There Are Only 5 Safe Places to Build in AI Right Now. Are You in One?

The Future of Web App Builders

The Shift in AI App Builder Landscape

• Many AI app builders, such as Lovable and Replet, are pivoting towards open-source models to stay relevant in a rapidly changing market.

• There is a critical question regarding the existence of safe spaces for building web applications that won't be easily disrupted by major players like OpenAI or Anthropic.

• This discussion centers on five aspects of web development that AI cannot replace, which may define future niches within the industry.

Understanding the Collapse of the Build Layer

• The current competition among app builder companies reveals a broader picture than what they publicly acknowledge; they are primarily focused on growth and valuation.

• A significant trend is emerging where numerous companies are racing to create platforms that can automatically generate applications from user descriptions.

• Lovable has achieved remarkable success with over 100,000 new projects created daily and substantial recurring revenue, highlighting its dominance in the space.

Competition Among Smaller Players

• Compared to Lovable's scale, other players like Verscel and Replet remain relatively small but are still competing for similar market share by offering AI-driven application development.

• The introduction of OpenClaw has shifted expectations from merely building apps based on ideas to creating entire businesses using advanced capabilities.

Challenges in Differentiation

• Most companies in this space offer minimal differentiation beyond basic functionalities due to reliance on common foundational models like Claude or ChatGPT.

• As these UI layers become easier to replicate, companies face challenges maintaining unique value propositions when their offerings depend heavily on external intelligence.

Strategies for Survival

• Conventional wisdom suggests that training proprietary models could help escape competitive pressures; however, this isn't always sufficient for long-term survival.

• Companies like Replet focus on owning their runtime environments rather than just competing through model performance. This ownership provides a distinct advantage over competitors relying solely on API calls.

Understanding the Role of Infrastructure in AI Development

The Unique Position of Notion and Other Companies

• Companies like OpenAI, Anthropic, Nike, and PayPal have established deployment infrastructures that support their production applications.

• Notion is building on top of existing AI models without training its own, offering users a choice between models like ChatGPT or Claude.

• The key insight is that as AI commoditizes production, companies must focus on layers of value that cannot be easily replaced by AI.

Five Durable Verticals of Value

1. Trust

• In an increasingly crowded web filled with indistinguishable AI-generated content, trust becomes essential for distinguishing legitimate services from scams.

• Companies providing verification layers will capture significant value by ensuring safety and accountability in transactions.

• Examples include Stripe's strong position due to processing over a trillion dollars in transactions, which enhances its trust signal.

2. Context

• Contextual data—specific to individual situations—is more valuable than general compute power or prompting ability.

• Companies that manage context effectively will control critical choke points on the internet as every agent needs access to this unique data.

• Notion has successfully built custom agents leveraging user-specific context, demonstrating how context can enhance the utility of AI tools.

Implications for Future Business Models

Importance of Contextual Agents

• As work increasingly relies on AI agents, those equipped with contextual information will gain competitive advantages over generic chatbots.

• The structural play around context is similar to successful models seen in Salesforce and other durable companies across various sectors.

This structured approach highlights the evolving landscape where trust and context are pivotal for businesses navigating an AI-driven future.

Understanding the Role of Context in AI

The Importance of Context in AI Development

• A dependable junior employee can significantly impact outcomes, emphasizing the importance of correct prompting in AI. Proper context and direction are crucial for effective AI performance.

• Various companies like Notion, Salesforce, Epic, Palantir, Snowflake, and Data Bricks are emerging players in the context space. Even tech giants like Apple and Google are exploring local AI applications to enhance their offerings.

Google's Multifaceted Approach to AI

• Google is not just a model player but also a foundation player with TPUs (Tensor Processing Units), a context player, an ecosystem player, and a devices player. This diverse strategy positions them advantageously in the competitive landscape.

The Challenge of Distribution in Software

Distribution as a Key Bottleneck

• First-time founders often overlook distribution challenges; building software is only part of the equation. Effective distribution is essential for success.

• In an era where digital products multiply exponentially, having an edge in distribution becomes critical for visibility and customer engagement.

The Role of Gatekeepers

• Major platforms like Google, Apple's App Store, TikTok, and YouTube act as distribution monopolies that gain power as more content floods the market. They dictate where users go for information or services.

Emerging Trends in Agent Discovery

The Need for New Distribution Mechanisms

• As businesses adopt AI agents extensively, discovering how these agents interact with each other and humans presents significant challenges.

• Content creators can benefit from establishing authority within niche areas of AI since they will help others navigate useful information amidst overwhelming data.

