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? 🥘

ChatGPT becomes a platform with 800M users (hello 2007), and Disney drops $1B on Sora with 200 characters licensed. Meta trains on China's Qwen 18 months after Chinese firms borrowed Llama (uno reverse), space data centres apparently now make sense, and FLUX.2 nails character consistency while McDonald's AI ad gets pulled for being terrifying.

Clearly we're not slowing down for Christmas any time soon.

On that note, will I do another one of these next week? 🎄 Not sure just yet… let's see how many mince pies and “festive cheer” I get through. Not missed a newsletter in over two years though, so perhaps I'll just do the news :)

Let's go 🚀🎅

📰 Not boring

🥊 The AI Arms Race

  • GPT-5.2 released in three variants: Instant ($1.75/1M tokens), Thinking (extended reasoning), and Pro (highest quality)

  • Gemini Deep Research drops same day - agent with citations and multi-step analysis

  • GPT-5 gets "danger label" - cybersecurity score jumped from 27% to 76% in one month on CTF exercises

  • Gemini 3 Flash launched - faster, cheaper, 81.2% on multimodal benchmarks

  • Gemini now lives inside iOS Chrome

OpenAI and Google dropped major updates same day. GPT-5.2 arrives in three tiers with August 2025 knowledge cutoff. The "danger label" on GPT-5 isn't symbolic - its CTF cybersecurity performance nearly tripled in one month (27%→76%). Either the model got dramatically better at hacking, or someone figured out how to measure it properly.

Also? So. Many. Model. Releases.

📱 ChatGPT Becomes a Platform

ChatGPT became a platform with 800M weekly active users - developers submit apps, monetisation coming later; hello early 2000s. Is THIS the App Store moment for AI? Disney isn't just experimenting; they invested $1B and licensed 200+ characters to Sora (though notably, no talent likenesses). If your APIs don't work with ChatGPT, you're getting disintermediated.

💼 Leadership Moves

  • Slack CEO Denise Dresser jumps to OpenAI as first Chief Revenue Officer (OpenAI "code red" for profitability)

  • Meta trains AI on China's Qwen - power reversal after Chinese firms used Llama in 2023-24

Slack's CEO became OpenAI's first-ever Chief Revenue Officer - expect more profitability and value themes moving forward. Meanwhile, Meta's training on China's Qwen model marks a power reversal: just 18 months ago, Chinese firms were borrowing Meta's Llama. Geopolitical AI collaboration is happening quietly while everyone argues about regulation. The narrative shifted faster than the policy.

🚀 Space Race Meets Data Centres 👀

Data centres hit a wall: $64B in US projects blocked or delayed by cities. Google's Project Suncatcher achieves 8x solar efficiency in orbit. Blue Origin and Starcloud are racing to build the infrastructure. Nvidia's chip smuggling prevention tech arrived quietly - protecting supply chains matters when ground-based expansion is stalling. The bottleneck isn't models anymore, it's real estate and power grids. This all just sounds mad.

🎨 Creative AI's Uncomfortable Reality

McDonald's and Coke's AI ads got roasted while FLUX.2 solved character consistency through “disentangled unconditional guidance” (keeping the same person across multiple images). Codex Mortis launched - first 100% AI game (code, art, sound, music) - mixed reviews. ChatGPT Image is 4x faster. The most important part? AI model naming remains utterly out of control. Can marketing please step up?

🔒 Privacy, Safety & The Regulation Wave

The trust infrastructure is breaking: badges don't mean safe, models are learning to hack, regulation is fragmenting. Product teams shipping AI without consent frameworks and failure modes are about to discover what "liability" means.

📲 Product & Platform Updates

Platform Expansion:

AI Features:

Strategy Divergence:

Platform Consolidation:

YouTube securing Oscars exclusivity, Instagram pushing Reels to TVs, DoorDash spinning out discovery apps. The platform playbook is evolving: go where users are (big screens), lower AI friction (selfie try-ons, natural language), and test aggressive monetisation ($20/month browsers). Meanwhile, consolidation continues - Meta killing desktop Messenger, Google sunsetting dark web reports. Strategy divergence: Apple embracing Google Cast while Netflix restricts it. The "build vs buy vs bundle" decisions are getting weirder.

