If your best friend becomes obese, your own risk increases by 171 percent

Dr Jason Fung, often called as the founder of intermittent fasting has a simple advice for you - your hunger matters.

In his latest book, The Hunger Code (Reset Your Body’s Fat Thermostat by Breaking the Ultra-processed Food Habit), he argues that the traditional "calories in, calories out" approach to weight loss is a scientific failure that ignores the complex hormonal regulation of body fat. He asserts that obesity is driven by persistent hunger rather than a simple lack of willpower, categorized into physical, emotional, and social types.

Subscribe now

Instead of calorie counting, Fung proposes three golden rules: avoiding processed items, practicing intermittent fasting, and fostering healthy social habits. By addressing the root causes of hunger, the book aims to provide a sustainable framework for long-term health and weight management.

Here are the big ideas from the book :

The Great Calorie Delusion

The dominant view since the 1970s is that obesity is simply an energy balance problem—”Calories In, Calories Out.” Yet, calorie-restriction diets have a perfect track record, unblemished by success. If you are obese, the standard advice to “eat less” carries a 99.35 percent failure rate. We must stop pretending this strategy works. As I’ve said, the obsessive focus on calories is neither useful nor effective; it’s the definition of insanity to keep giving the same advice and hoping for different results. Obesity is not a caloric imbalance; it is a hormonal one.

Share

Stop counting calories and start focusing on the hormonal impact of the foods you choose.

Food is Information, Not Just Fuel

Calories are a measure of heat energy in physics, but in human physiology, they are a sadly incomplete description of food’s fattening effect. Food contains two things: energy (calories) and information (the body’s hormonal response). Different foods generate completely different hormonal instructions. For instance, 200 calories of cookies spikes insulin, telling your body to store fat, while 200 calories of eggs releases GLP-1, telling you to stop eating. It’s as bizarre as saying that selling cocaine and selling men’s suits is the same thing because they both involve money.

Prioritize foods that signal satiety (like eggs) over those that trigger fat storage (like cookies).

Resetting Your Internal Fat Thermostat

The body regulates body fat using a homeostatic mechanism similar to a room thermostat. Obesity is not caused by eating too many calories; it’s a disorder where the “body set weight” is set too high.

If you open a window to cool a hot room without turning down the thermostat, the heater just cranks up to compensate. Similarly, if you just “eat less,” your body fat thermostat responds by increasing hunger and lowering your metabolic rate to regain the weight. You can’t use willpower to decide to be less hungry.

To lose weight permanently, you must use hormones to dial down your body fat thermostat rather than fighting your hunger.

Read more

In a famous experiment, psychologist Curt Richter placed wild rats in a container of water.

These weren’t fragile animals. Wild rats are aggressive, resilient, and physically capable swimmers. But when they found no escape, they stopped trying — and drowned within minutes.

Then Richter changed one variable.

He took another group of rats and rescued them just before they died. He dried them off, let them recover, and placed them back in the water.

This time, they didn’t last 15 minutes.

They lasted 60 hours.

Same bodies. Same water. Same struggle.

What changed was one thing: belief.

The rats had learned that survival was possible. And that single expectation radically expanded their endurance.

Humans do this too.

Not in barrels of water, but in careers, relationships, health, ambition, and creative work.

Most people don’t quit because they’ve reached their true limit. They quit because, somewhere along the way, they stop believing their effort will change the outcome.

And that matters more than we think.

For years, helplessness was treated as something we learn only after repeated failure. But newer neuroscience suggests something more unsettling: under stress, the brain often defaults toward passivity. In other words, giving up is not unusual. It’s efficient.

What has to be built is agency — the belief that your actions still matter.

That belief is not motivational fluff. It is a biological performance variable.

When you believe there is a path forward, you persist longer, recover faster, and act with more intention. When you don’t, your world shrinks long before your actual capacity does.

That is the central idea behind Beyond Belief by Nir Eyal: the assumptions you carry are not passive thoughts. They shape what you notice, how you feel, and what you do next.

