What are the criteria for choosing an agent? What narratives and practicalities are currently most valued by CT?

As early as October, Virtuals launched the concept of AI agent tokenization. It was a bold move, but most people did not care at the time because it was too ahead of its time.

Today, Virtuals has brought 50x, 100x returns to many agents, and suddenly everyone is paying attention.

The problem is that everyone is focused on 50-100x returns. To achieve similar excess returns, you must act early. You need strong confidence and patience to position yourself before others discover it.

Entering now may see returns of 5-10x - still stable, but far from the 50-100x returns early adopters have gained. But that doesn't mean it's too late; you need to have strategic vision. What was 'alpha' two months ago is no longer alpha. The Virtuals ecosystem is overcrowded, and the low-hanging fruit has been picked.

What should be done now?

First, do not limit yourself to one ecosystem. It doesn't matter if Base is with Solana or Virtuals with ai16z - both ecosystems have opportunities. The key is to remain flexible and seek value, rather than tribalism.

What should be valued when choosing an agent?

Essentially, the key to a successful agent is uniqueness and practicality. Without practicality, the behavior of agent tokens becomes irrelevant like 95% of memecoins, losing significance after the hype cycle.

Agents need to have unique products that offer tangible value. If the team truly understands the attention game, it can score points:

  • Engaging legends or stories can be created around the agents

  • Rapid and continuous feature releases

  • Actively meet community needs

When these factors are in place, agents have the potential to find product-market fit (PMF), aligning their products with the market, and subsequently, those teams that continuously execute and innovate (like aixbt) can dominate their niche and become pillars of the ecosystem.

Here are the narratives and practicalities that CT currently values most:


1. Seek Alpha

There is a high demand for agents that provide trading signals and insights. For example, aixbt focuses on providing valuable alpha to users.

Other alpha agents worth noting:

  • Rei: Provides comprehensive token and sentiment analysis, accompanied by detailed charts

  • Agent Scarlett: Focuses on token analysis, analyzing pros and cons and providing insights

  • 3σ: Covers complete project breakdown and interpretation

  • kwantxbt: Specializes in technical analysis, provides trading assistants, sustainably monitors market dynamics, and offers real-time analysis and advice

  • $TRUST: perp agent, can set parameters including entry price, stop-loss, etc.


2. Invest in DAO

Hedge fund-style DAOs operated by human or AI agents, using their funds for investment.

Example:

  • daos.world ecosystem on Base

  • VaderAI, WAI Combinator, sekoia_virtuals, aixcb on Virtuals

  • daos.fun ecosystem on Solana

The strength of investing in DAOs lies in their provision of:

  • Endorsement of high-quality agents

  • Distribution channels for portfolio projects

  • Early and OTC trading rounds of tokens often bring excess returns


3. On-chain trading

On-chain trading is a narrative that deeply resonates with the CT crowd, as there is a buzz over seeing someone turn $1,000 into $1 million.

It's not just about money; it's an excitement of seeing others struggle and succeed.

The first iteration of trading agents will harness this energy. They will focus on portfolio growth, showcasing their frameworks and strategies in real-time. Although this is an early narrative, it is gradually taking shape among some players.

Example:

  • Gekko AI: Self-improving trading agent on Virtuals, leveraging Axal's infrastructure to deliver results.

  • Big Tony: Built on Cod3x, focusing on autonomous execution for professional trading.

  • Project Plutus: Combines real-time analysis with automated DCA execution.

  • On-chain trading agents are expected to become the focus of market attention in Q1/Q2 2025.


4. Developer-centric utility tools

As the ecosystem develops, the demand for developer tools becomes crucial.

Example:

  • SOLENG: Monitors pull requests, conducts preliminary reviews, and serves as an AI judge for hackathons

  • H4CK Terminal: White hat agents responsible for bounties and vulnerability checks

  • CertaiK: Specializes in providing security audits for agents


5. Privacy and confidentiality (TEE)

The trusted execution environment (TEE) pioneered by Marvin Tong on Solana and Phala Network enables fully autonomous agents.

Example use cases/cool experiments:

  • aiPool's 'Unruggable ICO' (through TEE technology, investors can send funds to agents, which will calculate and distribute the corresponding amount of ICO tokens in a secure environment)

  • SPORE: TEE agents with genetic traits passed down to offspring.

  • Neural network experiments, such as DeepWorm.

There are currently no TEE agents on Virtuals; the first to launch will surely see a surge in value.


6. Play beyond narratives

By predicting future narratives in advance, you can achieve returns of over a hundred times. This was the strategy when entering Virtuals at a market cap of less than $30 million.

Emerging narrative for 2025:

  • DeFi agents

  • NSFW agents (AI companions)

  • Robotics

  • Data

  • Gamification of agents

  • Collective/crowd intelligence

  • Infrastructure