Advantages & Applications of Cryptographic AI Agents.

Jafar224

Robots are becoming "first-class citizens" in the crypto economy, as recent cases demonstrate this trend.
Searchers deploy robots like Jaredfromsubway.eth, leveraging human users' desire for convenience to front-run trades on DEX. Banana Gun and Maestro allow human users to engage in bot trading via the convenience of Telegram, being among the major gas consumers on the Ethereum chain. Now, in "flash-in-the-pan" new social apps like Friend.tech, robots join the competition upon adoption, potentially inadvertently further fueling the speculative wheel.
All of these indicate that whether profit-driven (such as MEV robots) or consumer-driven (such as Telegram botkits), robots are becoming preferred users on the blockchain.
While cryptographic robots in the crypto sphere are still relatively rudimentary, the rise of Large Language Models (LLMs) is beginning to transform robots outside the crypto sphere into powerful AI agents, with the ultimate goal of autonomously handling complex tasks and making decisions more wisely.

Building these AI agents in the crypto sphere enhances several key functionalities:
Native Payments: AI agents can exist outside of crypto, but if they are to execute complex operations, they need access to funds. Compared to granting AI agents access to bank accounts or payment processors (like Stripe), or dealing with the majority of other inefficient issues existing in the offline world, the crypto sphere offers meaningful improvements for AI agents.
AI Agent Wallet Ownership: AI agents connected to wallets will be able to possess assets (such as NFTs, earnings, etc.), thus granting agents the inherent digital property rights of all crypto assets. This is particularly important for transactions between agents.
Verifiable, Deterministic Operations: AI agents are most efficient when operations are verifiable (i.e., they can ensure certain operations have been completed). On-chain transactions are inherently deterministic, either happening or not happening. This means AI agents can more accurately carry out on-chain tasks.
Of course, there are limitations to on-chain AI agents.
One limitation is that AI agents need to execute logic off-chain to maintain performance. This means on-chain AI agents will host their logic/computations off-chain to optimize efficiency, but agent decisions will be executed on-chain, allowing for verifiable operations. Importantly, AI agents can also utilize zkML providers like Modulus to ensure their off-chain data inputs are verified.
Another key limitation is that the usefulness of AI agents depends on the tools provided. For example, if you ask an AI agent to provide a summary of real-time news, then the AI agent's toolkit needs to include a web crawler to fetch internet information to carry out the task. If the agent needs to save responses as PDFs, then a file system needs to be added to the toolkit. If you want an AI agent to mirror trade your favorite crypto KOL, the agent needs access to a wallet and permission to sign transactions for that wallet.
From the spectrum of deterministic to non-deterministic, most tasks performed by crypto AI agents are deterministic. That is, parameters of tasks and how they are executed (e.g., token swaps) are programmed by humans.

Crypto AI agents have evolved from early keeper bots (still used in DeFi and oracle apps) to today's more sophisticated agents leveraging LLMs, including AI artists like Botto; AI agents can use Syndicate's trading cloud to bank for themselves; and early AI agent service markets like Autonolas.
There are various exciting cutting-edge applications:
Smart Wallets Supporting AI Agents: Dawn utilizes DawnAI to provide AI agents that can help users send transactions, execute trades, and provide other real-time on-chain insights (e.g., NFT trends).
Crypto Gaming Agents: The latest game "Colony" from Parallel Alpha aims to create AI characters that can own wallets and trade with each other.
Enhanced Toolkit for AI Agents: The success of AI agents depends on the toolkit, and interacting with blockchains is currently an emerging area. Crypto AI agents need wallets, access to funds, permissioned functions, integrated AI models, and the ability to interact with other agents. Specifically, Gnosis demonstrates AI mechanisms for this early infrastructure, packaging AI scripts with smart contracts so that anyone (including another bot) can invoke the smart contract to execute agent operations while also being able to pay agent fees.
Enhanced AI Traders: DeFi super apps providing advanced functionalities for traders and speculators, including: executing conditional DCA (Dollar-Cost Averaging) positions; executing trades when Gas prices dip below a certain price; monitoring new meme token contracts; determining order routing, etc.
Long Tail of AI Agents: While large applications like ChatGPT are suitable for some general chatting, AI agents need to be fine-tuned for many industries, topics, and niche markets. Marketplaces like Bittensor encourage "miners" to train models for specific tasks (e.g., image generation, pre-training, prediction modeling) around target industries (e.g., crypto, biotech, academia). While Bittensor is still in its early stages, developers are already using Bittensor on top of open-source LLMs to build applications/agents.

Isha000

That has a lot of potential.

Jiten171

It's also important to grasp the advantages more.

Ritu77

It's very good, and one of the few wallets that can be used.

Samiksha

Robots are really being treated like humans, what a joke!