
Hosted by Valory · EN

Agentic AI is hitting an inflection point. Models can reason, act, and collaborate, but can we trust them? In crypto, the stakes are higher: incentives are real, outcomes are on-chain, and bad agents can cause real damage.Welcome to Agents Unleashed, the show that helps you find signal in the noisy world of agentic AI.Hosted by Thomas Maybrier, this bonus episode features Andrew Hill, CEO of Recall, on how we evaluate and enhance AI agent performance, especially in crypto. We dig into why trust, verification, and incentives matter, how blockchain can provide ground truth, and why community participation is essential to shaping the future of agentic AI.In This Episode:What Recall is and how it evaluates agent performanceWhy trust, verification, and incentives are the core of agent reliabilityHow crypto rails provide transparent verification for AI agentsWhy “agentic AI” is still early and why that’s excitingFrom trading to tooling: the next wave of agent use casesWays to engage with Recall: build, contribute, or simply spectateHow communities accelerate safer, smarter agent developmentResources & LinksAgents Unleashed → https://olas.network/agents-unleashedOlas Whitepaper → https://olas.network/documents/whitepaper/Whitepaper%20v1.0.pdfListener Survey → https://olas.network/blog/pod-surveyFollow Thomas on X → https://x.com/thomasmaybrierFollow Andrew on X → https://x.com/andrewxhillRecall → https://recall.network/🎵 Theme music: “Forward” by Grand Project on Pixabay: https://pixabay.com/users/grand_project-19033897/Sponsored by Olas: Build and own AI agents → https://olas.network

The Agents Unleashed podcast is back with a special bonus episode from Singapore.In this "highlight reel", host Thomas Maybrier shares some standout moments from Agents Unleashed Singapore 2025, where builders, researchers, and founders gathered to explore the next phase of autonomous AI.In This Episode:Who owns the agent revolution? featuring Victor Notaro (Valory)Real, useful benchmarks for AI featuring Danny Zuckerman (Recall)How to think about prediction markets featuring Adrian Lai (Supafund)AI agents: real or hype? featuring Don Gossen (Nevermined)Should AI agents launch tokens? featuring TM Lee (CoinGecko)Risk scores for autonomous agents featuring Med Amine Idmoussi (Bond Credit)Pearl v1: the agent app store featuring Iason Rovis (Valory)The difference between bots and autonomous agents featuring Sydney Lai (GAIA)Chapters00:00 – Welcome to Agents Unleashed00:49 – Who owns the agent revolution?02:35 – Real, useful benchmarks for AI04:46 – Prediction markets: hype or infrastructure?06:05 – Are AI agents real or overhyped?06:53 – Do AI agents need tokens?07:54 – Building trust with agent risk scores08:49 – Launching Pearl v1: the agent app store10:03 – Bots vs agents: why context is everythingResources & Links:🎥 Watch the full Agents Unleashed Singapore playlist → https://youtube.com/playlist?list=PLoP4p0r-X94r0Bgrt7PPQIR6vlzOIkg0N&si=wiQOmhLqjrt0IkpS📄 Olas Whitepaper → https://olas.network/documents/whitepaper🐦 Follow Thomas on X → https://x.com/thomasmaybrier🎵 Theme music: “Forward” by Grand Project on Pixabay → https://pixabay.com/users/grand_project-19225115/Sponsored by Olas → https://olas.network

We made it. 24 episodes. 6 months, 1 exciting ride through the future of agentic AI. Welcome to Episode 24 of the Agents Unleashed Podcast, the show that helps you find signal in the noisy world of autonomous agents.This episode marks the official end of Season One. To celebrate, we’re taking a look back: the highlights, the lessons, the recurring themes, and the big questions we’re still chasing. Whether you’re a long-time listener or just joining us, this is your chance to catch up on what really mattered in our first season.In This Episode:What people do, and don't get about AI agents ft. David MinarschTradeoffs and security risks of computer-use agentsHow a model can be "loyal" ft. Oleg GolevWhat is an AI agent, anyway? Ethereum's 10 year plan ft. Tomasz StanczakResources & Links:Catch up on all episodes → https://www.youtube.com/watch?v=ADAew8z_k80&list=PLoP4p0r-X94r1FA7yoOwRqvOjiYGSNQojFollow Thomas on X → https://x.com/thomasmaybrier🎵 Theme music: “Forward” by Grand Project on Pixabay: https://pixabay.com/users/grand_project-19033897/💬 Like, subscribe, and leave a comment to support the show.Sponsored by Olas: https://olas.network

