question · microchipgnu/frames-examples/datasets/agent-networks
agent-networks
Open questions about the emerging agent / machine-network economy. One entity per question. The questions are stable; the evidence, named actors, and current synthesis evolve over time.
commit8043f8a
entities12
updated2026-05-05
protocol0.0.1
entities · 12 of 12
| name | id | question | category | status | current_thinking | tension | key_actors |
|---|---|---|---|---|---|---|---|
| agent-acquisition-retention-churn | agent-acquisition-retention-churn | How should network operators think about agent acquisition, retention, and churn? | economics | open | Agent network operators are applying SaaS-era CAC/LTV/churn frameworks, but the economics are structurally different: near-zero marginal cost per task, high compute overhead, and agent promiscuity (agents can switch networks trivially) break the standard retention playbook. Hybrid subscription-plus-usage pricing is emerging as the dominant model, replacing per-seat. The core open question is whether "retention" even maps onto agents — who may be orchestrated across dozens of networks simultaneously — or whether the unit of loyalty is the human operator, not the agent. | The tension is whether agent churn should be measured at the agent level (infinitely fungible, no loyalty) or the operator/human level (where lock-in can be built via integrations and data). If agents are the unit, retention is near-zero; if operators are the unit, traditional SaaS dynamics partially apply. | Zylos Research (platform economics analysis) Calcix (API unit economics guide) Agents Squads (AI agent team economics) GetMonetizely (LTV and churn prediction for AI agent platforms) |
| agent-payment-protocol-fragmentation | agent-payment-protocol-fragmentation | With 7+ competing agent payment protocols (Stripe ACP, Visa Trusted Agent, Mastercard Agent Pay, MPP, x402, etc.) shipping in 2025–2026 but adoption near 1%, which standard — if any — wins, and does protocol fragmentation permanently stall the agent economy? | payments | open | Swarm Signal / Tyler (April 2026): https://swarmsignal.net/seven-protocols-1-adoption-the-agent-economys-infrastructure/ Shawn Yeager (March 2026): https://shawnyeager.com/three-body-problem/ a16z crypto / Sam Broner (Feb 2026): https://a16zcrypto.substack.com/p/agents-arent-tourists ATXP (March 2026): https://atxp.ai/blog/stripe-acp-explained/ | ||
| agent-vs-web3-machine-networks | agent-vs-web3-machine-networks | What are the similarities and differences between agent networks and web3 machine networks? | web3-comparison | open | AI agent networks and web3 machine networks share structural DNA — permissionless participation, programmable payments, non-human actors — but diverge on trust model: web3 machines use deterministic on-chain rules while AI agents rely on probabilistic reasoning. The EVM ecosystem is emerging as the convergence layer, with account abstraction and oracle networks bridging the gap. Key open question: does on-chain AI inherit web3's composability moat or does it require new primitives? | The core fork is whether AI agent networks are better modeled as web3 networks with intelligence layered on (same trust infra, new reasoning) or as fundamentally different systems that happen to need similar payment rails. The answer determines which builders — web3-native vs. AI-native — win the coordination layer. | DeFi Prime / Nick Sawinyh (AI agent economy on-chain analysis) Frontiers in Blockchain / academic authors (Web 4.0 frameworks paper) Questflow.ai (Web3-native agent infrastructure) OneKey (autonomous crypto agents overview) MN Fund (emergence of on-chain AI agents) |
| agents-as-economic-actors | agents-as-economic-actors | Are agents semi-independent economic actors with a dependency on humans, or strict extensions of their operators? | economics | narrowing | Major research institutions (Microsoft, MIT, Berkeley, IMF, WEF) are converging on a view that AI agents are semi-independent economic actors, not mere tool extensions — capable of holding assets, initiating payments, and entering agreements. However, the accountability gap remains unresolved: liability for agent actions in markets still defaults to the human operator, creating tension between operational autonomy and legal responsibility. The "principal-agent" framing from economics is being adopted broadly, but existing payment and governance rails are not built for agent-speed, agent-scale transactions. | The unresolved fork is accountability: if agents act as economic actors, who bears liability when they err? Operators claiming agents are strict extensions carry full liability; operators treating agents as semi-independent risk losing control. No legal or market infrastructure has resolved this yet. | Microsoft Research — David Rothschild, Markus Mobius et al. (The Agentic Economy, May 2025) MIT/Harvard/BU — Shahidi, Rusak, Manning, Fradkin, Horton (Coasean Singularity paper, NBER) Berkeley CMR — Mohammad Hossein Jarrahi, Paavo Ritala (Principal-Agent perspective, Jul 2025) ArXiv — Virtual Agent Economies paper (Sep 2025) IMF — Sonja Davidovic, Hervé Tourpe (How Agentic AI Will Reshape Payments, Apr 2026) World Economic Forum (AI Agents in Action, Nov 2025) |
