The SEOCover

The premise that you need a separate "AI search strategy" distinct from regular SEO is oversold. What actually matters is being findable and credible enough that AI systems training on the web pull your content into their reasoning—which mostly means doing the unsexy fundamentals better: depth, clarity, topical authority, and citations that matter. The real shift isn't creating new content tactics for AI; it's that every middling piece of content you've coasted on suddenly has to earn its existence again because an LLM will happily skip it for something better.

Read full take →

PostHog betting everything on "self-driving product" is a bold rebrand that papers over a harder truth: they're still selling product analytics infrastructure, just dressed in AI language that happens to move product right now. The shift matters tactically—it's a cleaner pitch than "we do session replay and feature flags and SQL queries"—but it doesn't actually change what the platform does or who buys it, and that messaging risk runs real if the AI hype cools and they're left explaining why you need them when Claude can't actually drive your product decisions. For practitioners, this is a reminder that repositioning plays best when it honestly reflects what changed about the product itself, not just what changed about investor appetite.

Read full take →