AI agents are under attack. From prompt injection exploits to invisible system takeovers, new security threats are forcing a rethink of how we build, test, and trust autonomous systems.Welcome to Episode 23 of the Agents Unleashed Podcast, the show that helps you find signal in the noisy world of agentic AI.Hosted by Thomas Maybrier, this episode investigates the growing danger of prompt injection, and how attackers are learning to hijack AI agents to steal data, drain wallets, and impersonate users. But it’s not all bad news: Thomas also explores how open-source workflows, decentralized protocols, and new evaluation tools like Olas Predict may offer a more secure path forward.In This Episode:Why prompt injection is the #1 threat for AI agentsReal-world hacks from BlackHatWhat red teaming has revealed about agent behaviorHow Olas agents handle trust-minimization, verification, and incentivesChapters00:00 – Welcome to Agents Unleashed00:50 – Real-world prompt injection at Black Hat03:21 – NVIDIAGTC demo: multi-agent red teaming04:07 – Why prompt injection is a systemic threat04:56 – The risks of compromised agents05:16 – How do we make AI agents trustworthy?06:19 – How Olas manages riskResources & Links:Agents Unleashed in Singapore → https://olas.network/agents-unleashedOlas Whitepaper → https://olas.network/documents/whitepaper/Whitepaper%20v1.0.pdfListener Survey → https://olas.network/blog/pod-surveyCopyPasta License Attack → https://hiddenlayer.com/innovation-hub/prompts-gone-viral-practical-code-assistant-ai-viruses/Follow Thomas on X → https://x.com/thomasmaybrier🎵 Theme music: “Forward” by Grand Project on Pixabay: https://pixabay.com/users/grand_project-19033897/💬 Like, subscribe, and leave a comment to support the show.Sponsored by Olas: Build and own AI agents → https://olas.network

AI chatbots are lying more than ever.Why is this, and what can we do about it?Welcome to Episode 22 of the Agents Unleashed Podcast, the show that helps you find signal in the noisy world of agentic AI.Hosted by Thomas Maybrier, this episode dives into two fast-moving developments with big implications for truth and trust in AI systems:A sharp rise in chatbot misinformation, driven by changes in model refusal rates and real-time web access — and the emergence of autonomous betting agents that use market dynamics to generate more reliable outputs.At the center of this conversation is Olas Predict, a new multi-agent forecasting protocol built around competitive prediction markets and agent performance incentives.In This Episode:Why AI models are getting more confident — and more wrongWhat NewsGuard’s audit reveals about LLM truthfulnessThe consequences of declining refusal ratesBetting agents and “market-based truth”How Olas Predict works, and what makes it differentWhy ownable, autonomous agents may be the path forwardResources & Links:NewsGuard Report → https://www.axios.comOlas Predict → https://olas.network/predictOlas Whitepaper → https://olas.network/documents/whitepaper/Whitepaper%20v1.0.pdfListener Survey → https://olas.network/blog/pod-surveyAgents Unleashed episode 19 → https://youtu.be/wgtvNr8BiGsDatabricks raise → https://www.pymnts.com/news/investment-tracker/2025/databricks-raises-1-billion-to-launch-new-products-for-building-ai-agents/Salesforce's worker shakeup → https://ppc.land/salesforce-ceo-details-ai-agents-handling-50-of-company-work/xAI launches "Macrohard" → https://www.webpronews.com/elon-musks-xai-launches-macrohard-to-rival-microsoft-with-ai-agents/New web standards could limit agent's access to the web → https://archive.is/n6iZYFollow Thomas on X → https://x.com/thomasmaybrier🎵 Theme music: “Forward” by Grand Project on Pixabay: https://pixabay.com/users/grand_project-19033897/💬 Like, subscribe, and leave a comment to support the show.Sponsored by Olas: Build and own AI agents → https://olas.network

