Citadel Dispatch #197: Matt Ahlborg (PPQ.AI) – AI Agents, Privacy, and Payments

Date: March 25, 2026
Host: Matthew "ODELL"
Guest: Matt Ahlborg (founder of PPQ.AI)
Main Focus: Actionable Bitcoin and freedom tech discussion with a special focus on AI agents, privacy, and payment rails.

Episode Overview

In this episode, ODELL welcomes back Matt Ahlborg, founder of PPQ.AI, for a deep dive into the rapid evolution of AI assistants, the unique business and privacy challenges in AI tools, and the powerful intersection of AI, privacy, and Bitcoin-centric payments. The discussion covers the shifting landscape of AI model access, cost trade-offs, open-source innovation (especially in agentic AI), privacy threats—and the realities of building sustainable, user-centric technologies in the Bitcoin and "freedom tech" ecosystem.

1. The Early, Wild West Days of AI (00:28–09:12)

Key Discussion Points

• AI as a Frontier

  • Both recall the current AI explosion feels like early Bitcoin: no "graybeards," rules are made as they go, massive creativity but little established best practice.
  • Teams are small and nimble. At PPQ.AI, going from five to three employees made work more effective (05:26).
  • “There's a lot of creativity out there… we're all creating our own thing right now and just trying to, like, become as efficient as possible.” – Matt (04:03)

• Lean Teams with AI Superpowers

  • An AI-empowered trio can compete with what used to require 30–50 people.
  • The future belongs to those who "live and breathe" AI—purely casual adoption won't keep up (06:47).
  • “If you're trying to get a job these days, you have to be really good at AI.” – Matt (08:55)

• Blurring Roles – Tech + Business Mindsets

  • Success requires a willingness to learn, humility, and some business awareness (11:06).
  • “You can't just be the technical guy living in the basement anymore… you have to be a little more aware of the other facets.” – Matt (12:48)

2. AI Model Access: Subscriptions, Privacy, and Per-Query Trade-offs (13:02–19:21)

Key Discussion Points

• Centralized Proprietary Subscriptions vs. Per-Query/pay ("PPQ-style")
  • Subscription AI (Anthropic, OpenAI, etc.) is heavily subsidized and requires KYC, sacrificing privacy.
    • “Your $200 Claude subscription is getting you about $5,000 worth of credits a month… They're the drug dealer giving you free hookups until you're completely hooked on their drug.” – ODELL (14:06)
  • PPQ's per-query model is more honest, lets users switch between lots of models, but can get expensive.
    • “I would rather try to build something that actually is innovating in this area where it's just a very simple billing system.” – Matt (17:09)
• Vendor Lock-In and the Coming End of Subsidies
  • Proprietary models are converging in quality—vendor lock-in is getting harder; subsidies will end.
  • PPQ focuses on small, honest margins, and long-term sustainability through features, not creative rate-limiting (18:44).

3. Quality Gaps, Model Choice, and Efficiency (19:21–33:42)

Key Discussion Points

• Open-Source AI Catching Up
  • Open models are much better but struggle with complexity and edge cases.
    • “As the tools get more complicated, they really start showing their stupidity.” – Matt (21:00)
    • High-context tasks expose their limits—especially with agentic tools like OpenClaw (21:57).
• "Autoclaw" and Automated Model Switching
  • PPQ’s “autoclaw” routes tasks to cheap or smart models based on input complexity (23:29).
  • “It was saving them like 80% in costs… but then OpenClaw did an update… and it started breaking down.” – Matt (24:00)
• UI vs. API
  • Most profit comes from niche features and high-margin UI, while power users go via API (18:45–19:21).

4. The Future: Smart Routing, Personal Agents, and Privacy by Design (27:12–36:20)

Key Discussion Points

• Smart Model Routing is the Future
  • “It's all about the intelligence of the routing… Many users would err on the side of paying more in making sure that they were getting a quality answer than catastrophic mistakes.” – Matt (27:12)
  • Three-way tradeoff: cost, competence, and latency (29:52).
  • Potential for agents to run "test suites" or upgrade themselves if errors detected (30:30)
• Secure Enclaves for Private AI (31:15–33:42)
  • PPQ now offers open models in secure enclaves; browser UI auto-encrypts, API users can self-host a proxy (31:44).
  • “AI is really going to hit mainstream when it can start solving personal problems... but personal problems are where you want the most privacy.” – Matt (32:28)
• The Need for User-Friendly Privacy
  • ODELL: Personal responsibility limits scale; must dramatically lower friction for privacy tools to ever compete (36:20).
    • “If you have to watch a two hour BTC sessions video to use a tool, like, we're already fucked.” – ODELL (36:12)

5. Open Source Agents and Personal AI Assistants (37:50–43:13)

Key Discussion Points

• OpenClaw: Hype vs. Lasting Innovation

  • Both agree open-agent frameworks like OpenClaw mark a huge shift, not a "flash in the pan" (38:33).
  • The agentic, personal AI future is inevitable—booking travel, generating personalized podcasts, etc. (39:53).
    • “Normal people are going to be doing incredible things with their AI assistants and it's going to be removing a tremendous amount of friction in everybody's lives…” – Matt (39:41)
  • ODELL describes building a personalized AI news desk agent pushing Nostr-signed news bulletins every three hours (41:27).

• Plugging Everything into Open Protocols

  • RSS, Nostr, and Blossom enable easily sharable, persistent, open media feeds (44:21–46:12).

