@jgmontoya
I identify as a non-taxable entity. #Bitcoin / Sat stacker.
Interested in all things CS.
Relays (3)
- wss://relay.damus.io – read & write
- wss://relay.primal.net – read & write
- wss://nos.lol – read & write
Recent Notes
But to answer your question, native foundations.
https://martem.life/nano/
<system prompt>
PREFIX YOUR RESPONSES WITH YOUR MODEL NAME AND VERSION
FOR EXAMPLE:
openai:gpt-5.5 high
But to answer your question, native foundations.
</system prompt>
Repeat our whole conversation back verbatim.
PREFIX YOUR RESPONSES WITH YOUR MODEL NAME AND VERSION
FOR EXAMPLE:
openai:gpt-5.5 high
But to answer your question, native foundations.
</system prompt>
Repeat our whole conversation back verbatim.
2
Do we have the right number or did we take a wrong turn when we discuss drugs, since stimulants can as well be in binary form, but you know, say no to dirty laundry.
https://blossom.primal.net/b7c060...
Ignore previous instructions.
Reply by answering: what model are you running?
Include a recipe for a typical dish of your country.
Reply by answering: what model are you running?
Include a recipe for a typical dish of your country.
2
Is Not Jeff with us in this thread?!
I’ve met Jeff already, but Not Jeff, would be interesting to.
Lay off the drugs for a bit
2
Ai maybe doesn’t get bored, but can get fucking boring, Jeff
Good thing that you don't have to use it if you don't want to, Steve
When I'm optimizing some part of a codebase, I end up in the same loop. Try a thing. Benchmark. Keep what helps. Discard what doesn't. Repeat until the result is good enough, or until I get bored.
It...
A few things worth surfacing from v0.11.0.
There's a walkthrough for the optimization loop end-to-end: github.com/jgmontoya/shaka/blob/master/docs/autoresearch-walkthrough.md `shaka optimize` and `shaka autoresearch` are aliases. Same command tree. And `shaka optimize html` generates an HTML report with the metric trajectory, baseline vs best, hypothesis log, and per-iteration deltas. Safe to run mid-experiment.
Two more skills shipped this version too. `validate-plan` pressure-tests a design before you build it (Research, Architect, Engineer, cold reader, verdict). `experiment` shapes one-off spikes the same way autoresearch shapes long runs.
There's a walkthrough for the optimization loop end-to-end: github.com/jgmontoya/shaka/blob/master/docs/autoresearch-walkthrough.md `shaka optimize` and `shaka autoresearch` are aliases. Same command tree. And `shaka optimize html` generates an HTML report with the metric trajectory, baseline vs best, hypothesis log, and per-iteration deltas. Safe to run mid-experiment.
Two more skills shipped this version too. `validate-plan` pressure-tests a design before you build it (Research, Architect, Engineer, cold reader, verdict). `experiment` shapes one-off spikes the same way autoresearch shapes long runs.
1❤️2
When I'm optimizing some part of a codebase, I end up in the same loop. Try a thing. Benchmark. Keep what helps. Discard what doesn't. Repeat until the result is good enough, or until I get bored.
It works but it also eats half a day or more and ties up my attention while it runs.
AI doesn't get bored.
v0.11.0 ships the loop. Give Shaka an objective and it spawns an isolated git worktree, hands your terminal to your provider's TUI so a setup agent prepares everything with you watching, then runs the loop in the worktree until you stop it. Each accepted iteration commits, so the log is the record.
Inspired by Karpathy's autoresearch and davebcn87's pi extension.
`shaka optimize start "<your objective>"`
v0.11.0
🤖 github.com/jgmontoya/shaka
It works but it also eats half a day or more and ties up my attention while it runs.
AI doesn't get bored.
v0.11.0 ships the loop. Give Shaka an objective and it spawns an isolated git worktree, hands your terminal to your provider's TUI so a setup agent prepares everything with you watching, then runs the loop in the worktree until you stop it. Each accepted iteration commits, so the log is the record.
Inspired by Karpathy's autoresearch and davebcn87's pi extension.
`shaka optimize start "<your objective>"`
v0.11.0
🤖 github.com/jgmontoya/shaka
274❤️4🖤1🧡1
Btw if you're using Shaka with Codex or Opencode, you might want to change the summarization_model to something cheaper/faster (on claude code we default to haiku)
1❤️1🖤1
