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! The tastiest bites from the world of Product and Tech.

If you’ve been forwarded this or just stumbled upon it, you’re in for a treat. For the best reading experience, check out the web version and sign up for future editions here.

What’s on the menu this week? 🧑‍🍳

📰 Not Boring – This week we’ve got it all, from Twitter-inspired mega hackathon to groundbreaking AI models that are reshaping what's possible. Nvidia’s making waves with their GTC announcements and Gemini is getting a serious glow-up with some game-changing collaboration features. Meanwhile, Apple's internal struggles with AI are becoming increasingly public. And that's just scratching the surface!

⌚️ Productivity Tapas – We’ve got AI-powered Figma heatmaps, simplified notification systems, and a great app for turning recordings into actionable insights. Plus, a pretty neat new AI image generation tools that promises to have text nailed.

🍔 Blog Bites – Stephen Coyner explores SaaS's future in an agent-driven world, Andrew Chen dives into "vibe coding," and Peter Ramsey shares essential micro-UX decisions for the AI era.

🎙️ Pod Shots – This week’s summary is of Instagram co-founder and Anthropic CPO Mike Krieger as he unpacks value creation in the AI world, sharing insights on differentiation, talent, and the future of work.

Plenty to get stuck into - off we go! 🚀

📰 Not boring

• Bolt CEO launches world’s largest hackathon off the back of twitter user’s suggestion

• Anthropic adds web search to its Claude chatbot [paid; but soon for free users]

• Adobe apps get an agentic overhaul for businesses to activate AI Agents in customer experiences and marketing workflows

• EXAONE Deep from LG released - setting a new standard for reasoning AI (beats R1 with a fraction of the size). Another day another model…

• Nvidia GTC 2025

  • New chip roadmap rolled out 20 petaflops this year 100 by 2027

  • New Groot N1 open sourced model AI model for humanoid robots

  • Nvidia announces two ‘personal AI supercomputers’

  • GM to use Nvidia AI chips, software to automate vehicles, factories

  • Partnership with DeepMind and Disney

• Google corner

  • Big Gemini overhauls announced

    • Canvas - brainstorm and collaborate on large projects from the chatbot

    • Audio Overview; turn text (docs, research, whatever) into an engaging audio ‘podcast’ using the Notebook LM tech directly from the chat

    • Gemini 2.0 Flash native image generation released, allowing you to merge text and images together and use conversational prompts to edit, add labels etc.

  • Gmail gains an ‘Add to calendar’ button, powered by Gemini

• WhatsApp’s testing a new feature to make your busy group chats less chaotic [YES PLEASE]

• OpenAI testing ChatGPT connectors for Google Drive and Slack so it can answer questions informed by files, presentations, sheets, and Slack convos

• Spotify payouts hit $10bn as highlighted in its Loud & Clear report

• Apple corner

  • “Something Is Rotten in the State of Cupertino” scathing report on Apple from John Gruber (well connected and respected Apple Blogger)

    • Whilst Apple’s Siri Chief Calls AI Delays Ugly and Embarrassing, Promises Fixes

  • But on the upside, Apple is reportedly bringing live translation to AirPods

• Patronus AI announces LLM to fact-check LLMs

• Current Y Combinator startups are fastest growing, most profitable in fund history because of AI

• Shopify has acquired Vantage Discovery, an AI search company founded by former Pinterest engineering leaders

• Rippling manager allegedly fed stolen customer and employee information to company’s rival. Rippling sues Deel, Deel denies ‘all legal wrongdoing,’ and Slack is the main witness. WILD

⌚️ Productivity Tapas: Time-Saving Tools & GPTs

• Attention Insight for Figma: AI powered heatmaps to predict where users will look first

• Novu: turns complex multi-channel notification delivery into a single <Inbox /> component. Add an inbox to your product with 6 lines of code

• ScreenApp: Turn your screen recordings into actionable insights and save hours daily

• Ideogram: AI image generator to create realistic pictures with accurate text (allegedly 🙂 )

• Zuni: access all of the most advanced AI models from your Chrome sidebar

• Mirage: Generate energetic, high-converting ads with people that don’t exist [not perfect, but holy crap]

Remember, as a Product Tapas Pro subscriber you can access the full time saving tools database for fast approaching 400 time-saving tools relevant for product managers and founders 🔥.

Check the link here to access.