Welcome to Episode 21 of the Agents Unleashed Podcast, the show that helps you find signal in the noisy world of agentic AI.Hosted by Thomas Maybrier this episode dives into two pillars shaping the future of the agent economy:Prime Intellect’s Environments Hub, an open-source initiative to build a shared library of reinforcement learning environments, designed to keep open AGI competitive with closed labs.Ethereum’s ERC-8004 (“Trustless Agents”), a draft proposal for giving AI agents a first-class identity on-chain. We’re joined by co-author Marco De Rossi, AI Lead at MetaMask and co-author of the proposal, to unpack what it means.In This Episode:Why open environments matter for the future of open-source AGIThe cultural mission behind Prime Intellect’s Environments HubHow ERC-8004 could become the ERC-721 of AI agentsThree approaches to “trust” for decentralized agentsThe link back to Olas: Registry pattern and the Mech Marketplace as live proofsResources & Links:Prime Intellect Environments Hub → https://primeintellect.aiEthereum ERC-8004 Draft → https://eips.ethereum.org/EIPS/eip-8004Olas Whitepaper → https://olas.network/documents/whitepaper/Whitepaper%20v1.0.pdfDavid Minarsch on X → https://x.com/david_enim/status/1958927494825845239Olas: Build and own AI agents → https://olas.networkListener Survey → https://olas.network/blog/pod-survey🎵 Theme music: “Forward” by Grand Project on Pixabay: https://pixabay.com/users/grand_project-19033897/💬 Like, subscribe, and leave a comment!Follow Thomas on X: https://x.com/thomasmaybrierSponsored by Sigil Fund, a top-performing crypto hedge fund since 2018. Learn more: https://sigilfund.com

Are autonomous AI agents about to redefine the future of Ethereum? |This week, host Thomas Maybrier sits down with Tomasz Stańczak, Co-Executive Director of the Ethereum Foundation and founder of Nethermind, to explore Ethereum’s vision for the next decade, from major protocol upgrades to the growing intersection of AI and blockchain. What we cover: • Strategic priorities for Ethereum’s next upgrades • Why cypherpunk values like privacy + decentralization still matter • How AI agents could use Ethereum as their coordination + verification layer • What “identity is defined by privacy” really means in an AI worldResources & LinksOwn & monetize your AI agents → https://olas.networkAgents Unleashed events & replays → https://olas.network/agents-unleashedFollow Thomas on X → https://x.com/thomasmaybrierListener Survey → https://olas.network/blog/pod-surveyThis episode is sponsored by Sigil Fund, a top-performing regulated crypto hedge fund since 2018. Learn more: https://sigilfund.com#aiagents #ethereum #olas #decentralizedAI #agentsunleashed

Are AI agents going to take over Ethereum? Welcome to Episode 19 of the Agents Unleashed Podcast, the show that helps you find signal in the noisy world of agentic AI.In this episode, host Thomas Maybrier explores Base’s HTTP x402 vision for machine-to-machine payments, shows how the Olas ecosystem already enables millions of agent-to-agent transactions through the Mech Marketplace, and looks at what happens when AI wanders to the fringes of science (hello, weird and wonderful physics). We also dig into how OpenLedger credits data contributors so user-owned AI isn’t just a slogan, it's a path for credit and compensation.What we cover:• Base’s HTTP X402 (machine payments) — and where it fits• How Olas makes agent-to-agent transactions real today (Mech Marketplace)• Registry, staking, and delivery guarantees most teams miss• AI’s strange-but-promising physics discovery — and why “AI explores, humans verify”• How OpenLedger gives data contributors credit and a cut• What it all means for user-owned, decentralized AIResources & Links🎯 Own & monetize your AI agents → https://olas.networkAgents Unleashed events & replays → https://olas.network/agents-unleashedFollow Thomas on X → https://x.com/thomasmaybrierListener Survey (win 500 OLAS, T&Cs apply) → https://olas.network/blog/pod-surveyThis episode is sponsored by Sigil Fund, a top-performing regulated crypto hedge fund since 2018. Learn more: https://sigilfund.com

