Who has been using claude code as a personal researcher? I've added various MCPs for search, deepsearch and basically use claude code all the time to do research on a topic, obviously use plan mode to create the research plan. I've been using the same for learning about a new GitHub repo too! Spawn multiple custom parallel agents and scan the whole repo. I'm on the 20x plan and it's definitely value for money. @AnthropicAI has definitely cooked with claude code.
#claude #anthropic #ClaudeCode #AIResearch
On this day in 1776 men signed their names to a document knowing it could get them killed.
Today, VCs sign a term sheet, ghost the founder in 6 months, spend all day posting on X, and call themselves “builders”
@rahulgs Nice! What are you able to do with inspect that you wouldn't be able to do with devin. I'm curious about build vs buy here, is it mainly about cost?
@jxnlco I queue by default, steer somehow feels to make codex focus on the latest message more than just using that as more info or add to the earlier structure, so if I do streer I say "continue doing what you were doing, and also do this, check this"
mythos will be bad ON PURPOSE on ai "frontier llm research" tasks, this is very very sad for the research community
also the fact that this is un purpose not visible to the user is crazy
@embirico Shifted to codex app, easier to multi task compared to terminal for sure, and easier to track sub agents, sub threads, worktrees. Have a doubt regarding memories, currently logged in via API key, memories won't work for it?