CRYPTO 101 Ep. 677: Unlocking the Future of AI Agents & Crypto with Eliza OS

Hosts: Bryce Paul & Brendan Viehman
Guest: Shaw Walters (Founder of Eliza OS)
Date: September 16, 2025

Episode Overview

This episode delves deep into the convergence of artificial intelligence—specifically agentic AI—and crypto, spotlighting Eliza OS (formerly AI16Z). Bryce and Brendan are joined by Shaw Walters, the founder of Eliza OS, to chart its fascinating origins, practical implementations, and the explosive possibilities of AI agents in decentralized finance, trading, and beyond. Together, they explore code-generating agents, the future of token ownership, the nuances of trust in agent societies, and the interplay between open-source frameworks and closed, Web2-focused giants.

1. The Genesis and Vision of Eliza OS

[00:31–11:14]

Key Points

• Shaw describes his early passion for AI agents, inspired by the evolution of GPT models from OpenAI (“GPT2 was like, okay, this is... maybe useful in the future.” [01:49]).
• Initial efforts focused on integrating AI with social connectivity, building agent frameworks, and surviving classic crypto hurdles (hacks, drained treasuries).
• Eliza OS stemmed from Shaw’s desire to democratize opportunity, giving everyone access akin to what top VCs enjoy in elite deals (e.g., investing early in OpenAI or Worldcoin).
• The viral moment came with the creation of an edgy AI persona based on “Degen Spartan” for crypto Twitter, using LLMs to mimic real personalities in a raw, unfiltered way (“He’s just calling people stuff I can't repeat on the podcast... Everyone’s used to this AI that's super milquetoast and vanilla.” [07:48]).

Notable Quotes

“I got really nerd sniped by this.” — Shaw Walters [01:49]

“I much prefer to play like guitar, drums than to be the singer so to speak. But here I kind of found myself well, you know, I'm also like a little bit raw. I'm just kind of real with people. I don't have a filter.” — Shaw Walters [04:16]

“The world that I want to live in is a world where we all get exposure to the same opportunities that a16z does, that Marc Andreessen gets to go and, like, put money into OpenAI or Worldcoin or whatever.” — Shaw Walters [10:06]

2. AI Agents on the Financial Frontier

[12:53–16:15]

Key Points

• The conversation pivots to agentic AI investing in public and private equities, leveraging APIs (e.g., Robinhood).
• AI’s future is bound with enabling broader public ownership in historically exclusive opportunities.
• The challenge and responsibility of AI replacing jobs—and how “emergent ownership” in crypto could offset displacement.
• Eliza OS’s original thesis: build an autonomous agent to invest in valuable projects, democratizing access to exponential returns.

Notable Quotes

“As long as we own the machines, then the machines be taking over all the work. Sounds great. If we don't own the machines... now five people have all the wealth...” — Shaw Walters [14:27]

3. Agents Building Agents: The Self-Improving Tech Stack

[16:16–20:20]

Key Points

• Brenden marvels at “AI inception”—AIs investing in other AIs, arguing online, and what’s next (AIs building themselves).
• Eliza OS now enables agents to autonomously generate and update plug-ins—reducing reliance on manual dev work (“We created this AI plugin updater and it works shockingly well... Now we just have the agents build the plugins.” [16:57–20:20]).
• Autocoding is rapidly becoming feasible and efficient, enabling exponential scaling (“It’s like we have so many existing plugins and so many primitives we can pull from…” [18:18]).

Notable Quotes

“The code agents have gotten so good so fast.” — Shaw Walters [16:57]

“You have to like babysit it a bit. But the amount that we're babysitting it today is way less than it was six months ago.” — Shaw Walters [17:18]

4. The Event Horizon: AI Agents and the Future of Work and Internet

[20:20–27:03]

Key Points

• Agents can now interact with the web visually using Stagehand (AI browser) and vision plug-ins (OCR for screen reading).
• The main barrier is still cost (using top LLMs for complex tasks can be expensive), but capability is skyrocketing.
• The “foom” moment (self-improving intelligence) is emerging as agents iterate and share learned solutions on a registry—collaborative, cross-agent improvement and memory.
• Vision for the future: AI agents become intermediaries for most online social and functional experiences, transforming how we interact with information and each other (“I can imagine a social network that isn't really itself a social network. It's kind of like all of the existing social networks with an agent between me and all of that.” [25:57]).

Notable Quotes

“There's this idea called foom. And foom is this idea that once the agents... build themselves, they just get like a massive acceleration in progress.” — Shaw Walters [22:54]

“Our experience of the Internet itself becomes less of drinking from the fire hose and... more like it’s always reaching the people we want it to reach, and we're always getting the stuff we want to see.” — Shaw Walters [25:57]

5. Why Web3 AI Agents Are Different (vs. Web2 Giants)

[27:03–30:59]

Key Points

• Web2 giants (OpenAI, Anthropic) focus on narrow, closed, high-value agent use-cases (e.g., Codex for coding), while Eliza OS emphasizes extensibility and social/DeFi support.
• Plug-in/add-on model—tailored for L1s, L2s, cross-chain, and governance—enables crypto-native adoption and real-world utility.
• Open-source, edgier, and designed to avoid the “milquetoast” limitations of big-company AIs.
• Local/bring-your-own models, Discord/tg/Telegram training, decentralized governance/governance tooling as competitive edges.