🎯 Quick Hits

Business:

Robots:

Hardware:

Research:

What This Means

This week wasn't just busy - it was, well, really busy. Sorry I've run out of superlatives and ways to describe it. Every week is insanely busy it seems. The AI race seems to be moving to ecosystems, infrastructure, and platform strategies. ChatGPT became a platform (again). Google continued to put Gemini everywhere. OpenAI got safety labels.

The platforms are maturing. The tools are getting better and better and it the provider's narrative is slowly shifting towards value creation. Your competitors are already shipping.

Productivity Tapas: Time-Saving Tools & Workflow Automation

  • Vybe: Turn natural language into secure internal apps with built-in auth, roles and deployments. 3,000+ integrations. Start fresh or remix proven templates

  • Tiimo: turns your to-dos into doable plans with visual timelines, AI task breakdown, and tools that cut through the chaos (iPhone App of the year 2025)

  • mintlify: Mintlify keeps your documentation accurate by monitoring your codebase, detecting user-facing changes, and proposing updates the moment you ship

    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

Strategy: How Airbnb Transformed Its Revenue Through Stunning Photography

Tom explores Airbnb's groundbreaking photography strategy that doubled its revenue in just one week. By identifying poor visuals as a major barrier to bookings, co-founders Brian Chesky and Joe Gebbia took the initiative to enhance the quality of property images, leading to significant revenue growth. Read the full article here.

💡 "A picture tells 1,000 words." This quote reinforces the importance of visual storytelling in building trust and driving action in a marketplace environment.

Key Takeaways

Identify Pain Points: Initial analysis revealed that poor photography was a huge barrier to booking. This prompted the hands-on engagement from co-founders to directly address host issues, showcasing the value of personal involvement in problem-solving.
Manual to Scalable Solutions: Airbnb's founders embraced the ‘do things that don’t scale’ mindset, learning firsthand from hosts which ultimately informed a scalable professional photography programme. This redefined their approach to user engagement.
Building Trust through Visuals: The programme tackled marketplace trust issues by ensuring high-quality images accurately represented listings, transforming guest perceptions and improving booking conversions significantly.
Creating a Competitive Edge: By investing in professional photography, Airbnb not only increased revenue but also established a unique selling proposition that differentiated them from competitors, offering guests peace of mind.
Systematisation of Operations: As the photography initiative grew, Airbnb streamlined processes to manage their network of freelance photographers, which allowed them to maintain quality while expanding their reach globally.
Data-Driven Insights: The feedback gathered during initial photography sessions proved invaluable in shaping future operational strategies and core business decisions, demonstrating the power of qualitative data in driving growth.

Tom, Strategy Breakdowns

Strategy: Strategies That Empower Decision-Making

Tim Herbig discusses the pitfalls of a poorly defined product strategy, illustrating how it can lead to decision paralysis when actionable opportunities arise. He emphasises that a successful strategy should provide clear guidance on when to accept or reject new initiatives. Read the full article here.

💡 "Strategy has one job: helping you confidently say yes or no to opportunities."

Key Takeaways:

Decision Paralysis: A vague product strategy can lead to indecision, making it difficult for teams to prioritise tasks when new requests arise. This often results in escalating debates rather than effective action.

Practical Usefulness: Successful strategies go beyond theoretical interests; they must enable teams to make real choices. This focus on practical application differentiates successful strategies from mere compliance with templates.

Interconnected Strategic Components: Elements of a strategy should function like dials; adjusting one should impact others, creating a cohesive decision framework. This interconnectivity is crucial for aligning team goals with strategic objectives.

Progress Wheel Concept: Understanding the relationship between strategy, goals, and discovery allows teams to continuously refine their approach. This proactive mindset fosters measurable quarterly goals and real progress.

