3 ms·

> What if Claude Code operated autonomously with massive parallel compute? Afaik this is not possible as LLMs have linear conversations.

• weiliddat 1y ago
  I guess if we interpreted it charitably, maybe every time there's a decision to be made, it just forked itself and ran with possible inputs it expects? I would say that's how some devs operate too. Instead of waiting for the product/customer to come back, let's predict how they might think and make a couple of possible solutions and iterate over them. Some might be dead ends, we can effectively prune them, some might lead to more forks, some might lead down linear paths. But we can essentially get more coverage before really needing some input. We might argue that it already does that in its chain-of-thought, or agent mode, but having a dedicated "forked" checkpoint lets us humans then check and rewind time in that sense.
• dbbk 1y ago
  Of course it's possible, you can do this today... just create 100 GitHub PRs and assign Copilot Coding Agent