Agents may be the platform era on steroids. Think about what that means for corporations. One logical endpoint is the last company: a singular firm whose agents do the work that used to need hundreds of thousands of businesses.
Even if Anthropic doesn’t get us to the end of corporations, Fukuyama style, the direction is clear. We won’t have big tech anymore. We’ll have mega tech.
The platform era worked because it was ruthlessly effective: aggregate demand, drive marginal costs towards zero, and let network effects do the rest. The dangers came bundled with that success: lock-in, rent extraction once competitors were gone, single points of failure, and power concentrated in a few boardrooms.
Agents amplify all of it. Hardening against that looks like open protocols over closed APIs, data portability, interoperable agents and antitrust that moves at the speed of software. For corporations, the counter-move is just as clear: own the orchestration layer, the trust and the distribution, and they’ll capture even more value than the last era did.
So the question is: what are the alternatives? The same agents that make mega tech possible also put enormous leverage in the hands of tiny teams, co-ops and open communities. A few counterweights they can reach for:
- Build on open protocols and open-weight models, so no single provider can switch one off or raise the rent.
- Pool agents and data in co-ops, so members share the scale advantages instead of renting them.
- Make portability a feature: let users leave with their data and their agents, and win them on trust.
- Go deep on niches the mega firms find too small or too local to bother with.
- Federate rather than aggregate, so many small services can interoperate like one big one.