🍔 Blog Bites - Essential Reads for Product Teams

Business Models: What Happens to SaaS in a World with Computer Using Agents?

Stephen Coyner explores how the emergence of computer-using AI agents could fundamentally reshape the SaaS industry. His analysis reveals how traditional SaaS products risk becoming commoditised as agents begin to mediate between humans and software, forcing a dramatic evolution in how enterprise software delivers value. Read the full article here 

💡 "In an agent-driven future, many of the elements that once differentiated SaaS products—polished user experience, brand loyalty, platform lock-in through muscle memory—lose their importance" as agents become the primary users of software, shifting value toward comprehensive platform solutions.

Technology: Vibe coding, some thoughts and predictions

Andrew Chen explores the emerging trend of "vibe coding," a new paradigm where AI tools like Cursor, Bolt, Sonnet etc. enable users to create software with minimal technical expertise. Coined by Andrej Karpathy, vibe coding shifts the focus from traditional programming to intuitive, AI-driven interactions, opening the door for a new generation of creators. Read the full article here. Also whilst on the subject, ICYMI Andrej’s YouTube channel is full of amazing videos explaining LLMs and how to get the best out of them.

💡 "I just see stuff, say stuff, run stuff, and copy-paste stuff, and it mostly works." This playful yet profound description of vibe coding highlights how AI is transforming software creation into an accessible, iterative, and creative process.

Design: UX Bites. Micro-UX Decisions in an AI-Driven World

Back again with another great UX Bites from Peter Ramsey / Built for Mars. His weekly UX Bites demonstrate how thoughtful micro-interactions can significantly improve user experience. In one of his latest he explores how AI's acceleration of development creates an imbalance in product teams, where developers increasingly make crucial UX decisions without corresponding design support. Read the full examples here 

💡 As AI enables faster development, the challenge shifts from technical implementation to design decision-making.

The overall theme (as ever) was around how small UX choices, from error states to component behaviours, become critical differentiators that every team member needs to understand.

🎙️ Pod Shots - Bitesized Podcast Summaries

Remember, Product Tapas Pro subscribers get access to an ever growing database of all the top Podcast summaries we’ve ever created. 

Check it out here

🌍 Where Will Value Be Created in an AI World? Insights from Mike Krieger, Instagram Co-Founder & Anthropic CPO

This week we cover the recent 20 VC podcast with Mike Krieger, the co-founder of Instagram and now the Chief Product Officer at Anthropic. He covers “Where will value be created in an AI-driven decade?” including a stack of really interesting topics such as model quality vs. Product UX, why models will become more different than similar and his take on the future role of software developers.

 🎥Watch the full episode here

📆 Published: March 3rd, 2025

🚀 Key Themes and Insights

🧩 1. Differentiation is Key to Value Creation

Mike emphasised that the most durable value in the AI era will come from businesses that leverage:

• Differentiated Go-to-Market (GTM) strategies

• Unique data pools or proprietary data

• Deep industry expertise

For example, industries like healthcare, finance, and legal services offer opportunities for startups to create value by solving complex, unsexy problems that require significant upfront work. These are areas where AI can be fine-tuned to deliver long-term impact.

Takeaway: Sectors where you can combine unique data, industry knowledge, and AI capabilities will allow you to create a competitive edge. The legwork you do now will pay off as models improve.

🌱 2. Startups vs. Established Companies: Who Benefits More?

Mike discussed whether the AI wave benefits existing vertical SaaS companies or new startups. His answer? Both, but with different challenges:

• Established Companies: Must integrate AI without alienating existing customers. They need to "skate to where the puck is going" while maintaining trust.

• Startups: Have the freedom to overpromise and experiment but often lack the data and relationships that established players have.

Takeaway: If you’re in an established company, focus on incremental AI improvements that align with customer expectations. If you’re a startup, paint a bold vision of the future and find early adopters willing to take a bet on you.

🔮 3. Build for Today, but Stay Ready for Tomorrow

AI startups often face a dilemma: Should they build for the capabilities of today’s models or anticipate future breakthroughs? Mike’s advice is clear: Don’t wait for models to be perfect. Instead:

• Explore the space aggressively.

• Build context and expertise in your domain.

• Be ready to capitalise on the next generational leap in AI models.

Example: Companies like Cursor, which iterated on their product for years, were well-positioned to succeed when model capabilities finally caught up.

