the very process of disenchantment. materialism is dialectically opposed to enchantment. dialectics bootstraps itself into existence by inventing materialism. but this is an enchanted process. science is a product of magic.

science invokes a domain of 'that which can be disproven' and equates it to 'that which can be called true.' eventually this turns into 'that which is real.' the concept of 'the god of the gaps' as a fallacy of informal logic promotes the domain of the disprovable as equivalent to the domain of the real. science denies the possibility of anything outside the realm of the disprovable by a logic as circular as the logic that says the bible is the word of god because god said so. all of this is shored up by science's commitment to reproducability, so that science only argues in favor of whatever has been successful. this is a fair strategy but victory isnt validity. do we really believe that reality adheres to human logic?

in any case, progress is only a noun in some people's imaginations. the actuallity of progress is as a verb. and that is within the context of an arbitrary goal. dialectics may progress us, but do we all approve of the goal?

2024-12-29 06:27:37 -0800

deleted the organizer agent directory and told cursor to build a new one from scratch. running it right now on my own, not through the manager agent. seems to have worked.

yeah there's a kind of tight-rope feeling to doing this.

2024-12-29 12:34:35 -0800

installing coquitts

2024-12-30 04:46:53 -0800

I'm going to have to fuss around with this a little bit to get it to work quite right. put that off a while. I tried to download a model for the tts yesteray. the model ended up being nearly two gigs. there's no way that I can see to check in advance how large a model's zip file is before you download it. this one I never managed to get onto my hard drive. kept getting 'connection interrupted.' I dont know. not worth my time at the moment.

it's monday. less than 48 hours left in the year. the regular world starts back up on thursday, more or less. I can't say that I know for sure how to proceed at that point. I have a little time to figure it out.

today, this morning, I'm going to see if I can get the manager agent and the organizer agent to work together. if I can do that, it ends a phase of that project. a number of ways I can go from there. I can poke around at some services that I've found, APIs, bittensor, whatever else. I can try to add more agents to the system I've started. getting the manager and organizer to work together is hardly a sure thing, though. maybe it fails. we'll see.