@ryanjhunter

exploring programmable companies, orgs and institutions. aka how does a human turn an intention into consequence in the world?

 event horizon++1mm
Joined February 2014
  • Tweets6.3K
  • Following1.5K
  • Followers727
672Media
i am still an effect-maximalist but am currently courting @sabine_s_ sorry i meant Ocaml
ryan j hunter@ryanjhunterJul 10
time pincer (things that become possible with @EffectTS_ ) ryanhunter.io/the-case-for-e…
3
130
Ai native producer right here
Christoph Nakazawa
@cnakazawaOct 15
We open sourced Void v1.1.11 and we made a music video for it. void.cloud
49
I've missed the launch build up but this tweets copy is meh
Garry Tan
@garrytanOct 17
Want to try GBrain but don’t know how? GBrain.io is the best easiest hosted way to get going

One brain for the whole team, and every AI they use

Everyone on a team is prompting alone, each AI with its own notes and its own sign-ins. GBrain gives the team one brain to prompt together, and the team syncs it to ChatGPT, Claude, Cursor or any AI...

gbrain.io
1
63
Openai "gonna connect ur mf" dots
“paula”
@paularamblesOct 17
just found out this ai agent conducted a forensic investigation into where i watched a movie tonight
97
ryan j hunter retweeted
Jacob Posel
@jacob_poselSep 17
If you actually want to master AI, here’s the playbook: 1. You’re only using Claude code or codex now. You’re not allowed to open any other apps. No cowork or chat nonsense either. 2. Every task needs to go through AI. No email, no browser, no figma, NOTHING 3. If AI can’t connect or do a task for you, you’re only allowed to give up after 20 attempts. Yep, 20. You need to learn how to figure shit out with AI. If you do all your work this way for a month, you’ll be a master of AI. Here’s the alpha: being good at AI means knowing how to figure shit out with AI. You need to train that muscle.
43
25
20
757
208,954
user program: print("hello world") compiler optimization: user seems lonely. schedule therapy. purchase self-help book. suppress original output. build succeeded ✓
2
80
deepseek has entered the chat all 21,389 of them
Loftwah
@loftwahOct 17
OpenCode Go is just too good of value to pass up right now
1
96
this man is destined for @ocaml_org i can sense it in the force
Matt Pocock
@mattpocockukOct 15
I think I'm addicted to getting agents to build state machines Every time I see an agent has built a state machine (or even just a reducer) I can feel a huge bump in the robustness And I know that future updates will be trivial and testable Would recommend for frontend
1
2
147
you can build a company whose decisions depend on your judgment. or you can build a company that progressively learns how to make appropriate decisions without demanding your attention every seventeen minutes.
45
i’m trying to institutionally provide for my family you probably have no idea what i mean by that
37
i have instituted a jane_street++ coding style requirement in my new repos, stay tuned
50
welcome to the age of AGI Artificial General Interruptions
227
Ordinal CLI schedule proof: X media attachment.
1
121
Ordinal CLI schedule proof: X media attachment.
119
Ordinal CLI schedule proof, part 1: docs, MCP, REST OpenAPI, and CLI surface are being compared before mutation.
1
138
Part 2: unsupported product formats are explicit residue instead of mystery failures.
27
time pincer (things that become possible with @EffectTS_ ) ryanhunter.io/the-case-for-e…
Guilherme Castro
@skastr052Jul 10
I’ve come from the future to share the message that markdown and prose are bottlenecks. Bet everything on @EffectTS_ as the de facto agent language. Go into debt if you have to
1
4
1,337
Ordinal CLI schedule proof: Happy Birthday USA. Testing X scheduling through the typed Effect CLI.
1
1
67
Happy Birthday, USA.
1
2
99
welcome to the agentic age
1
33
Load more