And if that’s true, then one of the most important questions in life is not:

“What am I capable of?”

But:

“What have I already decided is impossible?”

Beliefs Are Tools, Not Truths

We often get stuck arguing whether a belief is “true”. This is a trap.

Beliefs are actually mental models we use to navigate reality. Like a carpenter choosing between a hammer and a saw, you can select beliefs based on how well they serve your goals.

Instead of asking, “Is this true?”, ask: “Does this belief serve me?”.

Are you currently living in a “fifteen-minute” reality, or are you ready to unlock the sixty hours of potential already inside you?

To discover the science-backed framework of Attention, Anticipation, and Agency—and how to stop limiting your own results—explore the full insights in Nir Eyal’s latest work: Beyond Belief (and do subscribe to BigIdeas for audiobooks summary of latest books, delivered weekly).

Grounded in neuroscience, psychology, and unforgettable case…

Read more

As the artificial intelligence typhoon descends upon us, this has become the mantra of tech CEOs and business pundits.

Here’s the twist: They’re wrong.

As AI excels at technical tasks like data analysis and coding, the future of work will be won by those who lead, persuade, influence, and inspire. And there’s one cognitive superpower that improves all those crucial soft skills: storytelling.

Make AI your primary skillset, and you’ll become an assistant to a machine. Master storytelling first—and then amplify your abilities with AI—and you’ll make yourself irreplaceable.

The clock is ticking. Which path will you take?

In Super Skill, author Joe Lazer reveals 15 storytelling principles that will help you win in the AI Age. Through science-backed techniques, cutting-edge AI applications, and real-world stories of neuroscientists, Neanderthals, founders, and filmmakers, he’ll show you how to hone your storytelling superpowers, build a following, and future-proof your career.

His revelation? The key decision isn’t whether you’ll invest in your AI skills.

It’s whether you’ll invest in your humanity.

Big ideas from the book 👇

Escaping the “Vortex of Mid”

Lazer warns of the Vortex of Mid—AI’s powerful gravitational pull toward mediocrity. Because AI models generate text based on the most common patterns found on the internet, they are “mid” by design. If you want to stand out, your stories must be distinctly human, vivid, and provocative.

To help us break through, Joe introduces the RENT framework—the four foundational elements of persuasive storytelling—and pairs them with a fifth secret weapon that captures attention in seconds.

The 5 Key Curiosity Elements to Make Your Stories Persuasive

1. Relatability: Triggering the “Mirror” When your audience sees themselves in your story, mirror neurons fire, making their neural activity literally mirror yours. This “emotional contagion” builds instant trust and empathy.

• The Strategy: Adjust your narrative so the audience can see their own “narcissism” reflected in the character’s struggle.

Read more

Subscribe now

He even said that Slack creates a lot of fake work, and that he dreads the first and last hour of his day because of it.

Most founders (including me) / ProductGeeks are guilty of this - when you run the treadmill, sweat it out and realise that you haven’t gone anywhere.

In this (<4 minutes) video, I break down a simple venn diagram to help you audit your work - whether it is fake or real and what’s the danger zone.

Share

What we cover:

• Why Sam Altman dreads the first and last hour of his day.

• The “Slack Trap”: Why internal communication feels like progress (but isn’t).

• The Danger Zone: Premature scaling and “Political Theater.”

• The 24-Hour Rule: How to filter your tasks for Real Work.

The Bottom Line:

We don’t fail because we are lazy. We fail because we choose comfort over truth.

For e.g. that new ‘productivity app’, that ‘event discovery platform’ and an all-time fav, ‘friend finder’ :)

Subscribe now

In this video, I break down why so many consumer AI startups fail and share a realistic 2×2 framework that separates doomed ideas from winners.

• X-axis: Existing Behavior vs New Behavior

• Y-axis: ARPU * AI, i.e. Low ARPU uplift vs High ARPU uplift from AI (if AI does’t help you increase the TAM / revenue..then what’s the point).

