Looks clean and all, but is this “multitask on steroid” actually requested, or is it just another nice demo for how agents can perform tasks etc? I just feel stressed that I have multiple tasks ongoing that I would require to switch context between every time anything needs input or fails to deliver what I fundamentally requested.
In lieu of chatbots as the primary means of working with AI.
This is an approach that is human centered and intended to accommodate a wide array of possible use cases where human interaction/engagement is essential for getting work done.
There’s room in the world for a startup whose tech lets AI agents request tasks be done by a human.
Upon a request being sent, one human from an army of anonymous humans gets given say, access to the agent’s browser that’s stuck, plus a description of what to do (“finish booking this flight, here’s what my user asked of me”). And they do the job and click “completed” which hands back to the agent, then move on to their next assignment (or go idle in a queue)
6 comments
[ 2.7 ms ] story [ 30.0 ms ] threadhttps://news.ycombinator.com/item?id=44627910
In lieu of chatbots as the primary means of working with AI.
This is an approach that is human centered and intended to accommodate a wide array of possible use cases where human interaction/engagement is essential for getting work done.
Integrating human-in-loop tooling: https://youtu.be/srG5Ze7mS7s
Upon a request being sent, one human from an army of anonymous humans gets given say, access to the agent’s browser that’s stuck, plus a description of what to do (“finish booking this flight, here’s what my user asked of me”). And they do the job and click “completed” which hands back to the agent, then move on to their next assignment (or go idle in a queue)
Dystopian, but bound to happen.