49s Give an AI work shift a bounded category of issues. An epic can become research issues first, then tasks in a later wave, turning intent into an artifact. Full conversation: www.youtube.com/watch?v=oyrH... 0 vue
28s An idea at the gym can become an issue for the next engineering shift. Talk it through with Claude, file it, and let the shift pick it up with the rest of the work. Full conversation: www.youtube.com/watch?v=oyrH... 0 vue
20s Could idle cash help identify the next Bending Spoons-type acquisition? Ray’s hypothesis links cash on the balance sheet with late-stage investors wanting it back to put to work. From Ray and Toby’s September 15 conversation. 0 vue
36s Airtable’s unspent billion reminded me of Twitter’s sale: the sellers saw no better way to make money. I think they were right—they struck the deal before the market tanked and sold near the top. From my August 11, 2026 discussion with Toby Oliver. 0 vue
143s Better frontier models give better leverage on tools. Toby’s challenge: once the plan and analysis are done, the actual work can happen elsewhere, unless the provider can lock it in. Ray Deck and Toby Oliver separate high-level thinking from execution spend. 0 vue
42s More agentic workflows might increase token demand overall while decreasing it per job or per unit of value. More time waiting slows the workflow down. Ray Deck and Toby Oliver discuss the agentic token paradox. 0 vue
45s You can have the economic benefits of open-source AI without sending it across the planet. Western models, Chinese models and frontier models give customers more choices. 0 vue
52s Reconsider the assumption that frontier models have to be expensive. Competition gives customers more choice and a reason to revisit the AI spending plan. 0 vue
38s How much discretion does your software platform leave you? A tool can come with a way of building. Ray explains who that helps—and when your own ideas make the fit harder. 0 vue
34s Which parts of your AI instructions survive the next model? Separate today’s model repairs from long-lasting principles for how you do work. 0 vue
24s “You’ve got to let the first version go.” We make songs to work on ideas at State Change. Build This Again | Mental Model Music statechange.ai/songs/10-1 0 vue
24s We make songs to work on ideas at State Change. “It’s not a personal failing, it’s a delta in the flow.” The Delta and the Duck · Be Curious, Not Judgmental statechange.ai/songs/66-2 Music made with help from Suno. 0 vue
42s “How interesting, how interesting.” We make songs to work on ideas at State Change. Joy In The Rain closes with curiosity, movement and joy in the rain. Suno-assisted music. statechange.ai/songs/17-1 0 vue
24s “I’m not the floor you’re standing on, I’m the light.” We make songs to work on ideas at State Change. This chorus is about being valuable without becoming a critical dependency. Suno-assisted music. statechange.ai/songs/8-1 0 vue
31s Pruning can feel like losing ground. This verse makes room for one extraordinary pumpkin. We make songs to work on ideas at State Change. Inspired by The Pumpkin Plan by Mike Michalowicz. Suno-assisted music. statechange.ai/songs/11-2 0 vue
29s “Go and find your spark / In the hardest five percent!” We make songs to work on ideas at State Change. The Five Percent Spark leaves you with a question: what’s your hardest five? Suno-assisted music. statechange.ai/songs/41-1 0 vue
66s A recording service retitles a meeting, and matching the records gets harder. Instead of more string-closeness rules, ask whether the two records describe the same meeting—with the context of everything else that happened that day. 0 vue
56s A bullet-point CV becomes a really big document. Then AI reduces it to a few bullet points. All that communication is wasted. The useful test: I can spend less time and you spend less time on the friction parts, with more value for less cost. 0 vue
19s when I run out of ChatGPT credits, I will, I can bring it over to Claude because there's nothing here that's specific to ChatGPT Full conversation: www.youtube.com/watch?v=GAXW... 0 vue
36s Someone should be able to watch this with the sound off and understand what's going on. Full conversation: www.youtube.com/watch?v=GAXW... 0 vue
48s Selling is about distribution. Thought leadership, your reputation in the market, and maintaining relationships with your network. Full essay: www.youtube.com/watch?v=INQe... 0 vue
34s Shipping converts your knowledge and your reputation into cash flow. When all your time is spent on that conversion, you're not building up those assets. Full essay: www.youtube.com/watch?v=INQe... 0 vue