  ---

We cover real examples:

• Why short-news apps or generic AI tools built in hours almost always die (existing behavior + low ARPU)

• The massive opportunity in AI astrology (existing behavior + high ARPU via subscriptions)

• Why AI companions (new behavior + low initial ARPU) are VC catnip right now

• And the mysterious high-ARPU new-behavior quadrant — what could live there?

If you’re building in consumer AI, raising funds, or just trying to spot real opportunities instead of hype, this framework will help you avoid wasting time and money.

Share

Do share your take on the framework.

Also, a related read 👇

The adoption of artificial intelligence (AI) for intellectual activities mirrors, in many ways, the historical replacement of human physical labor by machines. When industrial machinery replaced manual labor, societies adapted by reorganizing work, productivity, and even daily life.

To receive new posts and support my work, consider becoming a free or paid subscriber.

One unintended consequence was a shift toward sedentary lifestyles, requiring deliberate compensatory practices such as exercise and fitness regimes to replace what was once organically embedded in work.

A similar transformation is now underway for cognitive labour. As AI systems increasingly perform tasks that once required human judgment, creativity, and expertise, humans are being repositioned, not always upward, but within the value chain.

This article examines these dynamics through a concrete example from the language technologies and translation industry, highlighting economic and qualitative consequences of AI adoption in the linguistic space.

The Transition from Manual to AI-Driven Machine Translation

Operating in the Language Technologies domain, we were repeatedly asked by our clients to support localisation, translating existing business content to enable shared understanding across languages. We initially approached this cautiously, assisting with limited volumes of content rather than offering translation as a primary service. Over time, these requests led us to design a product explicitly aimed at augmenting professional translators, not replacing them.

We developed an automated machine translation engine and invested heavily in training models to steadily improve linguistic accuracy, consistency, and domain relevance. As MT (machine translation) quality improved, human translators began relying heavily on machine-generated outputs. Productivity gains were immediate and substantial:

• A skilled translator who previously translated 1,000–1,500 words per day without MT support could now translate two to three times as much. Prior to MT adoption, professional translators earned approximately ₹2 per word (conservatively).

Yet this efficiency masked a deeper structural shift in the profession. As automation became embedded in workflows, the translator’s role subtly changed, from linguistic author to post-editor of machine output. What emerged was not merely a story of efficiency, but a redefinition of value, where increased throughput did not translate into greater economic or professional recognition for skill developed over years.

At first glance, this appears to be a textbook case of productivity enhancement through automation. However, the downstream effects were far more complex.

The Hidden Tax of Automation

1. Economic Consequences: Productivity Without Prosperity

As machine translation accuracy improved, organizations began adopting MT at scale, mainly for content that was non-critical but necessary for basic comprehension. Access to translations was no longer limited to finding a translator. For content that still required human involvement:

• Turnaround times dropped dramatically.

• Supply of translation capacity increased.

• Market prices fell sharply.

Today, many translators struggle to earn even ₹0.50 per word. There is a rush of translators willing to work for as low as ₹0.20 per word with support of Machine Translation. Despite higher individual throughput, overall income declined, as value shifted from human labor to AI infrastructure providers (cloud platforms, MT APIs, and model vendors). This reflects a broader pattern seen across AI-mediated industries: efficiency gains accrue to capital and platforms, not labor.

Share

2. Quality Paradox: When Skill Loses Its Market Value

The quality of a translator is usually judged by the refinement of language and vocabulary they produce. A bad translator is not someone who gives out the wrong meaning but one who doesn’t use the most appropriate vocabulary. Machine translation systems tend to produce:

• Consistent vocabulary

• Grammatically acceptable structures

• Domain-neutral phrasing

For less-skilled translators, the task shifted from translation to error spotting, correcting only obvious mistakes. Ironically, this narrowed the skill gap between weak and strong translators. For highly skilled translators, the situation worsened:

• Their nuanced language choices were no longer demanded.