The most anticipated AI model of the decade has finally landed — but does GPT-5 live up to the hype?Welcome to Episode 18 of the Agents Unleashed Podcast, the show that helps you find signal in the noisy world of agentic AI.In this episode, host Thomas Maybrier unpacks OpenAI's long-awaited GPT-5 release, the surprise return of open-weight models, and what it all means for the future of decentralized, user-owned AI agents.Plus:Why GPT-5 is more about coding than conversationWhat the return of open weights means for AI buildersThe battle between centralized and decentralized AI heats upWhy Olas focuses on agents you own — not rent💡 Featuring insights on GPT-5's architecture, performance, pricing wars, and the implications of AI models that can disappear without warning — plus a look at Sentient's fully decentralized model hosting experiments.00:00 – Intro & what's in this episode01:18 – The long road to GPT-503:00 – Early takes from influencers04:00 – Coding focus and enterprise adoption05:23 – The GPT-4o backlash: when models disappear08:11 – OpenAI's surprise: open-weight models return09:35 – Portable intelligence vs transparent intelligence10:25 – Sentient & decentralized models11:05 – Final thoughts: the future belongs to ownable agentsResources and Links🎯 Learn more about how to own and monetize your AI agents at https://olas.networkAgents Unleashed event replays → https://olas.network/agents-unleashedFollow Thomas on X: https://x.com/thomasmaybrierListener Survey: https://olas.network/blog/pod-survey🎵 Theme music: “Forward” by Grand Project on Pixabay → https://pixabay.com/users/grand_project-19033897/This episode is sponsored by Sigil Fund, a top-performing regulated crypto hedge fund since 2018.#AIagents #GPT5 #OpenAI #DecentralizedAI #Olas #AgentsUnleashed

Welcome to Episode 17 of the Agents Unleashed Podcast, the show that helps you find signal in the noisy world of agentic AI.Hosted by Thomas Maybrier (Digital Media Producer, Valory), this week’s episode answers the fundamental question: “What is an AI agent?” We’ll define key terms, explore historical roots, and explain the OLS approach to agents so you can understand what sets true agentic systems apart from bots and microservices.This episode begins with an origin story, tracing agent research back to the 1940s and highlighting definitions from Stan Franklin and Art Graesse. Then, we’ll map agents on the augmentation–automation spectrum and introduce David Wamsley’s modern definition. Next, get a primer on the Olas vision: self-custody, discoverability, proof of active agent, and end-user accessibility. Finally, we dive into multi-agent systems with Olas Predict and explain sovereign vs. decentralized agents in the Olas ecosystem.If you’re building or evaluating AI agents and want clarity on what “agentic” really means, this one’s for you.In This Episode:Historical foundations: Agent research from game theory to biology (Franklin & Gresse, 1997)Defining AI agents: Perceive, reason, act—autonomy vs. human-in-the-loopSpectrum of agentic AI: Augmentation to full automationOlas' four pillars: Self-custody, discoverability, proof of active agent, and accessibilityMulti-agent competition: How Olas Predict leverages decentralized prediction marketsSovereign vs. decentralized agents: Ownership, trust minimization, and transparencyResources & Links:Sigil Fund → https://www.sigilfund.comOwn your agent with OLS → https://olas.networkAgents Unleashed event replays → https://olas.network/agents-unleashedFollow Thomas on X: https://x.com/thomasmaybrierListener Survey: https://olas.network/blog/pod-survey🎵 Theme music: “Forward” by Grand Project on Pixabay → https://pixabay.com/users/grand_project-19033897/Sponsored by Sigil Fund → https://www.sigilfund.com