surely one of the USB°hard word drives around here has a Linux°hard word install°hard word drive on it?
💡 "get the c.2016 laptop running Linux°hard word, because Windows on an HDD°hard word is too slow" probably won't°hard word be done°hard word today.
----
I have come to the conclusion°hard word that, while the machine is good at some coding°hard word projects, it is uniformly°hard word bad 💡 or, at least, not°hard word good at sysadmin°hard word tasks. It's°hard word not°hard word useless°hard word for them, the results for specific°hard word questions are still better than a Google°hard word search°hard word. But if you ask it to do a mildly°hard word complex task, it will almost certainly get something wrong. 💡 there is no reason to think this is any "universal°hard word truth"; it is more a reflection°hard word of the available°hard word training data and the limitations°hard word of an LLM-only°hard word system
----
today's°hard word primary°hard word goal°hard word is the "frame-dispatch°hard word" database schema°hard word migration°hard word.
there are subgoals°hard word:
- set up the PRESUBMIT°hard word check for Python°hard word imports°hard word
- ensure°hard word the local°hard word environment°hard word can create the alembic°hard word schema°hard word migration°hard word scripts°hard word
- add some additional°hard word database fields for desired°hard word features°hard word 🔥 will they be "hashtags°hard word" or "channels°hard word"? or "topics°hard word"?
----
I'm°hard word not°hard word stating°hard word any LLM-specific°hard word goals°hard word with the schema°hard word migration°hard word. But it is possible either "LLM-generated°hard word topics°hard word" or "LLM-generated°hard word copy-edits°hard word" or something will suddenly become very easy.
----
non-Atacama°hard word goals°hard word:
- time-permitting°hard word, today is the day I should get a haircut°hard word. and, go shopping while i am out.
- nothing else remarkable°hard word is planned°hard word.