A new wave of startups, like ex-Twitter CEO's Parallel, is attracting significant investment to build web infrastructure specifically for AI agents. Instead of ranking links for humans, these systems deliver optimized data directly to AI models, signaling a fundamental shift in how the internet will be structured and consumed.

The internet was designed for human interaction, actively discouraging bots. The next evolution will reverse this, with AI agents becoming the primary users. This requires re-architecting everything from user interfaces to business models, with crypto likely serving as the native payment rail for these autonomous agents.

The frontier of AI competition is moving beyond raw model performance (e.g., Opus vs. GPT). The new battleground is the ecosystem of agentic 'harnesses'—specialized tools, workflows, and infrastructure—built around models. Anthropic's developer day focused entirely on these applications, signaling a major shift in where value is created.

AI labs like OpenAI and Anthropic are developing "super apps" that go beyond chat to take over your browser and computer for tasks like hiring or booking. This agentic model, where the AI acts on your behalf, could fundamentally shift power away from individual websites.

The next frontier of model development will come from AI agent companies. These firms, which currently build on top of existing models, have a strong incentive to create their own specialized models to power and distribute through their agents, opening a new competitive front.

The internet was built for human interaction. The rise of autonomous agents shopping for products and services on our behalf signals the dawn of an 'agentic web.' This will force a fundamental shift in marketing and sales, requiring businesses to learn how to effectively market to and be discovered by AI agents, not just humans.

As AI model performance commoditizes, the strategic battleground is shifting from models to platforms. Tech giants like Google are positioning their offerings not as features, but as the fundamental 'operating system' for the agentic enterprise. The new competitive moat is the control plane that orchestrates agents.

Obsessing over linear model benchmarks is becoming obsolete, akin to comparing dial-up speeds. The real value and locus of competition is moving to the "agentic layer." Future performance will be measured by the ability to orchestrate tools, memory, and sub-agents to create complex outcomes, not just generate high-quality token responses.

By shelving consumer-facing "side quests" like video generation, OpenAI's strategy now directly mirrors Anthropic's. This transforms the AI race from a consumer vs. enterprise competition into a direct fight to build the dominant "agentic" AI that can control devices and execute complex tasks for users.