From: aidotengineer
Munir, CEO and co-founder of Porsche AI, shares insights from his journey as a recent AI startup founder, drawing comparisons to his previous roles in product at big tech companies like Stripe, Google, and Amazon [00:00:03]. Porsche AI develops an open-source SDK for building production agents, with a particular focus on regulated industries [01:38:00].
User Problems and Discovery in AI
The landscape of user problems and discovery in AI is significantly different from traditional product development [02:11:00].
- Emergent Problems Unlike established product areas where user problems have a defined shape and adjacency to known issues (e.g., payment products at Stripe) [02:31:00], AI presents a “zero to one” scenario [02:57:00]. Users often do not precisely know their problems; instead, the problems are an emergent property of the AI space itself [03:25:00]. It’s akin to playing a “Soulslike game” for the first time, where nothing initially makes sense regarding how to combine elements or get to a desired outcome [03:00:00].
- Rapidly Changing Technology The capabilities and possibilities offered by AI technology change very quickly [03:41:00]. This means conventional three-to-six-month roadmaps are less effective. Instead, rapid iteration through putting products out there and observing user reaction is crucial [03:45:00]. The guiding principle becomes hypothesis-driven, deliberate iteration [04:06:00].
- Creating the Narrative Traditional tools like user storyboards and critical user journeys are largely diminished in their utility [04:24:00]. It’s more like starting the game No Man’s Sky, where there’s no clear story arc or mission [04:41:00]. Developers must actively help users overcome initial inertia to adopt AI for their specific use cases, collaborating to create narratives and anchor on a use case [04:54:00]. Once anchored, the process gradually shifts to more familiar territory of refining the product against that specific use case and its challenges [05:17:00].
Product Development Gratification and Velocity
The impact of AI on the development workflow can be highly gratifying due to the rapid product development cycle [06:01:00].
- Fast Iteration from Idea to Code Unlike big tech where product design, engineering design, and API design require numerous sign-offs across many departments, in an AI startup, the transition from an idea to a spec, design, testing, and release can occur within a few hours or days [06:06:00]. This direct transformation from concept to working code is exhilarating [06:47:00].
- Velocity as “Table Stakes” In AI, velocity is not merely a “nice to have”; it is a fundamental requirement for success. The ability to release quickly and effectively is essential [06:52:00]. Opportunities that provide significant competitive advantages (“power-ups”) emerge and disappear rapidly, such as the adoption of specific protocols (e.g., MCP, agent-to-agent protocol) or breakthroughs in models (e.g., diffusion models) [07:07:00]. Staying at the forefront of these trends is crucial for brand growth and establishing presence in the developer or user community [07:31:00]. This highlights prototyping and production in AI as a critical capability.
Outreach, Awareness, Traffic, and Adoption
One of the toughest and most underestimated challenges in starting an AI company is building scaffolding for outreach, awareness, traffic, and adoption from scratch [08:08:00].
- Navigating Hype and Noise The AI space is currently characterized by intense hype and “meme wars,” making it difficult to distinguish signal from noise [08:25:00]. Without the established brand advantages of a larger company, it feels like playing Crash Bandicoot without boosters or Mario Kart without power-ups [08:32:00].
- Leveraging People and Allies A key learning is that people primarily follow other people [09:04:00]. Unlike launching a product at a big company where a tweet from a leader can generate instant awareness, startups must find credible industry advocates to speak for them and act as force multipliers [09:09:00]. For broader brand awareness and traffic, finding allies is essential [09:33:00]. This includes partnering with startups at a similar stage or those slightly ahead who see synergy [09:51:00]. Strategies include cross-marketing, cross-selling, and getting products integrated into the documentation of other companies (e.g., BrowserBase) to drive traffic to GitHub repositories and websites [09:58:00]. This constant effort, though challenging, offers significant learning opportunities [10:24:00].
Munir emphasizes that he is still learning and hopes his shared experiences act as inspiration for others considering the leap into founding an AI company [11:18:00].