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!

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 cooking this week? 🧑‍🍳

📰 Not Boring – Is AI making us more productive, or just cognitively bankrupt? McKinsey drops a reality bomb on AI in the enterprise, while Alibaba-backed Minimax wants to be your new AI agent. Plus, Apple might be eyeing Perplexity, and OpenAI is reportedly gunning for Office and Workspace.

⌚️ Productivity Tapas: This week's lineup includes a collaboration platform that's about to change how teams work together, a tool that turns any messy document into crystal-clear diagrams, and an AI design assistant that transforms random thoughts into polished UI mockups. The future of workflow automation is getting wild.

🍔 Blog Bites: How Atlassian shapes new products with a "lighthouse customer" methodology, why the internet killed general-purpose products (but AI might bring them back), and Bill Atkinson's 10 timeless rules for making interfaces more human.

🎙️ Pod Shots: Windsurf's Varun Mohan shares the counterintuitive principles that helped him build AI's hottest startup—and why most startup advice about persistence is dangerously wrong.

Ready to separate the signal from the noise? Dive in! 🚀

📰 Not boring

The AI Reality Check

• MIT research suggests AI isn't making us more productive. It's making us cognitively bankrupt

  • But also worth reading “The Great PowerPoint Panic of 2003”, which sort of said the same thing back in the day…

• McKinsey report on Agents in the Enterprise drops a reality bomb: 80% of companies use gen AI but see no real business impact. Turns out giving everyone Copilot wasn't the magic bullet. Other highlights:

  • AI Agents are Different: Unlike basic AI tools, agents can work autonomously across multiple systems to complete complex business processes without constant human input

  • But Success Requires Process Overhaul: Companies need to redesign how work gets done around these agents, not just bolt them onto existing workflows

  • Plus New Challenges: Success depends on building proper oversight systems and getting employees to trust and adopt agent-driven processes.

AI Hype Zone

• You can now try Alibaba backed Minimax general agent for yourself

• AI avatars in China just proved they are ace influencers. It only took a duo 7 hours to rake in more than $7 million. The attention economy has found its perfect performers

• ElevenLabs releases experimental AI PA (free for next few weeks). PLUS a stand-alone voice-generation app

• Google brings new Gemini features to Chromebooks, debuts first on-device AI whilst Google Photos merges classic search with AI to speed up results

• Google’s new robotics AI can run without the cloud and still tie your shoes. What could possibly go wrong?

• Anthropic now lets you make apps right from its Claude AI chatbot. Slick.

• aaaand this AI-powered startup studio plans to launch 100,000 companies a year — really

Market Moves & Rumours

• Apple looking to buy Perplexity, but Ben Evans questions whether it’s bankers “flying a kite” given “it doesn’t have its own models, it struggles to rank in the top 200 in the App Store”

• OpenAI designs rival to Office and Workspace. Continuing to break every product ‘rule’ there is (and getting away with it)

  • Classic product strategy says focus: pick your ICP, nail a use case, expand methodically. OpenAI did the opposite - launched as a general "ask me anything" model, now sprawling into productivity software, search, coding tools, whatever catches their fancy.

    This should be a cautionary tale about lack of focus. Instead, it's working - for now. The question is whether there's any moat when everyone else can build the same general-purpose intelligence.

• Apple heard your complaints about the Liquid Glass Control Centre

Everything Else

• Here’s a great summary of Andrej Karpathy’s recent take on how sw is changing. Here’s the full video if you want to take it all in

• Coinbase Secures MiCA Licence: A Milestone in Europe’s Crypto Evolution

• The résumé is dying, and AI is holding the smoking gun

• Julie Bornstein’s Daydream is releasing an AI-powered chatbot for fashion-related shopping

  • Whilst Google launches Doppl, a new app that lets you visualize how an outfit might look on you

• Here’s a primer on getting up to speed with AI. Covering, AI Tools, AI Automations & Agents, Vibe Coding. From images, video creation, and general tools it’s a solid resource

• Chrome for Android now lets you move the address bar to the bottom, too (only 4 years after Safari)

• Kraken rolls out all in one global P2P money app ‘Krak’

Fact-based news without bias awaits. Make 1440 your choice today.

