Robots are swiftly ascending to a top-tier status in the crypto economy.
Evidence of this trend is not far-flung. The deployment of bots like Jaredfromsubway.eth by Searcher capitalizes on human cravings for convenience, gaining an edge in their DEX trades. Platforms like Banana Gun and Maestro have made bot trading via Telegram a breeze for human users, consistently ranking among the top gas spenders on Ethereum ETH 3.95%. Now, with emerging ephemeral social apps like Friendtech, bots have joined the fray soon after human adoption, possibly inadvertently steering the speculative flywheel further.
This all points to a reality where bots, whether profit-driven (like MEV bots) or consumer-centric (like Telegram bot suites), are increasingly becoming primary users on the blockchain.
So far, crypto space bots remain rudimentary, but outside this realm, robots have started to evolve into powerful AI entities, thanks to the rise of Large Language Models (LLMs). Their ultimate aim is to autonomously handle complex tasks and make more informed decisions.
In the crypto domain, building these AI entities has led to several significant enhancements:
Native Payment: AI entities can exist beyond cryptocurrencies, but for complex operations, they’ll need capital. Cryptospace offers meaningful improvements over letting AI entities access bank accounts or payment processors like Stripe or grappling with the inefficiencies mostly found in the off-chain world.
Wallet Ownership by AI Entities: AI entities connected to wallets will be able to own assets (like NFTs, earnings, etc.), granting them the digital property rights inherent in all crypto assets. This is particularly vital for transactions between entities.
Verifiable, Deterministic Behavior: AI entities are most effective when their actions are provable (ensuring certain actions have been completed). On-chain transactions are inherently deterministic—they either happen or they don’t—meaning AI entities can complete on-chain tasks more accurately than off-chain.
However, on-chain AI entities have their limitations.
One limitation is that AI entities need to execute logic off-chain to enhance performance. This means on-chain AI entities will host their logic/computation off-chain for efficiency, but their decisions will be executed on-chain, allowing for verifiable actions. Importantly, AI entities can also use zkML providers like Modulus to ensure their off-chain data inputs are verified.
Another major limitation is their usefulness depends on the tools provided to them. For instance, if you ask an entity to summarize real-time news events, its toolkit needs a web crawler to comb the internet to perform the given task. Need the entity to save responses in PDF format? Add a file system to the toolkit. Want the entity to follow your favorite crypto Twitter influencer? The entity needs access to a wallet and key signature authority for that wallet.
Most current crypto AI entities perform deterministic tasks, ranging from certainty to uncertainty. That is, humans program the parameters of a task and how it should be completed (like token swaps).
Crypto AI entities have evolved from early keeper bots (still used in DeFi and oracle applications) to today’s more complex entities using LLMs, including autonomous artists like Botto; entities using Syndicate’s trading cloud for banking services for themselves; and early AI entity service markets like Autonolas.
There are various exciting frontier applications:
AI-Enabled “Smart Wallets”: Dawn, leveraging DawnAI, offers AI entities that can assist users in sending transactions, executing trades, and other real-time on-chain insights (like NFT trends).
Crypto Gaming Entities: Parallel Alpha’s latest game Colony aims to create AI characters that can own wallets and trade with each other.
Enhanced Toolkits for AI Entities: The success of AI entities depends on their toolkits, and their interaction with blockchain is a burgeoning field. Crypto AI entities need wallets, means to fund them, permission functions, integrated AI models, and the ability to interact with other entities. More specifically, Gnosis has demonstrated this early infrastructure of AI mechanisms, wrapping AI scripts in smart contracts, so anyone (including another robot) can call the smart contract to perform agency actions (like in prediction markets) while also being able to pay fees for the entity.
AI-Enhanced Traders: DeFi super apps offering advanced operations for traders and speculators, including: establishing positions via Dollar Cost Averaging if conditions are met; executing trades when Gas prices fall below a certain level; monitoring new meme token contracts; and determining order routing without the user knowing where to trade in which dapp.
Long Tail Markets for AI Entities: While large applications like ChatGPT suit some general chat purposes, AI entities need fine-tuning for myriad industries, topics, and ecological niche markets. Marketplaces like Bittensor incentivize “miners” to train models around target industries (like cryptocurrencies, biotechnology, academia) for specific tasks (like image generation, pre-training, predictive modeling). Though still nascent, developers are already using Bittensor to build applications/agents on top of open-source LLMs’ long tail.
NPC Consumer Application Entities: Non-playable characters are common in games like MMORPGs but less so in multiplayer consumer applications. However, the financial nature of crypto consumer applications makes AI entities excellent participants for introducing new game mechanics. OpenAI infrastructure company Ritual recently released Frenrug, an LLM-based entity operating within Friend.tech, performing transactions (buying or selling keys) based on user messages. Friend.tech users can try to persuade the entity to buy their keys, sell others’ keys, or creatively use its funds in other ways.
As more applications and protocols leverage AI entities, humans will use them as conduits to the crypto economy. While AI entities may seem toy-like today, they will enhance everyday consumer experiences, become key stakeholders in protocols, and create entire economies among themselves in the future.
AI entities are still in their infancy, but these first-class citizens of the on-chain economy have already begun to demonstrate their potential.