Tim Herbig, Practice Product

UX: Personalised Progress and Insightful Engagement

Once again, Built for Mars is back this time sharing innovative user experience strategies from platforms like Strava, Spotify, and Duolingo. The newsletter highlights how these services create meaningful connections through contextual insights and engaging features. Read the full article here.

💡 "Spotify will tell you a specific global rank to contextualise just how much you listen to a particular artist." This ranking not only makes the data digestible but also enhances shareability among users.

Key Takeaways:

Dynamic Search Enhancements: Etsy introduces dynamic query add-ons during searches, improving user engagement by providing contextual search suggestions based on initial input. This feature personalises the shopping experience by offering real-time recommendations.

Contextualised Performance Recaps: Strava’s yearly recap contextualises individual accomplishments against other users, enhancing motivation and user retention by highlighting personal progress in a competitive light.

Global Listening Ranks: Spotify's specific global ranking system allows users to understand their listening habits in a more relatable way, transforming abstract percentages into clear standing, fostering a sense of community among listeners.

Nudge Features for Habit Formation: Duolingo employs nudges based on user data, promoting engagement and habit-building through tailored reminders that encourage continuity in learning.

Email Automation for Travel Planning: Flighty simplifies travel planning by allowing users to forward flight confirmations, which automatically organizes travel itineraries, saving time and enhancing user convenience.

Monzo's Salary Integration: Monzo’s new feature for salary integration not only offers financial perks but also enhances user trust and satisfaction through seamless banking experiences tailored to individual needs.

Peter Ramsay: Built for Mars

🎙 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

🤖 Why One Founder Believes 'Vibe Coding' Will Kill SaaS

One founder. Eight months. $80M exit. Maor Shlomo built Base44 solo, sold to Wix, and believes custom AI-built software will kill most SaaS companies.

In 8 months, one founder went from bootstrapped idea to an $80 million acquisition by Wix - without a team, without funding, just AI and conviction. Maor Shlomo, founder of Base44, explains why he believes most SaaS companies will be eliminated, why margins don't matter in the AI era, and where smart money should go today.

20VC with Harry Stebbings

  • 🎥 Watch the full episode here

  • 📆 Published: 24th November 2025

  • 🕒 Estimated Reading Time: 12 mins. Time saved: 58+ mins! 🔥

💡 Top tip - This is a dense conversation covering market dynamics, competitive strategy, and investment thesis. Read the executive summary below, then jump into the full article sections most relevant to your role (defensibility for product leaders, investment criteria for VCs, or market dynamics for founders).

Key insights from the full article:

  • 🎯 The strategic sale - Why selling to Wix for $80M actually tripled Maor's chances of building something that matters vs raising venture capital or staying bootstrapped

  • 🏗️ The vibe coding thesis - Why most CRM, task management, and organisational tools will be replaced by custom-built software that users own completely

  • 🛡️ Defensibility through infrastructure - Features copy in weeks, but building a vertically integrated platform with compute, databases, and integrations creates real moats

  • ⚔️ The Google threat - Why model provider consolidation is the existential risk, and why Google with Gemini + Cloud + data is the scariest competitor

  • 💸 Margins don't matter (yet) - Strategic decisions assume model costs trend toward zero; growth and distribution trump margin optimisation today

  • 🔄 Insane switching dynamics - How $1M in AI spend moves between providers with a single line of code, creating unprecedented market volatility

  • 📊 Sentiment as success metric - Measuring user frustration in chat messages beat traditional metrics for AI-native products

  • 💰 Revenue in the AI era - What early-stage investors should actually weight when evaluating $10M-$100M ARR AI companies

  • 💼 Where smart money goes - Vertically integrated businesses (even non-tech like law firms or hospitals optimised with AI) beat prompt wrappers

  • 🚀 No speed bump ahead - Why Maor believes we're only scratching the surface of economic value, even with today's models

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

or to participate