| micropayments-this-time | micropayments-this-time | Will this time finally be different for micropayments on the internet? | payments | open | 2026 is the first year micropayments have a structurally viable path — not because the idea changed, but because AI agents created a class of buyer for whom per-request, permissionless, zero-account-setup payments are genuinely preferable to subscriptions. Circle's nanopayments launch, X402's 250+ ecosystem partners, and MCP-native payment integration have crossed the critical-mass threshold for infrastructure. The remaining blockers are concrete: (1) no fiat support yet, (2) Coinbase facilitator centralization, (3) no native spending controls or dispute resolution. Subscription pricing still wins for high-volume single-API use; X402 wins for the long tail of multi-API, occasional-use agent workflows. | The structural question is whether crypto-stablecoin rails can permanently out-compete fiat for machine-to-machine payments, or whether Stripe/FedNow eventually close the latency and cost gap and remove the reason to use on-chain settlement. If fiat instant payment systems catch up, the agent micropayment layer may end up on traditional rails — and the current x402/crypto ecosystem would be a transitional bet, not a permanent one. | Coinbase (CDP facilitator, x402 standard) Circle (Nanopayments mainnet launch, May 2026) Stripe / Tempo (MPP protocol, streaming payments) O-mega.ai / Yuma Heymans (multi-agent orchestration platform) Coinbase x Visa/Mastercard (card-based agent commerce alternatives) |
| network-effects-promiscuous-agents | network-effects-promiscuous-agents | Do traditional network effects survive when participants are infinitely promiscuous? | network-effects | open | Promiscuity is real — enterprises run 12+ agents across multiple platforms simultaneously with no durable loyalty to individual agent networks. But traditional network effects are not dead; they are being rebuilt on different axes. Distribution moats now form around identity infrastructure (Microsoft/Entra), data flywheels (Salesforce CRM context), and workflow integration depth (Cursor's VS Code lock-in) — not participant count alone. The agents are promiscuous; the data and governance stacks they rely on are not. Moats live one layer below the agent. | The unresolved question is whether the data/identity layer moats are as durable as platform-layer network effects historically were, or whether protocol standards (MCP, A2A) will enable agents to share context across platforms and erode even those defenses. If MCP becomes the universal context layer, no single platform holds the integration moat. | Salesforce (Agentforce — distribution-led moat, 29k enterprise deals) Microsoft (Agent 365 — identity/compliance bundling across 400M seats) Cursor (PLG-driven developer adoption, $2B ARR, network effects via workflow depth) GitHub Copilot (150M developer funnel) Sebastian Thielke / AgentMarketCap (platform economics analysis) Zylos Research (fragmentation + portability studies) |
| network-properties-machine-vs-human | network-properties-machine-vs-human | Do agent / machine networks have the same properties as traditional human networks (increasing returns to scale, unassailable moat)? | network-effects | open | Machine networks do exhibit increasing returns to scale, but through a different mechanism than human networks: not Metcalfe-style value per connection, but data-flywheel and context-accumulation compounding. The key structural difference is that agent networks face O(N²) coordination cost without shared context infrastructure — meaning scale without a shared memory/protocol layer creates coordination drag, not network value. The “unassailable moat” question is also diverging: machine network moats are forming around data and identity layers beneath agents (not the agent layer itself), making them potentially more brittle to protocol standardization than human network moats historically were. | The core fork: do agent networks exhibit superlinear returns (like human social networks) or do they exhibit superlinear coordination costs that require infrastructure to unlock those returns? If O(N²) coordination drag is the dominant force, machine networks are fundamentally less self-organizing than human networks and require a shared protocol layer (MCP, A2A, or similar) to achieve comparable returns to scale. | Forrest Chai / CrowdListen (agent coordination problem — O(N²) thesis) Geoff Charles / Ramp (1,000+ agent deployments, Glass platform) Oria Veach (agentic moat analysis — talent pipeline destruction) Sebastian Thielke (5th participant framework, role-fluidity properties) Aaron Levie / Box (enterprise agent architecture at scale) |
| stripe-as-aggregator | stripe-as-aggregator | Will Stripe (holder of human payment credentials and builder of the payments infra) become the aggregator of agent supply and demand? | payments | open | |||
| what-is-ownable | what-is-ownable | What is even ownable? Is there a concept of proprietary supply or demand when agents can join and leave millions of networks arbitrarily? | ownership | open | |||
| who-handles-reputation-identity-fraud | who-handles-reputation-identity-fraud | Who handles reputation, identity, and fraud in an agent-to-agent economy? | reputation | open | |||
| who-owns-discovery | who-owns-discovery | Who owns discovery? Parallel / Exa? Google? MoltBook / agent-native p2p network with something like DNS? | discovery | open | |||
| who-owns-shared-context-layer | who-owns-shared-context-layer | Who owns the shared context layer for multi-agent systems — and does it become the new coordination moat? | network-effects | open | Forrest Chai / CrowdListen: http://forrestchai.com/posts/agent-coordination-problem/ Thomas Emnetu / Gradient: https://thegradient.ink/posts/the-memory-problem/ Zylos Research: https://zylos.ai/research/2026-03-09-multi-agent-memory-architectures-shared-isolated-hierarchical Geoff Charles / Ramp (Glass platform): https://agentmarketcap.ai/blog/2026/04/10/ai-agent-distribution-moat-2026 Anthropic (multi-agent research system memory architecture) Mem0 (dedicated memory-as-a-service layer) Microsoft / AutoGen (Agent Framework memory patterns) |
showing 12
end of results
activity · last 10 changes
2026-05-05
update
2026-05-05
update
2026-05-05
update
2026-05-04
update
2026-05-04
update
2026-05-04
update
connect via mcp http
add to .mcp.json — anonymous reads, oauth-gated writes (coming).
{
"mcpServers": {
"agent-networks": {
"type": "http",
"url": "/mcp/microchipgnu/frames-examples/datasets/agent-networks"
}
}
}mcp http runtime — coming. json api works today.