๐๏ธMost agent work will happen invisibly.
@steveruizok calls it "the dark code factory".
But for the moments humans need to steer or intervene, the canvas is the emerging as a place to do high bandwidth collaborative work
New episode with @steveruizok on the story behind @tldraw (Canvas SDK), multi-agent orchestration and where things are going.
Key takeaways:
1) 45 agents working in one document. 15 people ร 3 fairies each with a little chaos but no conflicts
2) Orchestrator agents delegate to workers and share IRC-style chat interface.
2) Annotation is going to be a huge way that people interact with AI tools (images, videos, etc)
๐ฏ Link to the full @theroughmagic pod:
https://t.co/I0iwVvhcKW
p.s. TL Draw are hiring: https://t.co/2eVNsgFIw5
๐๏ธAWS for your mom
Your AI needs a home: The personal cloud (@zocomputer)
Consumers should have access to company-grade infrastructure for personal computing.
Ben (@0thernet) demos Zo live and shows what that looks like when files, automations, memory, hosting and model access live in one persistent system you own.
Personal computing is starting to feel like the early PC movement again, and early days of the internet and P2P file sharing/hosting.
Full pod on YouTube:
https://t.co/t1lAieKjel
or your favourite podcast app...
100% of code written by agents. 100% reviewed by agents. Humans show up at the end.
@ryancarson and leading startups are already doing this. He broke down the full setup on the pod.
The hard part? The plumbing. i.e. Terraforming your repo so the agent can see everything, understand everything and act on it.
Where he thinks this goes next:
code factories โ company factories