In partnership with

Turn AI into Your Income Engine

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

  • A curated collection of 200+ profitable opportunities spanning content creation, e-commerce, gaming, and emerging digital markets—each vetted for real-world potential

  • Step-by-step implementation guides designed for beginners, making AI accessible regardless of your technical background

  • Cutting-edge strategies aligned with current market trends, ensuring your ventures stay ahead of the curve

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get Your Guide

We track Product so you don't have to. Top Podcasts summarised, the latest AI tools, plus research and news in a 5 min digest.

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? 🥘

Anthropic had the kind of week that would give a comms team a collective aneurysm: Claude moved into Slack as a permanent coworker, the NSA lost access to the company's best model over a contract dispute, the president called them a national security threat then un-called them a national security threat after a G7 chat, and the product team shipped a major design overhaul as if nothing was happening. Meanwhile, Meta announced $299 AI smart glasses (the "make everyone a glassh*le again" price point), brought a fintech founder in to run WhatsApp, and Zuckerberg told his recently-culled workforce to start having fun - which is definitely how fun works. Midjourney went from making art to making medical ultrasound scanners, which is either genius or terrifying, and Cloudflare quietly shipped the protocol that might finally solve the bot traffic problem nobody has managed to fix for a decade.

Your Menu

📰 Not Boring -> Anthropic's seven-day saga, Meta's prediction markets and $299 glasses, Apple's supply-side squeeze, agent infrastructure maturing, and Midjourney's medical moonshot
⌚️ Productivity Tapas -> agent-native Git, free Crunchbase alternative, LaTeX visual editor, real-time site presence, and a pen-testing model that doesn't refuse
🍔 Blog Bites -> why showing teams the truth changes nothing, product vision without death by committee, and the B2B launch that's just one post and one email
🎙️ Pod Shots -> "It's Never The Model" - seven podcast conversations on why the bottleneck is waste, not capability

Let's go 🚀

📰 Not boring

🤖 Anthropic's Wild Week: Slack Agents, NSA Standoffs, and a Presidential U-Turn

One company managed to embed itself in enterprise Slack, lose a government intelligence contract, get accused of being a national security threat, have that accusation reversed by the president, and ship a major product update - all in seven days. For PMs building on third-party AI, the NSA story is the one to sit with: your model provider's relationship with a government can change your access overnight.

🏢 Big Tech Power Moves

Meta had the busiest week in big tech: prediction markets, $299 smart glasses, a new WhatsApp leader from fintech, an employee surveillance blunder, and a "have fun" edict post-layoffs. The Kunal Shah appointment is the strategic tell - a payments founder running the world's most popular messaging app points directly at commerce, not conversation.

🍎 Apple's Hardware and Partnership Bets

Apple's week was about supply-side constraints, not demand-side innovation. The memory chip crunch driving price rises, a politically brokered Intel partnership, and a design team that needs rebuilding all point to the same reality: even Apple is hitting physical limits on how fast it can ship its AI ambitions.

🧠 AI Products and Agent Infrastructure

The agent infrastructure story moved from "can agents do things" to "how do you maintain and scale them." Prompt debt as a concept arrived at exactly the right time - most teams that started with agents six months ago are now drowning in prompts nobody owns. Chrome scoring sites for agent navigability is a quiet but significant signal: the web is being rebuilt for non-human users whether publishers opt in or not.

🔬 Hardware, Chips, and the Physical Layer

The physical layer is where AI ambition meets reality. Tesla selling Megapod hardware, Amazon competing with Nvidia on chips, and a 23-minute investigation of why data centres literally cannot plug into the grid - these are the supply-side constraints that will determine which AI products actually ship at scale and which stay in demo mode.

📜 Regulation, Privacy, and the Bot Traffic Question

Cloudflare's privacy-first anti-bot protocol is the week's most consequential infrastructure story for anyone building a web product. Bot traffic is no longer an abstract concern - it's half the internet. The protocol creates a new standard that PMs building web products will need to support, and the smart TV proxy story shows just how creative the bot economy has become.

