ChatGPT Dots has wasted more of my afternoons than any app I've ever used.
New rule for working with agents: one focused hour of prompting, then walk away.
I've lost whole afternoons babysitting "checking..." queues. The agents don't work harder because you're watching.
Give them a clear hour, close the laptop, go live your life.
Who else had to learn this the hard way?
it is so cool to practice long form video
where what I am developing is a) my curation of reference styles and b) casually speaking into the camera for 10-20 minutes
I know I have been exclaiming this from the rooftops but agentic video editing is such an unlock and I am very excited for my november daily long form challenge & to see what I learn
https://t.co/uceNbqgact
Great to hear you say that! I've been building a Twitter/Blog style journal to write more of these notes-to-self.
Any advice based on what yours have unlocked for you?
@ajambrosino idea: pokemon-style multiplayer game with human-agent native interactions.
> The game is an abstraction that forces simplicity and delight.
> Agents take action in the real world dynamically based on interactions in the game.
> Agents level up in tool capability, permissions and context as you work with them. Delight first, and build ongoing trust.
Why this interface?
1) Invokes a sense of nostalgic delight that is quite disarming to new users (85% of people don’t use an agent regularly yet).
2) The game is a shared workspace where game mechanics limit what an agent can do. A small set of moves makes every action easy to see, approve and trust.
3) This is not an enterprise workspace - it’s a personal one. There’s a lot of work to be done together when you really get into and start exploring someone’s mind palace.
4) Unlocks many multiplayer experiences as you + friends coordinate with your agents + theirs both in and out of the world (on a gameboy vs. groupchat).
i did just describe a poke-metaverse lol
Thoughts?