Author: Pan, founder of TP Wallet
For an AI agent to become “autonomous software that is widely deployable,” it needs two basic capabilities:
① Installability
② can be verified
These are precisely two points that traditional Web2 cannot provide, but cryptocurrency systems inherently have.
1. Artificial intelligence requires composability, and composability must be based on verifiability.
The future of AI agents is not one model, but:
- Automatically connect to other services
- Combine other agents
- Write code automatically
- Automated testing
- Automated decision making (including actions involving money).
This is called effective composability.
The question is:
If one agent contacts another, but you can’t verify that the other party actually performed what was expected, the automation ecosystem cannot be completely shut down.
Web2 infrastructure (API + SaaS) cannot provide this strong verification capability for the following reasons:
- The API is a black box
- Data can be manipulated
- The result of implementation cannot be proven.
- Central checking of permissions
- Funds cannot be settled automatically.
In other words:
A Web2 proxy cannot fully trust another Web2 proxy.
Therefore, automation stops at the “pilot level”.
2. Cryptocurrencies provide the “verifiable underlying layer” that AI lacks.
Crypto provides three key capabilities required for the future AI ecosystem at a fundamental level:
① Verifiable implementation
Smart contracts are transparent and provable.
ZK proofs allow verification of complex implementations off-chain as well.
Amnesty International can confirm that “the other side did exactly what I said.”
② Verifiable identity
The DID/key and proxy token are signable.
Solution: Who did it? Is it credible? Was he manipulated?
③ Verifiable value conversion
Smart contracts support:
- Automatic discount
- Automatic payment
- Profit sharing through multilateral cooperation
- Mortgage and lots
- Bond mechanism
This allows AI to manage and allocate funds directly without relying on institutions.
This is something Web2 cannot do.
3. AI programming requires “self-testability,” and when it comes to finance, “self-testability” becomes even more important.
Anthropic’s recent acquisition of Bun is a sign:
Artificial intelligence is about to upgrade from “automatically write code” to a full cycle of “automatically writing → automatic testing → automatic operation → automatic repair”.
But once the AI runs the code automatically and allows that code to directly impact the real world, for example:
- Executing the deal
- Management budget
- Pay API fees
- Participation in economic activities
Therefore, it is necessary to have a verifiable money security system based on cryptocurrencies.
Otherwise, AI will be just a toy and will not be relevant to the real economy.
Cryptography allows AI to:
- Automatic fines for “implementation errors”
- A “good contribution” is automatically rewarded.
- Automatic filtering during multi-agent collaboration
- Automatic settlement without human trust
This is essential for the future AI agent economy.
4. Summarize it in one sentence
AI enables software autonomy; Encryption makes independence trustworthy. Without the verifiability of cryptocurrency, AI agents cannot operate at scale in the real economy.
