Ryan Carson joined Intel because he expects compute to become one of the world's most valuable resources and wants to help developers use AI.
2
AI tools have made it much cheaper for people with domain expertise and beginner development skills to ship small products and test whether they solve a real problem.
3
People working in companies should connect their work to the company's top-level OKRs or KPIs, then break those goals into actions they can take.
Summary
Ryan Carson describes how his experience building Treehouse led him toward Intel and AI development. Treehouse began with a simple goal: give people a trusted, friendly way to learn coding without the cost of a computer science degree. After Treehouse was acquired in 2021, Carson relearned software development with ChatGPT, TypeScript, Next.js, Vercel, and OpenAI APIs, then shipped an app in about a month. He sees AI making it possible for people with deep knowledge in another field to build useful products without becoming expert ML engineers. Carson also discusses the limits of building products before finding a real problem, the value of communities and in-person relationships, and his hopes for AI that can help people reflect on daily conversations. At Intel, he focuses on growing an AI developer community and argues that competition in compute will expand access and lower prices. For employees, his advice is to connect personal work to company-level OKRs.
Treehouse was created to open coding careers to people outside the computer science club
Ryan Carson says his first web development job exposed an unfair shortcut. His computer science degree persuaded the employer that he could do the work, even though he did not know the specific ColdFusion language and was simply told to read a book. He realized that many people would never discover computer science or gain access to those jobs. That experience became the basis for Treehouse, which aimed to teach coding at a fraction of the cost of a computer science degree. The product used connected, carefully produced video lessons so learners could follow a coherent path instead of searching through disconnected material online.
AI let Carson return to hands-on development and ship a product quickly
After Treehouse was acquired in 2021, Carson missed writing code. He paid for ChatGPT Plus, asked how to build a Python app using the OpenAI API, and then taught himself TypeScript, Next.js, and deployment with Vercel. He shipped an app using OpenAI APIs in about a month. That experience changed his view of education because the tools delivered much of what Treehouse had hoped to provide, at a much lower monthly cost. Carson still sees education as part of his identity, but he chose Intel because he believes compute will create intelligence and he wants to help developers build with AI.
Many AI products are experiments, and that is a normal part of starting a company
Carson agrees with Demetrios Brinkmann that parts of the AI market resemble the Field of Dreams idea of building something before knowing who needs it. He includes his own app, Maple, in that category. He says this uncertainty is normal in startups because founders must discover which problems people actually have. The cost and speed of shipping have changed, though. Carson built Maple to test whether he could create an LLC, write the code, release the product, maintain it, and market it himself. He sees this as evidence that a one-person company is now much more practical.
People with domain expertise can build useful AI products without becoming ML engineers
Carson says people do not all need to become machine learning engineers. Someone can combine knowledge from their own field with beginner development skills and existing infrastructure. He gives the example of a musician who understands the science of music and could use AI to solve problems in that area. Demetrios Brinkmann describes Feathery as an example of a focused product that helps users complete complex forms by transferring information from documents they have already provided. Carson's advice is to start from a problem the builder understands deeply, rather than adding AI to an area where they have no direct knowledge.
AI can help people reflect on conversations and maintain relationships
Carson is interested in devices that can listen to daily conversations, with permission, and use that context to provide useful coaching. He imagines asking a phone how he could have handled a difficult conversation with his children differently. He also wants AI to remind him to reconnect with people and recall what they discussed. Demetrios Brinkmann connects this idea to coaching in a particular style, such as asking for advice modeled on Dr. Becky. Carson accepts that the technology could become unsettling, but he chooses to explore how a personalized assistant might improve health, communication, and relationships.
Communities grow through repeated value and personal connection
Carson says building a community helped create the conditions for Treehouse. He started small meetups called By Designers for Designers, expanded them to other cities, then added a website, news, and conferences. These activities made him a connector who knew many people and could later ask for help, make introductions, or explain what he was building. His relationship with Kevin Rose began after Carson invited him to speak at a conference, which eventually led to Kevin investing in Treehouse. Demetrios Brinkmann adds that founders can also join existing communities, learn the language people use, and listen for problems before bringing a product to market.
Competition in AI compute can expand access and lower prices
Carson joined Intel because he expects compute to become one of the world's most valuable resources and believes only a small number of companies know how to make the needed semiconductors. He wants to help developers use AI while also giving them more choices among compute providers. He mentions Intel's Gaudi 3, an ASIC designed for AI workloads, and says it is faster than an H100. Carson's larger point is that companies, countries, cities, and smaller communities will need to train, fine-tune, and run inference on models. He argues that more competition is needed so access grows and prices fall.
Employees should connect their work to the company's highest-level goals
Carson's advice for people in startups and large companies is to find the top-level company OKR or set of OKRs, understand how it flows down to their role, and work on moving that number. He says people who spend time on work outside the company's main goals will struggle to show value or advance. If the company has no clear OKRs or KPIs, he suggests asking leadership to define the strategy and the result needed over the next 12 months. He also recommends turning that result into smaller outcomes and tasks. Building helpful relationships matters alongside this alignment, since people prefer working with colleagues who have integrity and offer solutions.
"I would say just grab your favorite model and start talking to it about what you're thinking, what you're afraid of, what you're not sure about."Ryan Carson35:06
Who should watch
You are building a small AI product and need a practical way to test whether it solves a real problem.
You have expertise in a non-technical field and want to use AI without becoming a machine learning specialist.
You work in a large company and need to connect your projects to its broader goals and OKRs.
