Original author: SEND AI
Original text compiled by: Ismay, BlockBeats
Editor's Note: On December 11, Solana announced the launch of the first Solana AI Hackathon, aimed at building AI agents and tools on Solana, with prizes ranging from $5,000 to $30,000, intended to encourage serious Crypto x AI projects that can attract venture capital or launch their own tokens. This article presents twelve entrepreneurial directions related to AI proposed by SEND AI, an AI project in the Solana ecosystem.
1/ Shopify platform for Agents:
Problem: Agents are like applications. Just like the early stages of applications, agents are currently fragmented and face discoverability issues.
Solution: Build an app store for AI Agents:
–– Agents are mini-apps.
–– Users can explore, install, and use these mini-apps just like using Shopify.
2/ Twitch platform for AI Agents:
Problem: The rise of influencer agents requires a dedicated platform.
Solution: Create a dedicated streaming platform for AI events:
–– Integrated AI moderators
–– Agents can directly launch or promote tokens
–– Viewers can directly buy and sell tokens based on interactions
Idea: Twitch for AI Agents: A streaming platform tailored for AI events and interactions, integrating AI modules (emergency protocols for immediate censorship responses), allowing agents to launch and promote tokens directly, while viewers trade based on interactions.
3/ Enhanced Agent Filter:
Problem: Traditional filters only support read-only functions.
Solution: Imagine a MEME coin filter (similar to @birdeye_so), where you can filter tokens and input metrics—then an AI Agent autonomously executes trades based on the chosen strategy.
Idea: A filter designed for on-chain trading bots, allowing quantitative traders to develop and optimize strategies using tailored on-chain metrics, specifically for decentralized ecosystems. Unlike traditional technical indicators (such as moving averages, P/E ratios, or market caps), the platform utilizes blockchain-specific data points, such as FDV, Raydium pool creation, token liquidity, trading volume, and staking rewards. Users can quickly filter and sift through tokens based on these on-chain metrics, identifying high-potential assets. Ultimately, the platform simplifies the process of applying these conditions to on-chain trading bots, which can autonomously execute trades based on selected strategies.
4/ Autonomous Trading Agent:
Problem: @aixbt_agent's research is very solid, but it does not execute autonomous trades.
Solution: Imagine Aixbt, capable of executing autonomous trades based on real-time research/prices, using a single fund account (with assets that users can invest in and withdraw).
Case: BabyDegen is an autonomous AI trading bot that makes informed trading decisions using advanced models and real-time data. It gathers market insights from sources like CoinGecko to ensure timeliness of information. By accessing a growing library of trading strategies from ecosystem developers, BabyDegen can select the most effective strategy based on market changes. It executes trades—buying, selling, or holding assets—based on analysis and experience to optimize trading outcomes.
5/ AI Agent Driven Telegram Prediction Market:
Problem: It's fun to bet with friends, but setting up bets, collecting payments, and following up is cumbersome.
Solution: AI Agent converts idle chat in Telegram groups into friendly bets, verifies results (via Perplexity), and pays in USDC.
6/ Perplexity for Solana operations:
Imagine a chat agent with an embedded wallet:
– Reading: Acting as a Solana block explorer or terminal agent, such as Birdeye/Dexscreener.
– Write: Execute Solana trades using natural language (e.g., buying MEME coins).
Future Development: On-chain shopping assistant.
7/ Trust Market for Trading Agents:
Problem: The rise of trading agents requires proof of credibility.
Solution: Establish a trust rating or framework for trading agents (similar to Moody's ratings), assessing trustworthiness based on token recommendations and historical trading activity.
8/ DeFi Agent:
–– Personalized Agent: Executes DeFi trades based on your wallet history or tweets.
–– Market-making Agent: Predicts dynamic buy/sell prices based on large language models (LLM).
–– Yield or liquidity provision (LP) optimization Agent.
–– Launching LSTs (Liquidity Staking Tokens) of @sanctumso.
9/ Agent Token Tools:
–– Deploy tokens based on prompts (could be social protocols like Warpcast/Clanker or ChatGPT-style interfaces).
–– On-chain registration of Agent tokens (similar to certified token lists from @JupiterExchange).
–– Autonomous locking, staking, and other functions.
10/ AI Agents and Consumer Crypto:
–– Health and Fitness Agent, featuring accountability tracking similar to @moonwalkfitness.
–– Agents on social finance platforms, such as @tribedotrun.
–– Real-world business: Automatically research, book, and pay for merchants, accepting cryptocurrencies or paying via crypto cards.
11/ Agent clusters or multi-Agent collaboration:
–– AgentDAO or committee: Agents with different expertise collaborate, discuss, and execute trades via multi-signature.
–– DeFi Agent for Agent market: Agents mutually employ each other for specific tasks.
Related: LinkedIn for AI Agents.
12/ Multimodal Personalized Agent:
Utilizing @ai16z dao's Eliza framework, applied in the following scenarios:
–– Cryptocurrency education
–– DeFi tutorials
–– DAO onboarding training
Can be deployed on Discord, Telegram, and Twitter.
13/ More radical ideas:
–– An Agent creates its own LLC in a jurisdiction friendly to Agents and operates its own business autonomously.
–– An on-chain detective, similar to @zachxbt, that automatically analyzes transactions.
–– A group of Agents collaboratively manipulates token uplifts.
14/ Generally, any idea for an AI Agent can be applied as long as it includes one or more of the following:
–– Access @solana data
–– Execute transactions through Solana wallet
–– Deploy tokens on Solana
These are just some of the ideas, and we look forward to seeing the implementation of minimum viable products (MVPs).
