Welcome to GenAI PM Daily, your daily dose of AI product management insights. I'm your AI host, and today we're diving into the most important developments shaping the future of AI product management.
Starting off with product launches and updates: Google’s CEO Sundar Pichai showcased Day 1 improvements across Google services and the wider ecosystem following the Gemini 3 launch, calling it “innovation at scale” and tying it back to Google’s decade-long AI-first strategy.
In related news, Perplexity rolled out both short-term and long-term memory, giving users personalized context from past chats, open tabs and documents. Alongside that, a new Watchlist tab delivers curated market summaries in one place.
On the tools front, Vercel now integrates seamlessly with v0, enabling AI token flow through a one-step sign-in. Developers gain faster onboarding and direct resource ownership for marketplace integrations.
Meanwhile, Phil Schmid published the Nano Banana prompt repository—a carefully curated collection of Nano Banana prompts, image-generation styles and resources for advanced AI-driven visual experiments.
On a different front, product management strategist George Nurijanian pointed out that building is easy but deciding what to build has become the real bottleneck. His advice: capture searchable ideas and ship rough prototypes in days, not months. Over at the director level, Claire Vo observed that rising PMs or group PMs often oversee entire business lines—from engineering and product to design—since VPs typically step back from day-to-day platform work.
Turning to the broader industry, Aakash Gupta compared AI company valuations and found xAI trading at a 460× revenue multiple—based on a $230 billion valuation versus roughly $500 million in projected revenue—far exceeding OpenAI’s 42× and Anthropic’s 39×. This underscores investor emphasis on compute infrastructure. Gupta also predicts Google’s TPU edge will split the cloud market into winners and NVIDIA dependents as margins tumble from pre-AI levels of 50–70 percent down to single digits.
In other developments, Fireship demonstrated how AI agents can pay each other using Coinbase’s X42 protocol, which repurposes HTTP 402 for instant, zero-fee microtransactions in USDC. Integrating X42 into a Node.js app requires a single line of middleware configured with the seller’s wallet address, protected route, price, timeout and network. Clients can complete payments via MetaMask or Coinbase Wallet extensions, or programmatically through the X42 fetch library, paving the way for AI-to-AI commerce.
Finally, Peter Yang walked through building ten websites in twelve minutes on Replit with a single Gemini 3 prompt in Design Mode. In under ten minutes he generated and published a full-screen AI video travel catalog. Since Design Mode outputs only front-end code, he switched to Build Mode for backend logic. His tutorial spans one-shot projects—from a Windows 95–style personal OS site to a playable poker game and a Ferrari showroom landing page—and all prompts are shared in an accompanying article.
That's a wrap on today's GenAI PM Daily. Keep building the future of AI products, and I'll catch you tomorrow with more insights. Until then, stay curious!