Projects

There's a gap in our AI workflows. Humans communicate with agents with either text or audio, but with other humans, how often do we just grab a whiteboard and draw what we mean? We should be able to do the same with agents. As agents become increasingly autonomous, I strongly believe the human touchpoints become more, not less crucial. Models are getting better at understanding visual input and we can leverage that to strengthen communication. Hence I drove Snip to integrate tightly with AI workflows, giving it the skill to render and receive visual IO through Snip.

I find most new tab pages clunky, too many things, slightly jarring, so I made my own. If it looks good to you, give it a shot, give feedback through GH issues.