💡 Product Strategy and Founder Thinking

Midjourney going from image generation to medical ultrasound scanners is either a David Holz masterclass in platform thinking or a warning about what happens when AI companies chase markets they don't understand. Either way, it's the product strategy story of the week. PostHog's loops essay and the "you got faster, your company didn't" piece pair well - the organisations that capture AI productivity gains are the ones that redesign the process, not just the tools.

⌚ Productivity Tapas: Time-Saving Tools & Workflow Automation

  • StartupWiki: a free alternative to Crunchbase for researching startups, funding rounds, and founders. Useful for PMs doing competitive research or partnership scouting without a Crunchbase Pro subscription.

  • Oak: Git alternative designed specifically for AI agents. Handles the branching and merging patterns that emerge when multiple agents work on the same codebase simultaneously - solves a real pain point for teams running parallel agent workflows.

  • TikZ Editor: WYSIWYG editor for creating LaTeX figures. Turns the notoriously painful process of drawing technical diagrams in LaTeX into a visual drag-and-drop workflow. Niche but saves hours if you publish academic or technical content.

  • TownSquare: a tiny presence layer for websites that shows who else is currently viewing the same page. Adds real-time social context to any static site with minimal integration effort - useful for documentation, wikis, and community content.

  • Argus Red: a post-trained model specifically designed for penetration testing rather than refusing security-related prompts. CLI-first, built for security teams who need AI assistance without the safety guardrails that block legitimate security work.

    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.

🍔 Blog Bites - Essential Reads for Product Teams

Leadership: Why Showing Teams the Truth Rarely Changes Anything

John Cutler dissects a pattern every product leader has lived through: you surface the data, show the team exactly where they stand, and nothing changes. The problem is not the mirror - it is what people do when they see their reflection. Read the full article here.

💡 "The trouble with mirrors is that they don't come with instructions."

| | Key Takeaways

  • Mirrors provoke defence, not action : When teams see uncomfortable data about their own performance, the default response is to explain it away rather than act on it. The data itself is never enough to create change.

  • Context shapes the reaction : The same information presented in a safe, collaborative setting produces a different response than the same data dropped into a performance review. Framing determines whether a mirror becomes a tool or a weapon.

  • Change requires a next step, not just awareness : Showing the truth without offering a concrete, low-risk first action leaves teams stuck between knowing and doing. The gap between insight and behaviour is where most transformation efforts die.

  • Leaders own the environment, not just the information : The job is not to hold up the mirror; it is to create conditions where people can look at what they see and choose to move.

John Cutler, The Beautiful Mess

Why I picked this: Every PM who has ever built a dashboard, run a retrospective, or presented user research and watched nothing happen needs this piece. It names the failure mode precisely and offers a way past "but the data was right there."

Product Leadership: How to Build a Product Vision Without Death by Committee

Tim Herbig walks through a two-part case study on collaborative product vision development, tackling the tension between inclusive process and decisive outcome. The piece shows how one team built a vision that stuck by treating it as an emotional statement first and a deliverable second. Read the full article here.

💡 "A product vision that emerged from a committee process is almost never one anyone would fight for."

| | Key Takeaways

  • Vision is emotional, not rational : A vision that reads like a strategy document will be filed and forgotten. The ones that stick provoke a reaction - excitement, discomfort, ambition - and that requires a human author, not a consensus process.

  • Collaborative input, singular output : Gather perspectives broadly, then have one person write the vision. The collaborative part is the input; the craft is in the synthesis, and synthesis by committee produces mush.

  • Test with a "fight for" question : After drafting, ask the team: "Would you fight for this?" If the answer is polite agreement rather than genuine conviction, rewrite it until someone pushes back.