Takeaway: Start building now, even if today’s models aren’t perfect. The experience and insights you gain will position you to succeed when the technology advances.

🛠️ 4. The Three Pillars of Success in AI Models

Mike outlined three critical factors for long-term success in the foundation model space:

1. Talent: Attracting and retaining top talent is essential. Talent density creates a flywheel effect, attracting more great people.

2. Focus: Models will become more differentiated over time. Companies need to double down on areas where they excel (e.g., coding, agentic planning).

3. Partnerships: Success isn’t just about selling APIs. It’s about building long-term partnerships with customers, co-designing products, and becoming an indispensable AI partner.

Takeaway: Invest in talent, focus on your strengths, and prioritise deep customer relationships over transactional API sales.

🧠 5. The Future of Work: AI as a Collaborator

Mike shared his vision for how AI will transform work, particularly in software development:

• Today: AI tools like Claude Code and GitHub Copilot assist developers but still require human oversight.

• Future: Developers will shift from writing code to delegating tasks to AI, reviewing outputs, and focusing on higher-level problem-solving.

Takeaway : Start integrating AI into your workflows now. Train your teams to collaborate with AI tools and focus on tasks that require human creativity and judgment.

🌐 6. The Role of Open Source and Distillation

Mike addressed the debate around open-source AI models and distillation (simplifying high-end models for broader use). While open source can propel innovation, it also raises concerns about security and sustainability.

Takeaway: If you’re building on open-source models, focus on creating differentiated value through proprietary data, workflows, or integrations.

📈 7. The Speed vs. Stability Tradeoff

As AI companies scale, they face a tension between shipping quickly and maintaining stability. Mike highlighted how Anthropic balances this:

• API Products: Prioritise stability and predictability.

• Consumer Products: Experiment more and iterate quickly to learn from user feedback.

Takeaway: Tailor your release strategy to your audience. Enterprise customers value stability, while consumer users are more forgiving of experimentation.

🌍 8. The Global AI Landscape

Mike discussed the underestimated AI capabilities of countries like China, emphasising their ability to innovate at scale. He also highlighted Europe’s potential to lead in areas like privacy and ethical AI design.

Takeaway for Executives: Keep a global perspective. Don’t underestimate competitors in other regions, and look to Europe for leadership in ethical AI practices.

🔥 Key Takeaways

For Product Managers:

• Use AI to streamline workflows and focus on high-value tasks.

• Experiment with AI tools like Claude Code to accelerate development.

• Prioritise user feedback to refine AI-powered features.

For Founders:

• Focus on sectors where you can combine unique data, industry expertise, and AI capabilities.

• Start building now, even if today’s models aren’t perfect.

• Balance bold vision with practical execution to attract early adopters.

For Executives:

• Invest in talent and create a culture that attracts top AI researchers.

• Build long-term partnerships with customers, not just transactional API relationships.

• Stay informed about global AI trends and competitors.

🌟 Final Thoughts

Mike Krieger’s insights remind us that we’re still in the early days of AI. The companies that succeed will be those that combine technical excellence with a deep understanding of their customers and markets. Whether you’re a product manager, founder, or executive, the key is to start experimenting, stay focused, and build for the future.

AI isn’t just a tool—it’s a collaborator, a partner, and a force multiplier. The question isn’t whether AI will create value, but how you’ll position yourself to capture it.

 🎥Watch the full episode here

📅Timestamps:

• (00:00) Intro

• (00:50) Where Will Value Be Created and Sustained in a World of AI?

• (01:40) Are Foundation Models Commoditised Today?

• (04:31) Should Founders Build for the Models of Today or Build for Models of the Future

• (06:55) Why Will Models Become More Different Than More Similar

• (12:59) Will Human or Synthetic Data Be More Prominent in the Future

• (18:02) Model Quality vs. Product UX

• (20:12) The Competitive Landscape of AI

• (31:49) Do We Underestimate China's AI Capabilities

• (33:31) What Did Anthropic Learn from Deepseek

• (34:44) Is Deepseek a Sustaining and Credible Threat?

• (38:09) Transitioning from Model Provider to Application Provider

• (43:44) What is the Role of a Software Developer in the Future

• (48:31) Balancing API and Consumer Products

• (52:25) Is Europe Stronger or Weaker in a World of AI

• (52:59) Quick-Fire Round

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 🍽️.