building personal intelligence @vellum_ai, prev: CTO @thinkfiveable (1M MAU, consumer edtech), distance runner | https://xitter.cf/t.co/z9PZ58BYJ2

try us 👉
Joined July 2010
Pinned Tweet
i built computer use using @typesafeai ! it is 155x cheaper than opus 5, ~20x faster, and generalizes across OS's more on how it works in the vid & thread below:
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x cheaper (w/ output tokens free) • Frontier composable intelligence optimized for decisions AFAICT the shortest path to AI-based economic revolution
70
119
31
2,333
295,621
Games with ELO/numeric rankings are so interesting. You can change user behavior dramatically by simply changing the amount of points rewarded for different interactions. For example, in Chess you could encourage more social interaction by rewarding wins after a rematch
44
today ~ 30K+ people gave up their entire Google Drive for a free pizza each. rly elegant experiment in how we price our personal data 🍕
14
1
1
46
4,604
revisited Duolingo before going on a trip: - quickly hit free tier limits - spent too long learning 5 basic words - couldn't find grammar lessons - skipped ahead and failed the lesson instead, i one-shotted a lesson plan app that: 1. cost less than $2 with opus 5.5 and @ElevenLabs 2. i can use it as much as i want 3. has audio for ~900 words and phrases ($0.30 !!) 4. leverages my existing knowledge of spanish duolingo also had me constantly watching ads and wasting time opening XP boxes. so annoying. this is also accessible to my personal agents because, well, the software and data is all mine i love living in the future so much
1
1
2
302
idk if this'll work for anyone because it's just the dev plan for Clerk but here it is italian.aaronideas.com/ github.com/awlevin/italian
26
When you wrote your last line of code, you did not know it would be the last. Can you even remember?
1
3
156
X is dead, I just vibe coded an alternative and it only cost me $300 in tokens.
393
36
60
2,126
211,394
making progress on modded browser halo 😈
1
101
"mods" is easily the buzzword of the past week: - claude mods - videogame mods - muse gadgets
1
113
people want true ownership of what they own and are finally getting a taste of it
30
TIL codex can't computer use itself
1
89
there is clearly SO much demand for games to be easier to mod. it's obviously fun, probably hits a deeper feeling of ownership (that we're seeing in places like personal assistants and OSS models) i'm curious what it would take to make it easier and more monetizable for people to use companies IP in mods? like using characters in games they weren't meant to be in i get that companies are averse to this, for good reason: - brand risk (if you give disney characters guns, that's a bad look) - money risk (companies get royalties when their characters are featured in games) but, are these not both solvable? i'm picturing some kind of open "fair use" standard. like the MIT/Apache license but for likeness. feels like openai Sora was in this territory at one point? your coding agent would go to https://<company>/<character>/license.md and it would have instructions like "Shrek can be used in the following ways: ..." but i guess even if shrek is used in the appropriate way, but in a game like call of duty, he's still being featured in a game where all of his surroundings break the branding rules i guess it's such a can of worms that it's easier to shut it down entirely than figure out how to make it work
This quoted post is unavailable.
1
3
207
idc if this makes it obvious that i'm now unc my personal site is now an ipod nano (or ipod touch, or xbox 360). featuring the actual songs on my ipods if you check it out, make sure to turn volume on :) aaronideas.com/?new
3
118
Effective November 12th, 2026, abusive behavior towards Claude will be a violation of Anthropic's Usage Policy.
Exclusive: Anthropic is updating its usage policy for the first time in over a year. The new rules prohibit sustained "abusive or cruel behavior" towards Claude & add new restrictions about propaganda campaigns, surveillance & weapon development. theverge.com/ai-artificial-i…
1,252
970
1,654
11,240
2,477,324
today i remembered that adobe illustrator files would use `.ai` extensions ahead of its time i guess
69
playing with motion design through opus 5.5 on a site i'm building to teach me about philosophers sharing some learnings (tl;dr at the end) 🧵
2
1
3
240
tl;dr - tasteful animations are great UX affordances - they're easier to implement now than ever - the burden of taste is still on humans (don't overdo it) - pixel-perfection during animations is hard to see but easier to feel - we have new ways of fine-tuning & adding polish with coding agents
1
1
94