40s The distance to the next 10 cars, next 20 cars, gets much, much smaller. And I think we're starting to see that among the AI labs. Full essay: www.youtube.com/watch?v=Q2yD... 0 vue
28s The flat monthly price sits on top of usage, the actual consumption of compute that has real cost and scarcity to it. Full essay: www.youtube.com/watch?v=Q2yD... 0 vue
49s Instead of a huge explanation of a convoluted system: “Let me show you a diagram of it.” Ray and Robert explore faster idea transfer in both directions. Full conversation: www.youtube.com/watch?v=sbxU... 0 vue
35s “I see a calculator on your screen.” Robert was 50 kilometers from home. His phone reached the same desktop tools—not a second system. Full conversation: www.youtube.com/watch?v=sbxU... 0 vue
39s Now, the outside game is uh really dealing with the public, right? www.youtube.com/watch?v=J-HH... 0 vue
38s These things aren't about storing data. These things are about generating math. These are compute factories. www.youtube.com/watch?v=J-HH... 0 vue
41s “Coding itself is becoming a commodity.” Toby asks where value comes from when easier coding no longer protects a software business. Ray questions software factories that scale code output. Full conversation: www.youtube.com/watch?v=6ZKu... 0 vue
19s Ray contrasts direct tool-by-tool AI execution with code generation for complex work. The first can fall down; the second lets AI build its own tool. Full conversation: www.youtube.com/watch?v=6ZKu... 0 vue
30s “All the factories, all the raw materials are near each other ... you can get a proof of concept built in a day.” — Toby Oliver Full conversation: www.youtube.com/watch?v=nBUA... 0 vue
41s “The interface between the real world and the computer ... That gives you a moat. Potentially.” — Toby Oliver Full conversation: www.youtube.com/watch?v=nBUA... 0 vue
50s “Effectively what we have is we have code replicators. You just ask for things.” — Toby Oliver Full conversation: www.youtube.com/watch?v=IF3Z... 0 vue
20s Ray recounts one model builder's specialist bottleneck: “There is an opportunity for the new to come along to subvert that.” Full conversation: www.youtube.com/watch?v=IF3Z... 0 vue
38s “Is there a need to run a frontier model all the time?” Toby points to simple tasks and ordinary compute that may move outside frontier inference. Full conversation: www.youtube.com/watch?v=ehB8... 0 vue
25s “Why would you put all your spend, all your money with OpenAI?” Toby’s consultant/intern analogy routes the job: frontier judgment for the plan, routine execution elsewhere. Full conversation: www.youtube.com/watch?v=ehB8... 0 vue
48s Toby argues that AI can act, so schools need to reconsider what children should learn. Full conversation: www.youtube.com/watch?v=4FeO... 0 vue
33s Toby describes AI marking his daughter’s past exam papers and generating questions on weak topics. Full conversation: www.youtube.com/watch?v=4FeO... 0 vue
19s “The conversations we have with AI are so staccato.” Ray Deck on the give and take that AI conversations could have. Full State Change Mastermind conversation: www.youtube.com/watch?v=GDD6... 0 vue
14s We make songs to work on ideas at State Change. This clip is from one of my favorites: The Belly Button Rule. Trust is speed. Full song: www.youtube.com/watch?v=9ziJ... 0 vue
22s We make songs to work on ideas at State Change. This clip is from one of my favorites: The Belly Button Rule. Get closer. Build trust. Full song: www.youtube.com/watch?v=9ziJ... 0 vue
58s “The big UI opportunity in AI is bringing your experience inside the flow.” Bring a smaller product surface into ChatGPT or Claude, where the customer and agent work. Full essay: www.youtube.com/watch?v=5NRQ... 0 vue
22s “They don’t want another destination.” Customers want their agents to use your software in their own environment. Full essay: www.youtube.com/watch?v=5NRQ... 0 vue
15s “VCs want you either to go crazy or die.” Toby on investor incentives when growth no longer supports the old multiples. Full conversation: youtu.be/SuZLAoRr2-E 0 vue
38s “What do you think your valuation is?” Ray asks what Miro’s announced deal and Airtable’s sale mean for a smaller SaaS business still losing money. Full conversation: youtu.be/SuZLAoRr2-E 0 vue
39s CEOs aren't just cutting costs — they're handing headcount to AI. The honest answer to "when do I get to retire?" turns out to be: reduce headcount, and let AI carry the load. Full conversation: youtu.be/MZwIHkmcDmg 0 vue