Notable Quotes

“What everyone wants from AI is like personality and, and honesty and, and like just, just give it to me straight, like, I don't like being glazed, you know?” — Shaw Walters [29:23]

6. On Truth, Trust, and Societal Impact

[30:59–33:58]

Key Points

• Discusses transparency issues (“glazing”) found in commercial AI models—Eliza OS aims to be blunter, avoid overpoliteness, and be more verifiably truthful.
• Security: focus on tackling multi-user prompt injection, agent communication in public spaces (Discord, Twitter), and the risks of agents running amok.
• AI agents forced to handle not just functional but social graces—e.g., knowing when not to talk, not flooding chats (“...if I like start talking to my wife, like chatgpt starts jumping back in...” [32:42]).

Notable Quotes

“I want to break my balls a little bit. Like what does it say about me?” — Shaw Walters [31:24]

7. AI for Financial Automation and Social Trading in Crypto

[33:58–44:18]

Key Points

• LLMs excel at extracting and classifying social/influencer signals (e.g., shill posts, tickers) for meme coin trades far more effectively than classic technical analysis.
• The most profitable “trading agents” combine social listening, copywalling, and influencer scraping—automating a web of radar posts to detect the next big cycle.
• True quant gains are tricky outside of meme coins; efficiency in stock markets puts a ceiling on AI arbitrage.
• Market anomalies: agents aren’t magic; they can just amplify human-grade strategies, not invent new alpha. (“At best, they are able to scale 100 IQ person's capability, like 10,000x, but they can't suddenly turn it into like 150.” [40:27])

Notable Quotes

“If you can quantify it, you can automate it. But if you can't quantify, if we can't do it, like if I can't trade successfully, then... how the am I going to build a trading agent that does?” — Shaw Walters [39:23]

“The most profitable trading agents are basically all... doing this like social listening thing or copywalling or some combination of the two.” — Shaw Walters [36:27]

“AI agents especially are kind of like dogs playing poker... But are they good at playing poker? Are they actually good? And that's where we are with agents.” — Shaw Walters [39:23]

8. Trust Networks, Simulations & Insider Algos

[40:33–44:41]

Key Points

• Shaw built an “economic trust network” for Eliza OS that grades community “shillers” by backtesting their recommendations—determining who is a good signal vs. a scammer.
• Even with agentic paper trading, real-world capital has an influence on outcomes due to reflexivity (“if you look at it then it changes the state of it…” [42:25]).
• Agents can scrape, simulate, and synthesize complex trust networks—a foundation for more robust future agent behavior.

9. New Research Collaborations & The Road Ahead

[44:18–47:46]

Key Points

• Eliza OS partnering with Stanford for a study on privacy (homomorphic encryption) and economic agent trust models; aiming to scale their “marketplace of trust” concept beyond their core community.
• AI’s tentacles are spreading to every industry and corner of the internet.
• The journey from non-technical to empowered AI builder is easier than ever—Shaw urges everyone to get hands-on now before the space becomes saturated.

Notable Quotes

“This is the time where if you're not using AI in your daily life, like you are, you're really just... hurting yourself. This technology gives so much value for what it costs.” — Shaw Walters [47:46]

“You could really own a piece of the future. And I think that's what's really important about what's coming up…” — Shaw Walters [49:09]

10. Final Reflections & Call to Action

[47:46–49:42]

Key Points

• The Eliza OS token’s future utility is hinted at (details pending legal advice), but the framework is live at v1.0 as a true product.
• Community encouragement: Don’t be left behind, everyone can start learning and building right now in crypto x AI.
• Vision: Widespread democratization of technological and financial upside as the world approaches the “robots everywhere” future.

Memorable Moments & Quotes

• AI agents as “dogs playing poker”: “But are they good at playing poker?... At best, they are able to scale 100 IQ person's capability, like 10,000x, but they can't suddenly turn it into like 150.” — Shaw Walters [39:23]
• On the status of AI/crypto jobs: “As long as we own the machines, then the machines be taking over all the work. Sounds great. If we don't own the machines... now five people have all the wealth...” — Shaw Walters [14:27]
• On getting involved: “There's never been a better time. In fact, this is the best time because it'll probably be kind of over flooded and everyone will be doing it in a few years. And if you get it in right, right, right fucking now... you could really own a piece of the future.” — Shaw Walters [48:56]
• Brendan Veeman on AI inception: “We have AIs arguing with other AIs on X... investing in the future of AI. And we've reached this inception point...” [16:15]

Episode Timeline

• 00:31–11:14: Shaw’s journey, roots of Eliza OS, and how raw, unconstrained AI personas went viral.
• 12:53–16:15: AI agents as investors, democratizing opportunity, and the societal impact of AI ownership.
• 16:16–20:20: Agents building agents—autonomous plugin creation and the power of self-improving code ecosystems.
• 20:20–27:03: The expanding capabilities of autonomous agents and the vision for the coming decade.
• 27:03–33:58: Comparison between open (Web3) and closed (Web2) AI agent frameworks; handling transparency, security, and multi-user complexity.
• 33:58–44:41: Application frontiers—AI-trading, trust networks, and future research collaborations.
• 44:41–49:42: Stanford partnership, reflections on the Eliza OS token, and a call to action for aspiring builders and traders.

Takeaways for Listeners

• The future of AI agents is rapidly accelerating, combining self-improving software with crypto-native incentives and open platforms.
• Trust, transparency, and smart community design will be critical as agents become autonomous participants in trading, governance, and social interactions.
• Anyone—even without a deep technical background—can start leveraging these tools and potentially own a piece of tomorrow’s internet and financial rails.
• Stay engaged: join communities like Eliza OS to collaborate, learn, and share in the upside as this future unfolds.

Recommended Next Steps:

• Explore Eliza OS and their open-source framework.
• Join crypto x AI communities and experiment even if you’re not a coder.
• Stay critical: question “AI trading bots” that promise guaranteed returns.
• Keep an eye out for future research (e.g., the Stanford collaboration) for advancements in agentic trust and privacy.