--- schema: 1 type: video title: 'An AI Future Without the Lock-In — Remy Guercio, Tailscale' description: >- The era of unlimited AI budgets is over, and most teams are locked into whichever tools they picked first. Remy Guercio, who works on Tailscale's AI gateway Aperture, looks back at a year of AI lock-in and asks how teams get out of it while proving real ROI. He breaks an internal AI deployment into four layers (the models, the data connectors like MCP, CLIs and APIs, the interface, and the sandbox) and argues an AI gateway is what keeps them swappable. He shows how Aperture uses your Tailscale network for identity instead of shared API keys, sets budgets per model (unlimited for cheap models, tight for frontier ones), works as both an LLM and MCP gateway, handles authentication for Cursor, Claude Code and Codex in one place, and keeps logs you can use to diagnose what an agent did. In this talk: • Why unlimited AI budgets are ending, and what ROI maxing means • The four layers of an AI stack and where lock-in hides • How an AI gateway ties identity, budgets and logs together • Aperture as both an LLM and MCP gateway SPEAKER Remy Guercio, Strategic Projects, Tailscale X: https://twitter.com/remyguercio GitHub: https://github.com/remyguercio LINKS Aperture docs: https://tailscale.com/docs/aperture Aperture: https://aperture.tailscale.com Tailscale: https://tailscale.com CHAPTERS 0:00 Intro 0:45 A year of AI lock-in 2:05 The end of unlimited budgets 2:45 Wasted tokens 3:20 How do we get out? 4:44 Finance wants answers 5:14 ROI maxing 6:09 The four layers 8:09 Meet Aperture 9:24 How most AI gateways work 10:09 Identity from the tailnet 10:44 Built with tsnet 12:04 Budgets per model 12:44 LLM and MCP gateway 14:04 Auth in one place 16:49 Sandboxes 18:04 Why you need a gateway Recorded at the AI Engineer World's Fair 2026 in San Francisco. Subscribe for more talks from the engineers building with AI. AI Engineer: https://ai.engineer YouTube: https://www.youtube.com/@aiDotEngineer X: https://x.com/aiDotEngineer LinkedIn: https://www.linkedin.com/company/aidotengineer/ #AIGateway #Tailscale #AIEngineer canonical: >- https://liquidgold.cc/channel/aidotengineer/explore/video/an-ai-future-without-the-lock-in-remy-guercio-tails-562b69479ad3 markdown: >- https://liquidgold.cc/channel/aidotengineer/video/an-ai-future-without-the-lock-in-remy-guercio-tails-562b69479ad3.md llms: null parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-23T22:19:21.256Z' language: null children: [] videoId: s-glGHAa1a4 durationSeconds: 1178 publishedAt: '2026-10-11T00:00:35.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 47 endSeconds: 59 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=47' heldReason: rights - ordinal: 1 text: null startSeconds: 59 endSeconds: 127 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=59' heldReason: rights - ordinal: 2 text: null startSeconds: 181 endSeconds: 203 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=181' heldReason: rights - ordinal: 3 text: null startSeconds: 214 endSeconds: 290 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=214' heldReason: rights - ordinal: 4 text: null startSeconds: 301 endSeconds: 323 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=301' heldReason: rights - ordinal: 5 text: null startSeconds: 335 endSeconds: 378 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=335' heldReason: rights - ordinal: 6 text: null startSeconds: 498 endSeconds: 509 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=498' heldReason: rights - ordinal: 7 text: null startSeconds: 585 endSeconds: 619 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=585' heldReason: rights - ordinal: 8 text: null startSeconds: 664 endSeconds: 730 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=664' heldReason: rights - ordinal: 9 text: null startSeconds: 785 endSeconds: 849 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=785' heldReason: rights - ordinal: 10 text: null startSeconds: 860 endSeconds: 1012 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=860' heldReason: rights - ordinal: 11 text: null startSeconds: 1099 endSeconds: 1178 watch: 'https://liquidgold.cc/watch/s-glGHAa1a4?t=1099' heldReason: rights --- # An AI Future Without the Lock-In — Remy Guercio, Tailscale The era of unlimited AI budgets is over, and most teams are locked into whichever tools they picked first. Remy Guercio, who works on Tailscale's AI gateway Aperture, looks back at a year of AI lock-in and asks how teams get out of it while proving real ROI. He breaks an internal AI deployment into four layers (the models, the data connectors like MCP, CLIs and APIs, the interface, and the sandbox) and argues an AI gateway is what keeps them swappable. He shows how Aperture uses your Tailscale network for identity instead of shared API keys, sets budgets per model (unlimited for cheap models, tight for frontier ones), works as both an LLM and MCP gateway, handles authentication for Cursor, Claude Code and Codex in one place, and keeps logs you can use to diagnose what an agent did. In this talk: • Why unlimited AI budgets are ending, and what ROI maxing means • The four layers of an AI stack and where lock-in hides • How an AI gateway ties identity, budgets and logs together • Aperture as both an LLM and MCP gateway SPEAKER Remy Guercio, Strategic Projects, Tailscale X: https://twitter.com/remyguercio GitHub: https://github.com/remyguercio LINKS Aperture docs: https://tailscale.com/docs/aperture Aperture: https://aperture.tailscale.com Tailscale: https://tailscale.com CHAPTERS 0:00 Intro 0:45 A year of AI lock-in 2:05 The end of unlimited budgets 2:45 Wasted tokens 3:20 How do we get out? 4:44 Finance wants answers 5:14 ROI maxing 6:09 The four layers 8:09 Meet Aperture 9:24 How most AI gateways work 10:09 Identity from the tailnet 10:44 Built with tsnet 12:04 Budgets per model 12:44 LLM and MCP gateway 14:04 Auth in one place 16:49 Sandboxes 18:04 Why you need a gateway Recorded at the AI Engineer World's Fair 2026 in San Francisco. Subscribe for more talks from the engineers building with AI. AI Engineer: https://ai.engineer YouTube: https://www.youtube.com/@aiDotEngineer X: https://x.com/aiDotEngineer LinkedIn: https://www.linkedin.com/company/aidotengineer/ #AIGateway #Tailscale #AIEngineer Canonical: https://liquidgold.cc/channel/aidotengineer/explore/video/an-ai-future-without-the-lock-in-remy-guercio-tails-562b69479ad3 · Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks > Missing: no stored document chunks are available.