Key Takeaways
- Base44's $80M acquisition by Wix in 8 months highlights rapid AI startup potential.
- "Vibe coding" is predicted to disrupt traditional SaaS by enabling highly customizable, user-owned software.
- AI coding platform defensibility requires vertical integration and complex ecosystem development.
- AI model costs are trending to zero, shifting focus to user adoption and distribution.
- Investing in AI for non-software sectors like finance or healthcare offers significant opportunity.
Deep Dive
- Maor Shlomo's Base44 was acquired by Wix for $80 million, scaling from idea to millions in revenue within 18 months.
- The acquisition enabled Base44's product team to remain lean while leveraging Wix's marketing and operational resources.
- Post-acquisition, Base44's revenue has grown to over $100 million.
- Shlomo initially hesitated but was persuaded by Wix's CEO, Avishai Abrahami, to join, making it a "perfect match" for growth.
- "Vibe coding" is projected to become the largest software category, integrating functions like CRM and project management.
- This shift envisions liquid software with templates and open-source projects, customized by users with ownership of code and data.
- The guest predicts that the pre-2022 "bloated SaaS" model will struggle, with smaller CRM companies being particularly vulnerable due to a lack of strong moats.
- The future size of Salesforce in 10 years is questioned by the guest.
- Base44 focuses on allowing non-technical users to build software via prompting, avoiding direct code interaction, differentiating from developer-focused tools like Cursor and Replit.
- The guest expresses less concern about competitors like Replit and Lovable, and greater worry about major model providers like Google dominating the market.
- Google is considered the most likely to dominate AI coding due to its compute power, rapid progress with Gemini, and existing infrastructure including Google Cloud and data access.
- The speed of copying features in AI coding is low, indicating that the technical moat has been eroded.
- The market for AI models is dynamic, with companies rapidly shifting spending between providers like Anthropic (Claude), Google (Gemini), and OpenAI based on performance and cost.
- The ease of switching between LLM providers, in contrast to cloud services, creates an "insane market dynamic" with volatile spending patterns.
- AI coding agents, while offering autonomy, carry risks such as significant costs incurred from incorrect user prompts for free users.
- User success for AI apps is measured by direct usage, sentiment in chat interactions, and improvement of specific metrics like prompt sentiment.
- While margins are important for a healthy business, rapid growth and market capture are primary goals in the current AI market, with model prices decreasing.
- Founders must account for drastic and rapid changes in model-related spending, which constitute the majority of costs and lead to uncertain revenue streams.
- Key metrics for AI apps emphasize rapid revenue growth, such as reaching $10-20 million ARR quickly, rather than absolute ARR.
- True defensibility in AI lies in vertically integrated businesses, such as infrastructure providers or non-software entities like law firms and hospitals optimized with AI.
- The low cost and speed of copying technical advancements mean businesses solely focused on LLM prompting may struggle to build a lasting moat.
- The guest identifies investment opportunities in ambitious, near-impossible goals within non-software sectors like finance, restaurants, hospitals, and banks.
- He advises caution against investing in AI agent companies or startups with rapid revenue growth unless the investor has deep industry understanding.
- The guest believes the software building market is not a winner-takes-all scenario, favoring multiple successful companies like Cognition over Cursor.
- AI development is expected to benefit the entire market, not just model companies, preventing a significant AI bubble.
- Base44's successful $80 million acquisition highlights that vast economic value can still be created with existing AI models, leading to more efficient companies and better products.
- The LLM era enables leaner company building; Base44 was built by one person, contrasting with capital-intensive predecessors.
- Base44 benefits from operating outside Silicon Valley, experiencing reduced talent competition and support from Wix acting as a proxy for Silicon Valley resources.
- By 2026, the guest predicts 95-100% of code will be AI-written, with humans prompting agents, a reality Base44 already approaches.
- For non-technical individuals, the primary advice for building with AI tools is to address personal problems to ensure usability and solve genuine pain points.
- This approach emphasizes an iterative and experimental mindset, encouraging users not to fear discarding code and starting over.
- The guest prefers Anthropic over OpenAI, citing its faster progress in coding, focus on B2B value, superior models, and stronger growth trajectory despite OpenAI's first-mover advantage.
- He would choose Jack Dorsey as a board member for his consistent builder mindset demonstrated across multiple successful ventures like Twitter and Square, aligning with Base44's ethos.
- Reflecting on past ventures, the guest regrets missing significant personal time with family and friends due to inexperience and overworking.
- Amid increasing AI competition and eroding technical moats, the advice is to focus on one's own business and select partners with a shared mindset and complementary activities.