Hey Product Fans!

Welcome to this week’s 🌮 Product Tapas.

New here? We're your shortcut to staying sharp. Essential stories, practical tools, real insights.

For the best reading experience, check our web or app version and sign up for future editions here.

What’s cooking this week? 🥘

The coding agent wars officially spilled into the rest of the office this week - Anthropic shipped Claude Cowork to mobile and web, and the usage data is telling: a third of all Cowork tasks are business process automation, not code. Meanwhile Figma absorbed a Y Combinator vibe-coding team (making their design-to-deployment ambitions extremely clear), Meta dropped v2 of a model that literally reads brain imagery, and X quietly launched an MCP server that makes the platform accessible to every AI tool using the protocol. Oh, and Hopper just paid $35M to the FTC for hidden fees, which should send a chill through every PM running drip pricing on a marketplace.

Let's go 🚀

The battle for AI distribution has shifted from "best model" to "most surfaces." Anthropic's Cowork usage data shows agents are eating business operations, not just code - and the companies building agent entry points at the keyboard, browser, and platform API level are the ones positioning to own daily workflows.

Meta is running three parallel experiments - neuroscience, AI-generated gaming, and an expanding app portfolio - and each one is a different bet on how people will interact with software in five years. The Pocket launch is the PM-relevant one: it's Roblox meets ChatGPT, and the creation-to-sharing loop is tight.

Figma's acquisition spree makes their direction clear: the design tool wants to eat prototyping, code generation, and deployment. For PMs, the practical question is whether your design system is structured for AI consumption - if it's not, the design-to-code pipeline breaks at the handoff.

Apple and Google are both treating AI assistants as primary OS surfaces now, not bolt-on features. Discord's moderation failure and WhatsApp's username impersonation problems show the other side of rapid AI deployment - the trust layer hasn't caught up with the capability layer.

The operational side of AI products is maturing fast. DoorDash's simulation platform is a good signal: the companies shipping AI at scale are building testing infrastructure, not just prompts. If your team is still evaluating AI products by model benchmark scores, you're measuring the wrong thing.

The a16z, MIT, and JPMorgan reading lists rarely overlap - that's the point. Cross-referencing them tells you which themes are consensus (AI infrastructure, energy) versus which are domain-specific bets.

The TV Time shutdown is a cautionary tale for PMs: pivoting to AI by killing your existing product only works if your data asset transfers. Riverside's move is smarter - extending the creator ecosystem into adjacent formats without burning what already works.

Hopper's $35M settlement is the kind of enforcement action that changes product roadmaps overnight. If your pricing has any dark-pattern adjacent mechanics - drip pricing, hidden fees, urgency timers on non-scarce inventory - this is your compliance wake-up call.

Over 3.5 billion people open WhatsApp, Instagram, or Facebook Messenger every day. Your customers are already there, asking questions and comparing options.

Wati puts your business across WhatsApp, Instagram DM, Facebook Messenger, SMS, RCS, and web chat in one AI-powered inbox.

Automations instantly respond, route conversations, and keep every interaction tracked in one place.

Meet customers where they already are, before your competitor does.

See All Wati Channels

Remember. Product Tapas subscribers get our complete toolkit - 550+ personally tailored, time-saving tools for PMs and founders. Your shortcut to efficiency and what's hot in product management

Check the link here to access.

John Cutler explores why dashboards, metrics, and transparency initiatives consistently fail to drive the behavioural changes they're designed for. The problem isn't information access - it's that showing people a mirror doesn't give them a reason to change what they see. Read the full article here.

The real blocker to change is rarely awareness - teams usually already know what the data shows. What they lack is the structural permission to act on it.

Crown & Reach reframes the classic Xerox PARC narrative: the failure wasn't a lack of business acumen, it was poor navigation of uncertainty - treating research breakthroughs as finished products rather than staged bets. Read the full article here.

The Growth Playbook dissects the structural patterns that separate founders who build audience on LinkedIn from those who generate noise - positioning, content sequencing, and the counterintuitive role of vulnerability in B2B credibility. Read the full article here.

Remember, we've built an ever-growing library of our top podcast summaries (120 or so). Whether you need a quick refresher, want to preview an episode, or need to get up to speed fast - we've got you covered.

Check it out here

Every playbook tells you to de-risk: run the test, read the data, hedge the bet, ship the MVP. This week, seven of the best conversations in tech - a chip founder, the man who built the iPod, the creator of FarmVille, Uber's CEO, Anthropic's own PM, the founder of Clay, and the ghost of Walt Disney - all landed on the opposite instruction: concentrate, commit, and refuse to hedge. The scarce input was never the analysis. It was the nerve to hold a non-consensus bet long enough for it to pay off.

Featured in this round-up:

Estimated reading time: 8 mins. Time saved: 11+ hours!

Key insights from the full round-up:

That’s a wrap.

As always, the journey doesn't end here!

Please share and let us know what you would like to see more or less of so we can continue to improve your Product Tapas. 🚀👋

Alastair 🍽️.