Overwhelmed by biased news? Cut through the clutter and get straight facts with your daily 1440 digest. From politics to sports, join millions who start their day informed.

Sign up now!

⌚️ Productivity Tapas: Time-Saving Tools & GPTs

• Liveblocks 3.0: gives you ready-made features like AI Copilots, Comments, and Multiplayer Editing to make your product more engaging and grow your business

• Fluig: turn any documents & ideas into a diagram (mindmaps, flowcharts, cards +++)

• Flowstep: Turn your thoughts into beautifully crafted designs. AI Design assistant for transforming simple text prompts into UI designs, wireframes, and flows

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

Product Strategy: How Atlassian Shapes New Products

Love it or hate it Jira is still pretty ubiquitous amongst product teams… Tom Alder recently interviewed Tanguy Crusson, Head of Product for Jira Product Discovery (JPD), revealing Atlassian's pragmatic approach to product development that balances opinionated design with customer flexibility. Read the full article here.

💡 "If there's one methodology I'll take with me everywhere, it's this one: The biggest challenge when building new products in big companies is that teams are often way too slow because they over-validate with quantitative data instead of qualitative insights... instead of building for 100s of customers from day one (based on data), what if we build for just 10 (based on conversations)?"

Strategy: The Internet Killed General-Purpose Products. AI Will Bring Them Back.

Dan Hockenmaier explores how the internet enabled specialised products to outcompete general-purpose ones by excelling at specific value propositions, and argues that AI will reverse this trend by enabling products that can be great at multiple things simultaneously. Read the full article here.

💡 "AI's core strength is the intelligence and flexibility to tailor outputs to each customer's individual needs. This will enable products that can be great at multiple value props simultaneously, causing a re-centralisation of industries around fewer clusters led by new, massive general-purpose products."

Design: Bill Atkinson's 10 Rules for Making Interfaces More Human

Figma commemorates the Apple pioneer whose QuickDraw and HyperCard programmes made the Macintosh intuitive enough for nearly anyone to use, deriving 10 timeless principles from his approach to computing and creativity. Read the full article here.

💡 "He knew the mark of great interface design was when a user could stop thinking about the tool and focus entirely on the work itself." This captures Atkinson's fundamental belief that the best interfaces become invisible, allowing users to concentrate on their creative goals rather than wrestling with technology.

🎙️ 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

🚀 Windsurf's$3B Journey: How to Build AI's Hottest Startup Without Falling in Love with Your Ideas

Varun Mohan has built one of Silicon Valley's most talked-about AI startups, but his path to Windsurf wasn't linear. The CEO and co-founder has pivoted twice, changed company names twice, and learned some hard lessons about the difference between being right about technology trends and being right about market timing.

In this recent conversation with Harry Stebbings, Mohan reveals the counterintuitive principles that helped him build Windsurf into an AI coding platform that's reshaping how developers work—and why he believes most startup advice about persistence is dangerously wrong.

 🎥Watch the full episode here

📆 Published: June 2nd, 2025

🕒 Estimated Reading Time: 3 mins. Time saved: 60 mins🔥

💡 The Pivot Paradox: Why Being Wrong Fast Beats Being Right Slow

Mohan's journey started with Exa Function, a GPU virtualisation company built on a prescient but incomplete thesis. "We basically said Nvidia is going to sell a lot of GPUs," he explains. They were right about that part—spectacularly so. But they were wrong about something crucial: they thought hundreds of different model architectures would emerge, requiring diverse GPU workloads.

Instead, everything became transformers. "In a world in which all the architectures look the same, there's very little reason for us to be a differentiated infrastructure provider," Mohan admits.

The lesson? "One of the weird things about startups is that you don't win an award for doing the same wrong thing for longer. Down the line when you fail, none of them will care."