  • The vision earns its authority over time : A new vision has no credibility on day one. It earns credibility by being referenced in real decisions - and by the leader being willing to say no to things that contradict it.

Tim Herbig, Product Management Newsletter

Why I picked this: Vision-by-committee is the default failure mode at most companies, and it produces documents nobody references six weeks later. Herbig's "emotional statement, not a deliverable" reframe is the kind of practical correction that changes how a team actually works.

Growth: The B2B Launch That's Really Just One Post and One Email

The Growth Playbook breaks down how a handful of B2B founders launched products successfully using nothing more than a single LinkedIn post and a targeted email - no Product Hunt, no press cycle, no ads budget. The piece argues that founder-led content is the most underrated distribution channel in B2B. Read the full article here.

💡 "The best B2B launches look nothing like launches. They look like a founder sharing what they built and why."

| | Key Takeaways

  • One post, one email, done : The founders who converted best skipped the multi-channel launch playbook entirely. A single LinkedIn post that told the founding story and a targeted email to 50-100 people in their network outperformed elaborate campaigns.

  • Founder credibility is the distribution : In B2B, the founder's reputation and network are the channel. The content works because the person behind it has skin in the game and the audience knows it.

  • Authenticity beats polish : The posts that performed were honest, specific, and rough around the edges. Overly polished launch content triggered the same scepticism as a cold outreach email.

  • Repeat the format, not the message : The founders who sustained traction kept posting - not the same announcement, but ongoing behind-the-scenes updates that turned a launch moment into a narrative arc.

The Growth Playbook

Why I picked this: This cuts through the over-engineered launch playbook that most B2B teams default to. For any PM or founder preparing to ship something, the data here says your personal network and one honest post will outperform the 47-step launch checklist.

🎙 Pod Shots - Bitesized Podcast Summaries

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

🤖 It’s Never The Model

Everyone's chasing a smarter model. The teams actually pulling ahead are winning somewhere far less glamorous: utilisation, craft, cheap experiments and the discipline to copy what works. Across seven recent conversations - from the people behind Anthropic's compute strategy, Microsoft, Zynga, Clay, the iPhone and xAI - one uncomfortable pattern keeps surfacing: the bottleneck is almost never the capability. It's the waste sitting right next to it.

Key insights from the full article:

  • 🏗️ Unlimited GPUs still fail - Anjney Midha's diagnosis of why cash-and-compute-rich labs ship nothing: it's the culture, not the hardware. "Culture is not a set of beliefs. It's a set of actions."

  • ⚙️ You're probably wasting most of your compute - Google runs 95% node utilisation as standard; most clusters run far below their potential. The margin for error shrank and common sense became a competitive advantage.

  • 🐛 The gains hide in the boring layer - Ethan He's claim from building Grok Imagine: "A lot of the improvements do not come from new algorithms. It comes from finding small bugs in the data pipeline."

  • 🔎 Intelligence isn't knowledge - Will Bryk's contrarian line: we'll have super-intelligences still bottlenecked by bad search. A brilliant model can't surface a fact it can't find.

  • 🧭 Stop polishing one idea - test a hundred cheap ones - Mark Pincus: your instincts are right 95% of the time, your ideas right at best 25%. Use AI as a failure machine, not a one-idea builder.

  • 🧱 Your private evals are the new IP - Satya Nadella's control test: if you can't swap one model for another on your own eval and keep climbing, you don't control your stack.

  • 💸 AI-generated code is a debt, not a free lunch - Tony Fadell's warning: "short-term gain for very, very long-term loss." The waste you can't see yet is the most expensive kind.

  • ⚖️ Capitalism rewards risk, not hard work - Kareem Amin's founder philosophy: effort is table stakes; the reward is for the brave, non-obvious bet held with conviction.

EXPLORE THE FULL PODSHOT →

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 🍽️.

Reply

Avatar

or to participate

Keep Reading

View more
caret-right
© 2026 Product Tapas..
beehiivPowered by beehiiv