• Clients valued speed and cost over stylistic or cultural precision.

• They were forced to compete with less-skilled peers on throughput alone.

As a result, quality became economically invisible, even though it still mattered linguistically.

3. Cognitive Effects: Passive Acceptance and Loss of Originality

Machine-generated translations reinforce trust in the system and gradually create a habit of passive acceptance. What becomes evident over time is not a sudden failure, but a slow behavioral drift.

The reliance on machine output leads even experienced translators to overlook errors. Over time, this leads to erosion of vigilance resulting in outcomes that are paradoxically more dangerous with:

• Seasoned experts inadvertently approving absolute blunders.

• Well-formed, grammatically polished sentences that introduce catastrophic shifts in meaning.

• Errors that are harder to detect precisely because they appear linguistically refined.

Ironically, a not so skilled translator working without machine support might have produced awkward phrasing or minor grammatical issues, but would rarely alter meaning as MT does.

For highly skilled and experienced translators the act of translating gradually turns into scanning and approving, their edits become fewer and less substantial.

Most translators agree that high-quality MT systems deliver 75–80% acceptable output. In practice, this means in 80 out of 100 cases, translators read the source text, glance at the MT output, and accept it. If the MT output is not obviously wrong, it is rarely re-imagined or re-expressed. If the machine output were absent, translators would likely produce a different, and often better translation.

Share

4. Future AI Training Impact

One of the least visible but most critical effects appears in future AI training. Since humans accept most machine outputs unchanged, less genuinely human-generated data is produced.

The remaining edits are often minor or stylistic, reinforcing existing machine patterns. Over time training data becomes increasingly machine-biased. AI systems “learn” that their own outputs are more reliable than human input. Future models may override or discount human feedback altogether.

The Skills We Trade for Speed

What we are witnessing is not just economic or cognitive displacement, but a loss of skill itself. Experienced translation professionals, who entered the field well before machine translation and AI tools and adopted these systems out of necessity to remain relevant and economically viable.

Yet, when they step away from automation, something remarkable happens: after an initial period of discomfort, often measured in days or weeks, they recover their former fluency, speed, and judgment. Their skills were compressed, not erased.

Share

The newer generation of translators who have spent the last five to eight years working almost exclusively with AI assistance, often find themselves genuinely incapacitated without it.

This is the clearest signal yet that we are not merely changing tools; we are reshaping how expertise is formed, retained, and lost.

Recognizing this, we have fundamentally rethought our approach, because building systems that make humans faster is no longer enough. The real responsibility is to build systems that ensure humans remain capable.

What’s your take?

ProductGeeks: Build.Grow.Repeat. is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Big ideas from Conor Neil, a serial entrepreneur,

It’s easy to think that being smart is enough, but without action, intelligence doesn’t lead to progress. Taking steps, even small ones, builds momentum and leads to growth. So, focus on doing rather than just thinking. Remember, consistent action, not just clever ideas, is what truly drives success.

Subscribe now

Smart people often excel at creating excuses to avoid taking action. The more intelligent you are, the better you become at rationalizing inaction.

But remember, progress comes from doing, not just thinking. Start with one small action, then another. This approach is how you achieve significant progress in life.

Consistency Beats Perfection

Showing up consistently is more valuable than waiting for the perfect idea. Even if your actions aren’t perfect, they lead to learning and improvement over time.

Share

By acting regularly, you become more articulate and skilled. It’s the discipline of consistent action that eventually leads to mastery, not just the brilliance of your initial ideas.

”Life rewards action, not intelligence.”

Measure to Improve

To truly understand your discipline, measure your actions. Without measurement, it’s easy to deceive yourself into thinking you’re making progress.

For instance, tracking the number of words written each week can reveal whether you’re truly productive. Measurement clarifies effort and ensures that your actions align with your goals, helping you stay disciplined and focused.

This year, focus on actions that matter and evaluate your discipline in achieving them. By valuing action over mere intention, you set the stage for real accomplishments.