Future Considerations for Agent-Friendly Businesses

• There’s potential for developing an "agent-native app store" to facilitate agent discovery of utilities or businesses that cater specifically to their needs.

Rethinking Commerce with Agents at Core

Viability Factors for Agent Transactions

• Understanding what makes a business viable for agent transactions involves considering transaction speed, clarity of offerings, ease of selection by agents, and simplicity in receiving goods/services.

Current Business Landscape Insights

• Few businesses currently consider these factors; notable players include Google (search/discovery), Apple/Google (app stores), TikTok/YouTube/Substack (content), and Amazon (commerce).

The Concept of Taste in Software Production

Defining Taste Beyond Personal Quality

• Taste should be viewed as a separate vertical on the web rather than merely a personal quality. It encompasses product decisions and design sensibilities that shape what gets built.

The Impact of Free Software Production

• With free software production becoming commonplace, choices about what to produce become paramount—this includes accountability over generated content's accuracy.

Analogy with Music Production

• Just as music production tools have democratized music creation leading to saturation but rewarding those with unique taste or vision—similar dynamics apply to software development today.

Understanding the Future of Software Production

The Role of Taste in Software Development

• Production is becoming increasingly accessible, with creators possessing a unique taste and understanding of audience connection.

• Successful software must resonate with its audience through strong design and a clear value proposition that addresses felt needs.

Balancing Design and Value Proposition

• While prioritizing one aspect, it’s often recommended to focus on the value proposition for product success.

• The best products combine both strong design sense and a compelling value proposition, acting as dual engines for achieving product-market fit.

Human Oversight in Agentic Systems

• In agentic systems, success relies on human expertise to fine-tune prompts and workflows rather than just superior underlying models.

• Humans remain accountable for guiding agents' actions within the economy, emphasizing the importance of oversight despite advancements in automation.

Liability Concerns in AI Applications

• As AI becomes more integrated into various sectors, liability issues arise regarding accountability when AI-generated outputs lead to negative outcomes.

• Professionals in regulated industries (healthcare, finance, legal) are particularly focused on managing liability as part of their service offerings.

Governance Layer for AI Agents

• Effective management of liability will become essential as AI agents autonomously execute complex tasks; defining boundaries is crucial.

• Companies specializing in liability management will play a significant role in shaping governance structures around AI technologies.

The Future Landscape of Web Intelligence

• Major model providers like OpenAI and Anthropic are positioned to dominate foundational intelligence capabilities for future web applications.

• The influence of these core model makers will persist even with the emergence of open-source derivatives based on their technologies.

The Future of Platforms and AI

The Role of Companies in the Evolving Landscape

• Many rapper companies, such as Lovable, Bolt, and Shipper, lack durable assets. While some may get acquired, most are likely to fail. Only a few that gather sufficient user data and momentum will have a chance to evolve into platforms.

• Lovable aims to become "Shopify 2.0," positioning itself as a key player in future business infrastructure alongside companies like Vercel, Replit, Stripe, and Shopify that own essential trust and execution layers.

Trust and Context in the AI Economy

• Context owners like Notion, Salesforce, Snowflake, and Databricks control data gravity. They provide necessary context for agents to function effectively within the agentic economy if they manage their roles correctly.

• Distribution gatekeepers such as Google, Amazon, and Apple influence how attention is directed online. Their strategic moves could determine how agents interact with information on the internet over the next few years.

Human Contribution Amidst Automation

• Humans—founders, professionals—play an essential role by providing judgment and accountability. Understanding web structure allows us to guide AI agents strategically rather than being replaced by them.

Building for the Future: Key Considerations

• Entrepreneurs should evaluate what aspects of their business remain relevant even if AI improves significantly. If nothing holds value against better models, it's crucial to pivot now.

• Conversely, if advancements enhance your product's value (e.g., owning parts of trust or liability layers), you are positioned well for growth without fear of obsolescence from new models.

Importance of Distribution in Product Development

• As we focus on productionalizing code for MVPs (Minimum Viable Products), distribution must not be overlooked. Engaging customers early is vital for validating product-market fit—a fundamentally human task.

• Despite tools facilitating creation (like Lovable and Replit), many apps go unnoticed due to insufficient thought given to customer needs during development—a lesson that remains critical today.

Lasting Principles in an AI-Dominated World

• Core principles such as trust, context, distribution, liability, and taste have always been significant on the web but are becoming increasingly vital due to AI's influence; these elements cannot be replaced by technology alone.