In partnership with

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? 🥘 

AI's having an identity crisis: it's simultaneously going to destroy everything AND you should treat it like a boring tool (classic tech contradiction). Google's playing monopoly games the moment judges let them keep Chrome (what timing), while media companies watch their traffic evaporate faster than a crypto crash (FT down 30%, Daily Mail down 90%).

• 📰 Not Boring → AI reality checks, workflow land grabs, creator economy enshitification, media apocalypse economics

• ⌚️ Productivity Tapas → Parallel problem-solving agents, unified LLM APIs, pixel perfect product prototyping tools

• 🍔 Blog Bites → Voice-first productivity transformation, AI agent adoption gaps, corporate process vs efficiency trade-offs

• 🎙️ Pod Shots → Google's ultimate innovation factory strategy and the Gmail revolution that changed the web forever

Let's go 🚀

Training cutting edge AI? Unlock the data advantage today.

If you’re building or fine-tuning generative AI models, this guide is your shortcut to smarter AI model training. Learn how Shutterstock’s multimodal datasets—grounded in measurable user behavior—can help you reduce legal risk, boost creative diversity, and improve model reliability.

Inside, you’ll uncover why scraped data and aesthetic proxies often fall short—and how to use clustering methods and semantic evaluation to refine your dataset and your outputs. Designed for AI leaders, product teams, and ML engineers, this guide walks through how to identify refinement-worthy data, align with generative preferences, and validate progress with confidence.

Whether you're optimizing alignment, output quality, or time-to-value, this playbook gives you a data advantage. Download the guide and train your models with data built for performance.

Download the guide

📰 Not boring

Crikey it's been a busy old week in news, loads to get through

The Great AI Reality Check

• Anthropic co-founders warn AI could wipe out 50% of office jobs within 1-5 years, and that there's a 25% chance that it goes really, really bad. Brilliant.

• In this week's 'this changes everything' AI is turning traditional corporate org charts upside down.

• This is an excellent post by Sanya Ojha from Bain, talking about the challenges of implementing AI at scale. TL;DR, start with tangible use-cases and things that need problem solving. Avoid fancy non-scalable shiny things.

The Future of AI, LLMs, and Observability on Google Cloud: For tech leaders, effectively leveraging LLMs means starting with simpler use cases like summarisation, prioritising robust observability, and using advanced monitoring to balance quality, cost, and performance.

There's a delicious contradiction here. AI is simultaneously disrupting everything, solving the world's problems, and turning org charts upside down—yet the most practical advice remains refreshingly mundane: treat AI like a tool, not magic. What's fascinating about Anthropic's brutal honesty isn't the prediction itself, but that they're saying it out loud while others whisper it in private. The enterprise deployment wisdom is consistent across sources: start small, measure everything, focus on real problems rather than shiny demos.

The Workflow Wars Heat Up

• Perplexity pro users can now connect their email, calendar, Notion, and GitHub to Perplexity. Enterprise Pro users can also connect Linear and Outlook. Workflow automation updates keep coming.

• Perplexity also launched an AI email assistant for Max subscribers.

• Notion releases a load of AI agents to help you build pages and databases, analyse feedback, and draft emails. But in an uncanny resemblance to last week's OpenAI MCP data leak news, one user immediately got Notion to leak all of your private Notion pages. "Move fast and break things"

This is the productivity stack land grab in action—whoever owns your data connections owns your workflow. Perplexity's aggressive expansion makes sense when you realise they're not just building a search engine, they're building the nervous system for knowledge work. Too bad Notion's "move fast and break things" approach hits different when the things being broken are your users' confidential data.

The Creator Economy Gets further Ai-fied (should that be enshitified?)

