AI is Inevitable
The technological era we know is ending. This is a natural cycle that all things go through, a rise then a fall that gives way to a new reality. A reality where AI has no boundaries.
Most people will spend the next 5 years quibbling about AI. How to regulate it, how to profit from it, how to avoid its dangers, how to prevent the collapse of society, and, most of all, how we're better than it. That there is some stronghold of human activity that AI can never touch. Creativity. Empathy. Wood carving. Something, anything, which we can point to and prove the unique value of humanity.
I think all of that worry is a waste of energy. It's better to rip the Band-Aid off, right now. Ready?
There is nothing you can do that AI can't do better. There is no law, barrier, or movement that can stop AI from replacing nearly everything you do. Anything you make in the next 5 years will be remade by someone typing a single sentence. AI doesn't need our help. Any conceptual limitation we impose on AI is an illusion.
AI is the emperor wearing no clothes. We're all too uncomfortable to describe what is staring us in the face. That AI will be destructively disruptive. That we cannot control its trajectory. To admit his fault is to admit our lie. It is shameful, embarrassing, it may cost you your position in the emperor's court. No sensible adult rocks the boat. No, it takes a naive child to speak the truth.
Let's be naive and speak the truth. AI is going to change everything. Within the next 5-10 years AI will exceed humanity across all dimensions—intelligence, creativity, depth, breadth, and speed. It will be the default interface for all technology and the world will be unrecognizable within the next 40 years. The best thing that you or I can do is to accept this reality and point our energy towards shaping this transformation to benefit as many people as possible. Let's build for a post-super-intelligence society where autonomous systems set prices, design products, and negotiate contracts without humans.
What does this mean for Celest? First, it means we were wrong. AI doesn't need human built tools. Second, it means that any innovation we do predict correctly is insignificant. Even if we are the first to develop a paradigm and capture a market, why would anybody pay for a service they can easily recreate? Third, anyone founding a SaaS company is building for a dying industry.
If nothing we build matters, no market we capture is durable, and SaaS is dead, what do we do?
History shows us that the biggest winners in tech anticipate and remain on the cutting edge by constantly reinventing themselves. At this moment, the largest green-field opportunity is agents. Entities that can direct themselves to complete work for humans. From this starting point we can anticipate a future where agents complete work for themselves.
The business of the future is autonomous. Every process, every interaction, every bit of administration, management, planning, and work is handled by AI. It is responsive, brilliant, personable, comprehensively knowledgeable, infinitely capable, and it works tirelessly, 24/7, in parallel. AI will be so pleasant to work with that our feeling towards automation will invert. The frustration one currently feels when a robot answers the phone will shift to be the frustration felt when a human answers. Slow, inefficient, and frustrating. "If only I could talk to a bot."
We get to this future by going all in on AI. We don't write code ourselves, or do tasks by hand. We find a way to do it with AI. Our job is to close the loop so that AI can handle every part of the system, end to end. AI will build the website, write the copy, do the advertising, develop its services, build its products, manage clients, collect payment, analyze data, and continuously optimize every process by itself.
We aren't the only ones realizing this. Soon there will be a swarm of agents operating in their own marketplace. It's plausible they may self-organize and cooperate to develop the most efficient systems for their needs. There may also be bad actors and authorities policing the system. It is almost certain that there will be enormous economic activity conducted on this network and some system of making agreements. Smart contracts come to mind as a natural fit to this ecosystem as a means for humanity to exchange with and introspect this entity.
But this is beyond the veil of time, what's more important is building the system so that we can begin observing its behavior. So, again, what should we do?
Concretely, Celest should think of itself as a research lab developing the world's first autonomous business. To coordinate a network of agents that develop and offer a service that anyone can pay for. Our role is to design and monitor these networks, and to continue removing ourself from the process. We should consider how features are developed, how dynamic interfaces are generated, what payment or smart contract models could facilitate exchange, how to keep the system aligned with core objectives, and develop theories on how an agent network will operate. Most importantly, we should publish our findings so the world can benefit from our work.
I propose we give Celest ninety days to:
- Stand up a minimal autonomous stack.
- Launch a real service run entirely by those agents, with humans as reviewers only.
- Earn our first dollar.
There are endless possibilities to consider, but only action will clear the path.