ai learning resources i actually use

this is a living list of the ai resources i keep bookmarking and coming back to.

it is not meant to be exhaustive. it is the set of classes, docs, tools, and x threads that have actually helped me understand what is going on under the hood, especially around coding agents, claude code, skills, and agent workflows.

i will keep updating this as i find better resources.

classes

• backprop.ai
  one of the better structured learning paths i have found for getting sharper on ai fundamentals.

claude code

these are the links i have been using to understand claude code better, especially how people are structuring workflows around it.

• claude code skills article from an anthropic engineer
  useful if you want to understand how people are thinking about skills in practice.

• cc.bruniaux.com
  open-source learning platform focused on claude code.

openclaw

openclaw is interesting if you want to self-host or go deeper on coding-agent infrastructure instead of staying fully inside hosted tools.

• openclaw docs
  the main place to start if you want to understand setup and architecture.

• hetzner
  useful if you want a vps to run openclaw on.

• nvidia neoclaw thread
  worth reading for ecosystem context.

• minimax
  cheap provider option if you are testing different setups. minimum top-up is $25.

acp, sub-agents, and agent teams

this is the bucket i have been using to understand how multi-agent systems are getting structured in real tools.

• acp + agent teams thread 1
  good overview of the direction.

• acp + agent teams thread 2
  useful follow-up thread with more detail.

• claude code docs: sub-agents
  official docs for how claude code handles sub-agents.

• claude code docs: agent teams
  official docs for orchestrating teams of agents.

• codex subagents post
  useful comparison point for how this idea is showing up across tools.

skills

skills are turning into a real layer in agent workflows, especially if you want reusable behavior instead of repeating the same prompts over and over.

• cryptoskills.dev
  interesting if you care about crypto-specific skill patterns.

• skills.sh
  useful for understanding the broader skills ecosystem.

how i am using this

right now, i am mostly using these links for three things:

• understanding how agent workflows are being designed
• getting better at claude code and coding-agent tooling
• figuring out which patterns are durable enough to use in products, not just demos

if you are building with ai and want practical resources instead of generic hype, this list should be a good place to start.