• Google Puts Its Popular AI video generator Veo3 into YouTube Shorts and YouTube gets a host of AI tools help creators give viewers what they want* (*debatable I am sure)

• Google Research announces Learn Your Way, an AI-powered experiment that reimagines textbooks into personalised, multimodal learning experiences; raising retention test scores by 11%

That asterisk is doing heavy lifting - because what viewers want and what's good for the creator economy aren't necessarily the same thing. When AI can generate infinite content, the question isn't whether it can, but whether it should. The 11% retention boost in Google's learning experiment is promising, but education and entertainment operate by different rules.

The Media Apocalypse Accelerates

• Media appears to be in an existential crisis as Google's shift to AI has completely upended the online news model (Search referral traffic to the FT is down 25-30%, Daily Mail c-90%)

• Reddit wants to get paid more by the AI model builders for access to all us handy humans on the platform

The numbers are stark and the economics brutal. When AI Overviews answer questions without requiring clicks, the fundamental business model of digital journalism collapses. Publishers are calling it an "existential crisis," and for once, the hyperbole might be justified. Reddit's play is smart - if your content is training the models destroying traditional media, at least get paid properly for it.

Hardware Experiments and Wild Bets

• OpenAI might be developing a smart speaker, glasses, voice recorder, and a pin 2026-2027 release. Back to today, OpenAI Plus, Pro, and Business users can set the time GPT-5 thinks. Plus OpenAI announces new guidelines on teen safety

• Meta releases new wearables device access toolkit to let external developers build apps for its new smart glasses

• Facebook is getting an AI dating assistant

Apparently the lesson from Humane's $700 AI pin disaster was "we need more AI pins." The hardware race is heating up, but it's unclear whether any of these form factors will stick. An AI dating assistant from Facebook feels like the logical endpoint of our digital dystopia.

Odds and Ends

• Google adds Gemini to Chrome for all users in push to bolster AI search. Literally 1 week or less after a judge rules Google doesn't need to sell Chrome. What a coincidence…!

• Tether CEO [stablecoin if you didn't know] confirms major capital raise at a reported $500 billion valuation. That's roughly on par with OpenAI. In related news x402 is an open protocol for internet-native payments that enables users to pay for resources via API without registration or fees.

• This github has a library of 90+ prompts for Nano Banana

• Former NotebookLM devs' new app, Huxe, taps audio to help you with news and research

Nothing says "we're not a monopoly" quite like Google immediately integrating AI across Chrome the moment you're legally allowed to keep it. Meanwhile, a stablecoin company (Tether) valued like the most important AI company in the world tells you everything about how big (and profitable) stablecoins are going to be. The NotebookLM team spinning out to build Huxe makes sense - audio-first research tools feel like the natural evolution of how we consume information. Sometimes the most interesting developments happen in the margins.

⌚️ Productivity Tapas: Time-Saving Tools & GPTs

• Roma: build your own parallel problem-solving agents

• Alloy: Prototype with your real product. Capture your product from the browser and make changes with chat

• Sudo: unified API for LLMs - quick and cost-effective way to access OpenAI, Anthropic, Gemini, etc. One endpoint for lower latency, higher throughput, and lower costs than many alternatives

  Remember. Product Tapas subscribers get our complete toolkit - 460+ 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

Productivity: Voice-First AI Removes the Invisible Friction Between Brain and Page

Katie Parrott explores how switching from typing to voice dictation with AI has fundamentally transformed her writing process and work patterns. I am 100% a fan of this - this is the single most productivity improvement I’ve found in the last 2 years. I even wrote a post about it a few months back here.

Clearly she’s more articulate than I and demonstrates how removing the micro-frictions of keyboard input allows ideas to flow more naturally whilst creating new challenges around work-life boundaries. Read the full article here.

💡 "When I type, I'm always self-editing—backspacing, rephrasing, or policing awkward syntax. When I talk, the ideas tumble out in real time. It's less linear, but that turns out to be great for brainstorming."

BINGO: typing creates cognitive overhead that we don't notice until it's removed, freeing mental bandwidth for higher-order thinking.