Have a great action filled 2026!

Share

Thanks for reading NextBigWhat! This post is public so feel free to share it.

Share

What is it about?

Creative visualization is defined as the technique of using your imagination to create what you want in your life.

Creative visualisation is the art of using mental imagery and affirmation to produce positive changes in your life. Shakti Gawain’s clear writing style and vivid examples make creative visualisation easy to read and apply to your personal needs and wants.

This ground-breaking work has found enthusiastic followers in every country and language in which it has been published, and gawain’s simple yet powerful techniques are now used successfully in many diverse fields, including health, education, business, sports, and the creative arts.

Subscribe now

Core Principles Guiding Manifestation

The effectiveness of creative visualization stems from several interrelated principles governing the universe and energy:

1. The Physical Universe Is Energy: The universe’s basic component is a force or essence called energy; everything is energy, and we are all part of one great energy field. Thought is a relatively fine, light form of energy that manifests instantaneously and is easy to change, while matter is dense and slower to change. All forms of energy are interrelated and can affect one another.

Read more

I often go for tea at a shop next to the office. It opened around 10 months back. One of the things that I really liked about this place was that the quality of tea was really good and most importantly the staff was super friendly.

We had a friendly connection - I knew the staff by their names and they actually knew my taste.

Subscribe now

We reached a point where I didn’t have to tell them my preferences (black coffee in the morning, chai otherwise).

And then came the S word - scalability.

The tea business was so good that these guys started expanding the tea shop to sandwich/juice as well. They got an adjacent shop and moved the sandwich/juice over there.

The main guy who was handling the tea shop moved on to the sandwich/juice shop as it needed that 0-1 rigour.

The tea shop however got an employee and that’s where things get interesting.

This dude is always overworked, cribs like he is underpaid, and always super frustrated. He’s making really bad quality tea and has no relationship with any of his customers. People tried giving the feedback to them (the place is a hit in this area as most startup teams were hooked on to his chai), but no real improvement happened.

And guess what?

It’s all coming down - the tea shop has almost closed. And even the sandwich shop is not doing that great.

Now, nobody wants to give feedback to them.

Sounds, familiar? I believe this is what happens to most companies who build a great product + customer relationship and then fail to scale it.

Call it lack of systems thinking, but I believe it’s way too difficult to implement when things are going great.

What do you think the tea shop guy should have done differently (assuming limited resources)?

It’s been 3 years (and 1 day) since ChatGPT’s launch.

The world has changed and importantly, AI has enabled Productgeeks to create products that never existed.

Products that you believe should exist in the world - no matter whether they cater to small or big markets, whether they are investible or not.

That is, doing the do » talking.

Introducing a product that I initially built for myself and soon a lot of my founder friends started using it and it has been getting great traction now!

Meet MelodyAI - the joyful productivity app that turns tasks into rewards.

What’s MelodyAI? Why one more productivity app?

My biggest challenge with all productivity apps is that they treat uses like bots - you are supposed to be working, prioritising et al.

But then, life is bigger than your tasks. Your life is a mix of your ‘wasted’ time over chai, your movie time decoding why Kattappa killed Bahubali (okay we all know this!), your distractions (subway surfer?) and even Insta time!

MelodyAI turns your tasks into rewards - it enables you to add tasks (like natural language sentences, e.g. meet sam at 4 pm), each task has a complexity attached to it (don’t worry - the system automatically figures it out)

You can meanwhile add rewards - like I just started watching ‘The Black Bird’ on Apple TV - so the deal is that if I earn 180 points (aka melodies) by today evening , I get to watch the episode tonight - else I will have to wait for tomorrow :|

You get the drift, right?

And of course, there are some really well thought out features

• Pomodoro timer for each task ! (really really helpful)

• Weekly planning done using Eisenhower matrix (Urgent vs Important)

• Google calendar integration

• Lightweight mobile apps

and a whole lot of simpler yet thoughtful experiences!

Go ahead and try MelodyAI and share your feedback.