Key Takeaways:

• Being partially right about a big trend isn't enough—you need to be right about the specific implementation

• Startups require "irrational optimism" and "uncompromising realism" simultaneously

• The market doesn't reward persistence with bad ideas, only persistence with good execution

🎯 The Focus Doctrine: Why One Thing Done Well Beats Five Things Done Okay

When Windsurf pivoted from their GPU business—which was generating "a couple million in revenue"—they made a radical decision. Over a weekend, Mohan and his co-founder decided to work on Kodium (their previous product). By Monday, the entire company had shifted focus.

"Companies don't succeed because they do many things well. They succeed because they do maybe one thing really well," Mohan explains. "It never makes sense to be diverting your resources to work on two different products that have different exponential growth curves."

This philosophy extends to hiring. Mohan waits until he's "drowning in a role" before making hires, contrary to conventional wisdom about hiring six months ahead. "Startups don't fail because they look like messes inside. Startups fail because they don't do the right thing well enough."

Key Takeaways:

• Revenue cannibalisation is scary but necessary when pivoting to a better opportunity

• Focus beats scale in early-stage startups—one great product trumps multiple good ones

• Internal chaos is acceptable as long as you're executing on what matters most

⚡ The Speed Advantage: Why Velocity Trumps Moats in AI

In a world where "time to clone" is shrinking rapidly, Mohan argues that traditional moats don't apply to startups. Even Nvidia, despite CUDA's apparent dominance, succeeds primarily through execution speed, not switching costs.

"If CUDA didn't exist, [large companies] would find a way to write assembly, the lowest level code, and make that run on the GPUs," he argues. "The reason why people use it is it's just awesome."

For Windsurf, this means shipping major releases every 1-2 weeks and maintaining what Mohan calls "existential dread"—the constant fear that drives rapid iteration. "If we don't ship something amazing to our users in three months, we will become irrelevant."

Key Takeaways:

• Speed of learning and iteration matters more than traditional competitive advantages

• Large companies struggle with velocity because they lack existential pressure

• Being first to market provides learning advantages that compound over time

🤖 The AI Development Reality Check: Why Agents Aren't Ready to Replace Developers

Despite the hype around AI agents replacing software engineers, Mohan offers a more nuanced view. When Devon launched with claims about replacing junior developers, "everyone was giddy with excitement," but the reality fell short.

"People think these systems are way more capable than they are today," he explains. Current AI excels at reading and understanding code but struggles with complex, multi-step tasks that require high accuracy and easy correction.

For asynchronous agents to work, "people's expectations on the output are going to be really high. Your quality has to be really, really high. And if the quality is not high, it better be easy to correct."

Key Takeaways:

• AI agents are better at augmenting developers than replacing them

• Asynchronous AI work requires near-perfect accuracy due to slow feedback loops

• The technology is improving rapidly, but current limitations are real and significant

🏢 The Enterprise Insight: Why Java Developers Matter More Than You Think

One of Windsurf's key advantages comes from understanding enterprise reality. Over 50% of their revenue comes from enterprise customers, many of whom have large Java codebases using JetBrains IDEs rather than VS Code.

"JP Morgan Chase, who is a customer of ours, over 50% of their developers are JetBrains users," Mohan reveals. While Windsurf forked VS Code to create their own editor, they also built plugins for JetBrains to serve enterprise customers completely.

This enterprise focus also shapes their view of AI's future impact. While some developers will operate purely through natural language, "there will always be production critical applications" that require deep technical expertise.

Key Takeaways:

• Enterprise customers often have different tooling needs than Silicon Valley startups

• Supporting existing workflows can be more important than forcing adoption of new tools

• The future will likely include a spectrum of technical roles, not wholesale replacement

🔬 The Product Development Philosophy: Small Teams, Big Bets, Fast Failures

Mohan's approach to product development centres on small, opinionated teams working on unproven ideas. "When a product or an idea has no legs and you're proving it out, you should actually have very few people working on an idea," he explains.

Teams of 3-4 people work on new concepts until they can prove that "even the crappy version of that idea is already amazing." Only then do they resource up.

