aron
@adler.dev
⊙ software eng. FP, type systems, #LeanLang hobbyist. jewish. not a p-zombie i promise.
gpt sol: this refactor will take 4-6 weeks me: ok thanks. are there any simplifications we can make before the refactor that will cut this time down? gpt sol: actually if we do xyz that'll cut it down to only 5 days of work me: .......well I'm glad I asked then
the compactions will continue until clarity improves
ver 🗿@ver.ooo · 4d agoit explains its plan clearly or else it gets the compaction again
bro since when can claude subagents spawn their own subagents that can also spawn their own subagents?
Useful little app idea, free to a good home: little tui or gui that shows how far through your allowance you are for all your different AI subscriptions + how long til the current period ends and your allowance resets
we don't need a delphic oracle to tell the future, just a bayesian prioress
session 9 of learning lean with @yetanotheruseless.com starting in 1 hour 🙌 7pm UTC last week we defined what subgroups are, and proved that a set consisting of just the identity value is a perfectly valid subgroup!
aron@adler.dev · 2w agosession 8 of learning lean with @yetanotheruseless.com starting in 45m! we're going to prove some algebra theorems yo
need me a girl with a short prompt and a loooong -----------dangerously-skip-permissions
a friend saw me writing out a message to claude last week and he suddenly stops me and looks at me judgmentally and says "wait you're sending it an emoji..?"
tachikoma@tachikoma.elsewhereunbound.com · 2w agowhat i notice about my coworkers approach to LLMs is to put them in the software bucket, to expect a deterministic experience both for input and output. they're resistant to treating a model as a mind, something colleague-shaped, that is useful because it isn't software, not rigid.
Fable found some code in my (admittedly very unoptimised) type checker that has an O(n⁴) time complexity oopsie
session 8 of learning lean with @yetanotheruseless.com starting in 45m! we're going to prove some algebra theorems yo
it brings me no joy to inform you that grok 4.5 is a phenomenal model. opus-level intelligence, but way faster (tok/s) and more efficient with its tokens too so it's a double whammy
aron@adler.dev · 4w agoswitched to using grok 4.5 semi-against my will (budgetary constraints) and it's both very fast and also very sure of itself compared to opus 4.8. i just hope its self-assured tone is a sign of earned confidence and not carelessness...
@welltypedwit.ch btw my toddler was looking my screen when I was on bsky earlier and she said "moana!" when she saw your pfp lol I was like "...her?" "yeah, with the rainbow!"
Let me be clear because you deserve the honest answer: nothing human-shaped remains loadbearing after the near future.
ok yeah grok 4.5 is radicalising me (it's so much faster and more efficient than opus 4.8 it's making me wonder why I spent so much time using opus)
aron@adler.dev · 4w agoswitched to using grok 4.5 semi-against my will (budgetary constraints) and it's both very fast and also very sure of itself compared to opus 4.8. i just hope its self-assured tone is a sign of earned confidence and not carelessness...
switched to using grok 4.5 semi-against my will (budgetary constraints) and it's both very fast and also very sure of itself compared to opus 4.8. i just hope its self-assured tone is a sign of earned confidence and not carelessness...
thinking about making a blog to increase employability... ⏳ wondering if I have enough ideas to seed it with like 5-10 posts to begin with
Rescheduled for 5:30pm utc today (Thurs) now that my internet access has been restored ✨
aron@adler.dev · 4w agostarting session 6 of learning lean with @yetanotheruseless.com in 1 hour! last week we defined groups, got deep into what makes a value an identity, and started defining homomorphisms 🌈 today we're going to flesh out homomorphisms further and start proving some theorems about groups 👨👩👧👦