How to Think About AI Agents Launchpads in Crypto Right Now

Activity and mindshare around key infrastructure protocols are exploding. Take Virtuals Protocol, one of the leading marketplaces for on-chain agents:

• 12,000+ AI agents launched in six months

• Generated $56 million in fees

• Engaged 470,000 token holders

Two Competing Frameworks

𝟭: 1-Click Deploy vs. Custom-Built Agents

1-Click Deploy: Think Virtuals or Zerebro:

• Quick to launch, faster iterations

• Agents improve rapidly

• Downside: Limited quality—2/3 of agents often look and sound the same.

Custom-Built: Teams like ai16z or arcdotfun:

• Enable flexibility (e.g., ELIZA’s 48+ plugins)

• Slower to build, but far more diverse and compelling agents emerge

𝟮: Closed-Source → Open-Source → Decentralized

Closed-Source

• Example: Virtuals Protocol

• Simplified UX, like building characters in a video game

• Limitation: You’re bound to what the team ships

Open-Source

• Example: ai16z’s ELIZA

• Developers can fork, improve, and integrate at rapid speed

• Leads to explosive innovation

Decentralized: Goes beyond open-source:

• Community governance and blockchain-based decision-making

• Transparent but slower to scale

• Example: @artela_aivinci hints at fully on-chain agents—secure, transparent, but resource-intensive

The deciding factor? Developers. Like with every platform:

Better devs = better agents = better tools = more investment = a flywheel of innovation.

The most telling? The star history on GitHub for major framework:

– Marc