About 50% of internal projects fail, but "the one thing that works pays for the hundreds of things that fail." This high-failure rate is intentional—it's better to fail fast on bad ideas than to slowly succeed with mediocre ones.

Key Takeaways:

• Small teams avoid the coordination problems that plague larger groups working on unproven concepts

• Early versions of great ideas should be obviously valuable even when rough

• High failure rates are healthy when you're pushing the boundaries of what's possible

🌐 The Remote Work Reality: Why In-Person Still Matters

Despite the trend toward remote work, Windsurf operates as a 100% in-person company. Mohan sees this as an "unfair advantage on speed" that enables rapid pivoting and coordination.

"Right now if it's 2 or 3 PM everyone at the company can be called, can be brought into a room at that very moment," he explains. This flexibility becomes crucial when ideas need to change quickly or when the entire company needs to pivot direction.

While he acknowledges remote work can succeed with "much more principled" companies, the in-person advantage is particularly valuable for startups operating in rapidly changing markets.

Key Takeaways:

• In-person work provides coordination advantages that matter for fast-moving startups

• Remote work requires more structure and discipline to achieve the same velocity

• The choice between remote and in-person should align with your company's strategic needs

🔮 The Future of Software Development: More Human, Not Less

Looking ahead, Mohan envisions AI making software development more accessible without eliminating the need for deep technical expertise. "The expectation for PMs is not going to be 'I tell someone to go and build something.' It's going to be 'you should go out and build it too.'"

But this democratisation has limits. For mission-critical systems like JP Morgan's transaction processing, "you probably don't want to vibe code." There will always be applications that require developers who can "go all the way down to the weeds."

The result will be a spectrum: more people able to build software through natural language interfaces, but still a need for experts who understand systems at the deepest levels.

Key Takeaways:

• AI will expand who can build software, but won't eliminate the need for technical expertise

• Different types of applications will require different levels of technical depth

• The future includes both democratised development and specialised expertise

💰 The Billion-Dollar Reality Check: Why Solo Success Is a Myth

When asked about predictions of solo billion-dollar companies, Mohan is skeptical. "We live in a capitalist market so there's competition. If you're able to do something with one person, unless you're just a thousand times smarter and more capable than everyone else, someone with two people that are as smart as you are probably going to be able to do your idea."

This perspective reflects his broader philosophy about competition and market dynamics. Success comes from execution speed and learning velocity, not from trying to build impenetrable moats or work in isolation.

Key Takeaways:

• Competition is inevitable in valuable markets—plan for it rather than trying to avoid it

• Solo success stories ignore the reality of market dynamics and competitive pressure

• Building great companies requires teams, not just great individual contributors

So lots of interesting takeaways - some I’m less sold on - but I do like his view that success comes not from falling in love with your ideas, but from falling in love with solving real problems. His willingness to pivot quickly, focus ruthlessly, and maintain "existential dread" about staying relevant has created one of the most compelling companies in the AI coding space.

 🎥Watch the full episode here

📅Timestamps:

• 00:00 Intro

• 01:28 When to Give Up vs When To Stick at It

• 03:33 “Never Fall in Love With Your Idea”

• 07:42 What Founders Get Wrong About Being First

• 11:10 What Would Windsurf Do If They Had Unlimited Resources

• 14:40 Will Lovable and Bolt Ultimately Compete with Windsurf and Cursor

• 17:31 The Product Development Rule That Breaks All Startup Rules

• 20:26 The Cold Truth About Moats in the AI Era

• 25:20 Remote vs. In-Person

• 31:59 Who Actually Counts as an Engineer in 5 Years?

• 33:54 Will Product Managers Even Exist in 2030?

• 37:10 Async Agents Are Coming But Most Will Fail & Why?

• 41:03 The Truth About Agent-Only Workflows

• 44:26 The One Area of Engineering That AI Will Eat Next

• 42:54 What Cursor Got Right (That Windsurf Didn’t)

• 50:28 Are LLM APIs Already Commoditized?

• 53:31 Should Model Companies Own the App Layer?

• 59:01 What Does Varun Want to be Remembered For?

• 01:00:42 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 🍽️.