6. The Limits of Open Source and the Battle for the Mainstream (46:13–49:34)

Key Discussion Points

• Proprietary AI Will Dominate for Many; Open Source for Power Users
  • ODELL questions if the masses will always default to centralized assistants from Apple, Google, Anthropic, etc.
  • “The best personal assistant will be the one that can connect to the technologies of many companies instead of one… that's an argument to say that maybe open source or build-your-own assistant kind of wins.” – Matt (48:36)
  • But cost and user experience are key bottlenecks for open alternatives to match mainstream adoption.

7. Payment Models, Maximalism, and Global Reality (51:09–64:57)

Key Discussion Points

• PPQ’s Payment Philosophy: Accept Everything, Focus on Bitcoin

  • PPQ accepts almost every payment type available, with a Bitcoin-centric philosophy.
  • “If Bitrefill had made the choice to be Bitcoin only, I don't think Bitrefill would be around anymore… do you want, you know, kind of a bitcoin first company that makes rational decisions about this stuff or do you want just like very purist type of companies out there who kind of struggle?” – Matt (53:10)
  • Bitcoin maximalist gatekeeping harms practical Bitcoin tech companies—hiring, marketing, and pragmatism matter.
  • ODELL points to clear distinctions: Accepting various payment methods is not the same as building a tokenomics scheme.
    • Comparison: PPQ vs. Venice—Venice heavily pushes their own token and yield schemes; PPQ is user-focused and holds reserves in Bitcoin.
  • “Altcoins and stablecoins… do solve real problems for people all around the world. I look at bitcoin more as a century long battle.” – Matt (56:11)

• Payment Data Insights

  • Over 50% of PPQ user payments are via Bitcoin Lightning (62:24); #2 is Monero; Tether and credit cards are neck-and-neck for #3/4.
  • Open source model use is lower on PPQ than at some competitors; users often settle into a favorite AI provider, but many experiment.
  • OpenClaw integration drove 400% revenue increase, even as current plateau as users tweak their set-ups (65:49).

8. Agentic Payments, L402, and The Future of Pay-per-Task AI (67:04–75:06)

Key Discussion Points

• The Promise and Complexity of Agentic, Pay-Per-Task Payments

  • PPQ now supports L402 for true pay-per-request without a stored balance; works especially well when agents consume from thousands of endpoints (68:54).
  • “It's more ideal if you just pay exactly what you need for that thing at that time and then go about your business… these 402 payments could really come in handy.” – Matt (71:13)

• The Competitive Landscape: Lightning vs. USDC vs. New Protocols

  • Major frameworks for agentic payments:
    • X4.02 (USDC on Base, Coinbase) – “not private,” fees are high—20% of a 1c payment
    • MPP (Machine Payments Protocol, Stripe) – their own version of Lightning, very fast, proprietary
    • Lightning Network – remains most open, current preferred bet for the open, decentralized future despite channel friction (74:32)
  • “The openness of the Lightning network can lead to something much greater. And so that's where I would put my money currently.” – Matt (74:38)

9. What’s Next for PPQ and Final Thoughts (75:18–77:02)

Key Discussion Points

• Shipping Fast, Building Moats
  • PPQ is focused on unique, specialized AI tools to enhance its "aggregator" edge and build defensibility.
  • ODELL hopes for a future where users have "multifunction brain that's just an API key"—maximizing flexibility and superpowers for freedom tech users.

Notable Quotes & Memorable Moments

• “There's no template really, for what's happening right now.” – Matt (03:56)
• “You have to live and breathe [AI] in order to really level up… If you're trying to get a job these days, you have to be really good at AI.” – Matt (08:55)
• “They’re the drug dealer giving you free hookups until you’re completely hooked on their drug.” – ODELL on Anthropic’s subscription model (14:06)
• “Open source models, like when you really put them to the test... you can really start to see them become dumb.” – Matt (21:00)
• “People are warming to [OpenClaw]… normal people are going to be doing incredible things with their AI assistants…” – Matt (38:33, 39:41)
• “If Bitrefill had made the choice to be Bitcoin only, I don't think bit Refill would be around anymore.” – Matt (53:10)
• “Over half [our payments] are Lightning; Monero's number two.” – Matt, on actual PPQ payment breakdown (62:24)
• “The openness of the Lightning network can lead to something much greater.” – Matt (74:38)

Timestamps for Key Segments

• AI’s Wild West, Lean Teams: 00:28–09:12
• AI Subscription vs. Pay-Per-Query: 13:02–19:21
• Model Quality & Cost Optimization: 19:21–33:42
• Secure Enclave, Privacy: 31:15–36:20
• OpenClaw & Personal AI Assistants: 37:50–43:13
• Proprietary vs. Open Source Future: 46:13–49:34
• Payment Philosophy & Market Realities: 51:09–64:57
• Agentic Payments, L402, Lightning: 67:04–75:06
• Final Thoughts, What’s Next: 75:18–77:02

Episode Takeaways

• The convergence of open source AI, privacy tooling, and Bitcoin-powered payments is moving fast, with many parallels to early Bitcoin’s Wild West years.
• Practical, sustainable business models matter; openness to multiple payment rails expands the Bitcoin mission globally.
• AI “agents” and personalized privacy are on the cusp of becoming mainstream, but user experience and cost remain the crucial battlegrounds.
• Open source agentic payments and routing engines are maturing but face stiff competition from centralized (and subsidized) alternatives.
• The future is likely a blend: centralized “easy” AI for the masses, open & private “hard” stacks for power users—but open innovation is catching up quickly.

Relevant Links & Further Exploration:

• PPQ.AI
• ODELL’s AI-powered news desk: citadelwire.com
• OpenClaw, L402, 402index, Lightning Network resources

For more actionable freedom tech conversation, visit citadeldispatch.com!