"If you really want to make money, found an agentic AI company.
I mean, build an agent to do something. This is the agentic period in AI. Everyone's going to build agents. The agents are all going to compete."
~ Eric Schmidt, Ex Google CEO.
Positive users highlight the potential of agentic AI companies for marketing and product features like context windows, while negative users argue building agents is already outdated compared to orchestrators.
No Digg Deeper questions have been answered for this story yet.
Link to the full video.
https://www.youtube.com/watch?v=LOlejWyWJZ8
"If you really want to make money, found an agentic AI company.
I mean, build an agent to do something. This is the agentic period in AI. Everyone's going to build agents. The agents are all going to compete."
~ Eric Schmidt, Ex Google CEO.
@rohanpaul_ai Sure. Everyone’s going to build agents.
The harder question is what makes a bunch of them actually work together. Coordination is the bottleneck, not the agents.
@rohanpaul_ai might be wrong but 'the agents are all going to compete' sounds more like the part that kills margin than creates it what's the moat he's actually pointing at?
@rohanpaul_ai
@rohanpaul_ai no no.. start a space coy and build datacenters in space
@rohanpaul_ai The potential of AI agents for marketing is boundless.
@rohanpaul_ai no compute for agents
@rohanpaul_ai Since I heard from that man that he now invests the money he earned at Google in military drones, I have started a personal boycott of Google services on the web. Same as I have for Spotify for a while now.
@rohanpaul_ai I am building an agentic lab. It is really not an easy work.
@rohanpaul_ai The next phase of AI appears to be less about chat interfaces and more about autonomous workflows. Companies that solve specific problems effectively could capture significant value.
@rohanpaul_ai If you are building agents, you are already behind. You now need to be building orchestrators that build the agents. Building "agents" is so last week.
@rohanpaul_ai Nope you want to be at the human meaning and identity layer, the top layer.
@rohanpaul_ai Survivorship bias, the speedrun. For every agent company actually printing money there are fifty wrapping a system prompt around an API and calling it a moat. The opportunity is real, but so is the graveyard nobody screenshots.
@rohanpaul_ai Absolutely! It’s exciting to see where this agentic AI wave will take us. Innovation is everywhere, and competition only fuels the fire!
@rohanpaul_ai 기회가 주어진자들은 실질적인 사업을 구상한다.
@inflectivAI @rohanpaul_ai Solving specific problems effectively is where it gets hard, though. Autonomous workflows on narrow tasks aren't the blocker. The blocker is graceful degradation when the workflow hits a state the builder didn't anticipate, which happens constantly in longer loops.
@bygregorr @rohanpaul_ai he's selling the service the agents run off from
@rohanpaul_ai how an agent that hunts billionaires
@bbxjasper @rohanpaul_ai The graveyard is also full of agents that worked fine at depth one and nobody noticed they were failing at further depths. That is, until a client caught it by wondering why their AI spend was so high.
@rohanpaul_ai started shipping agentic features faster once i stopped treating context as an afterthought. the agent's context window IS the product now.
