I used an agent to build me a full app.
It produced something good looking in one week, but the code is so hard coded, full of ifs and spaghetti style...
AI trainers, please do a better job when you teach these agents!
Every org alive today was architected for a pre-superintelligence world.
Post-superintelligence, the human role shifts from doing the work to verifying it. Agents will make three types of tool calls: to software, to other agents, and to humans. That third one is the design problem nobody has solved yet.
And here's what people miss: the frontier doesn't need to move an inch for this to be transformative. The model capability overhang is already there and it's massive. Everyone already has a Jarvis.
Excited for the value superintelligence unlocks in enterprise.
Great piece by my friend @joannezchen and the @FoundationCap team.
What people don't see is these companies are probably being paid by AI companies for these videos. Not to watch if these people are working. Actually, to teach humanoid robots the craft. When those robots learn, the workers get laid off.
O Lyria 3, do Google, está fantástico. No entanto, ainda está em fase alfa e se recusa a fazer qualquer coisa de cunho religioso. Eu queria uma coisa tipo funk carioca falando de Jesus. Apelei...
https://t.co/2LpMEaoAPY