Product: Why Your AI Agent's Superpowers Don't Guarantee User Love

Product Curious recently explored the critical gap between AI agent capabilities and actual user adoption, focusing on the architectural decisions that make or break product success. He breaks down the essential components that product managers need to understand to build AI agents that users actually want to use. Read the full article here.

💡 "The most sophisticated AI agent in the world is worthless if users can't figure out how to interact with it or don't trust its outputs."

This quote perfectly captures why so many AI products fail despite impressive technical capabilities - the human experience matters more than the underlying intelligence.

Strategy: Why Software Companies Choose Process Over Efficiency

Sean Goedecke has an interesting take that large tech companies deliberately sacrifice engineering efficiency for organisational control, drawing parallels with James C. Scott's theories on state governance. He reveals why companies maintain cumbersome processes despite knowing they slow down development, and how engineers navigate both formal and informal systems to get work done. Read the full article here.

💡 "The processes that slow engineers down are the same processes that make their work legible to the rest of the company. And that legibility (in dollar terms) is more valuable than being able to produce software more efficiently."

🎙️ Pod Shots - Bitesized Podcast Summaries

Remember, we've built an ever-growing library of our top podcast summaries. 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

🎯 How Google Built the Ultimate Innovation Factory While Wall Street Watched in Horror

It's been a little while since we covered some of the stuff from Acquired but a few weeks ago they put out this massively comprehensive episode on Alphabet and Google, which is one of their best yet, so had to put it in the newsletter.

In 2005, Google's stock plummeted 27% after announcing they were investing heavily in products beyond search. Wall Street saw a "pure play" search company suddenly "tossing balls into the air like a drunken juggler." But what looked like reckless diversification was actually one of the most brilliant strategic plays in tech history.

This is the untold story of how Google transformed from a search engine into the platform company of the web era, launching 15 products with over half a billion users each—seven with over two billion users. That's over 25% of humans using seven Google products. From Gmail's revolutionary Ajax technology to YouTube's $50 billion revenue empire, every move was calculated to defend against Microsoft whilst fulfilling their mission to organise the world's information.

🎥 Watch the full episode here
📆 Published: September 2025
🕒 Estimated Reading Time: 8 mins. Time saved: 245+ mins! 🔥

🚀 The Gmail Revolution: How One Engineer Changed the Web Forever

Gmail wasn't just an email service—it was the birth of the modern web application. When Paul Buchheit launched it on April Fool's Day 2004, offering one gigabyte of free storage when competitors offered 2-4 megabytes, it seemed too good to be true.

The breakthrough wasn't just storage. Buchheit had discovered Ajax (Asynchronous JavaScript and XML), enabling dynamic web pages that didn't reload with every action. "This is almost like Google search all over again," explains the hosts, describing how people realised you could create applications that looked and felt like installed software but ran entirely in a browser.

The strategic motivation was crystal clear: Google's entire business depended on Microsoft's Internet Explorer and Windows. Over 90% of Google searches happened on Windows PCs using Internet Explorer. "Google exists at the pleasure of Microsoft," the hosts note. Gmail was designed to make consumers demand rich web applications, creating leverage against any future Microsoft threats.

Key Takeaways:

• Gmail pioneered Ajax technology, enabling the first truly dynamic web applications

• The invite system created viral growth whilst managing infrastructure costs

• Strategic defence against Microsoft's platform dominance was the primary driver

• Gmail grew from 1,000 beta users to over 2 billion today

📔 Want to read the full breakdown?

This is just the beginning. The complete analysis covers:

• 🗺️ Maps: How the API launched a thousand startups and became a $10B business
• 📝 Docs & Sheets: The Trojan horse that distracted Microsoft
• 🎬 YouTube: From $1.65B "mistake" to $500B empire
• 💰 DoubleClick: The $3.1B chess move that blocked Microsoft
• 🔍 Search Dominance: Becoming the world's largest ad seller
• 🌐 Chrome: The browser that saved Google's future
• 🎯 The Innovation Playbook: What this means for today's leaders

Continue reading the full article →

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 🍽️.