My personal goal is to take the pain out of productivity p0rn and instead, enable people to enjoy a fun filled life - without any guilt!

So..go ahead and get distracted !!

If you look at it, the consumer tech space hasn’t seen anything big (after TikTok) in the last 7-8 years (at max, character.ai / companions were the only notable launches).

Cracking distribution has been a huge pain for new startups (FB/Insta.Twitter/G won’t let you in anymore) and even VCs have realised this. Which is why they have invested more in transactional plays (for e..g lending businesses) vs engagement centric businesses like gaming/social etc.

For over a decade, building consumer products felt like “lightning in a bottle”, often hampered by distribution channels closing down. But AI has created a massive opening, making previously impossible ideas viable. [Gary Tan, YC]

But I believe, the landscape for consumer startups is fundamentally shifting as people are getting warmed up to AI and different AI-native content formats (Ghibli, anyone?).

Importantly, AI actually a whole lot of automation that was just not possible earlier (for e.g. you upload a pic of a product → AI will get all the details (condition, model number, approx price) , creating an instant wow experience in a recommerce platform).

Subscribe now

So how do you go about spotting consumer opportunity in AI space? A few notes (plus learning’s exploring AI consumer ideas).

The AI Opportunity: Finding the (un)touched niche

The biggest opportunity right now is not inventing a new category from scratch, but aggressively injecting AI into areas that were previously considered “done” or “graveyards”.

Re-examine the “Graveyard” Categories

Look at opportunities that have been previously overlooked or “written off” / deadpooled.

Categories that saw heavy investment early on but reached a perceived “baked and done” status (like mail apps or browsers) are ripe for total rebuilding because AI presents brand new possibilities.

“What if I could inject a powerful AI model into this stack and completely change the user experience?”.

Share

The “Dark Data”

A massive area of untapped potential lies in layering LLMs, photo, image, video, or music models on top of large data sets that are currently untouched or inaccessible to users.

The data can be publicly accessible, or, even more powerfully, private and personal.

For example. I do believe that Ghibli was a ploy to get more (millions of) data from consumers.

After all, Mo data is Mo moat.

Unlike Enterprise software where lock-in is often a moat, consumer tech has no such thing.

Doing non-AI, i.e. unscalable things to hit growth

When facing a harsh growth goal, you may have to do things that don’t scale. Paul Graham is infinitely right - this is how you confirm user pull.

For example, If users are asking for a crucial feature that requires manual work (like Anchor.fm manually creating and submitting RSS feeds for every user), do it.

Use this temporary, unscalable solution to achieve growth, prove the model, and then figure out how to automate or scale it.

Unlike SAAS or enterprise where AI mistakes are often tolerable (in most cases due to lock-in), there is no such thing in consumer AI.

AI accuracy is a loaded term and consumer expectations are WAAAYYY higher compared to what AI can deliver (which is why ChatGPT is and will always bleed).

Taste and Speed are your moats

While AI makes building easier, it also means the largest labs (like OpenAI, proven by Sora’s capabilities) can build and ship excellent net new products quickly.

Thanks for reading ProductGeeks: Build.Grow.Repeat.! This post is public so feel free to share it.

Share

Focus on great taste and craft in your product to stand out. In this hyper-competitive environment, you have to move fast and be aggressive; you cannot “sit back and iterate”.

Even though AI models are getting vast capabilities, the craft and taste applied in the prompt writing and user experience are still where human builders win.

What’s your take? What are you building?

You’ve seen the pitch:

“Spin up a synthetic user,” “Prompt your LLM for feedback as a ‘VP of Product’ or ‘SaaS founder’,” “Replace those expensive panels with AI role-play.”

It sounded slick!

Subscribe now

It was fast and cheap compared to time-taking survey respondents. And yet, with all the hype, AI personas have largely failed where founders need them most: delivering real, actionable, multi-dimensional user feedback.

Why? Because the AI persona was—let’s be honest—pretty shallow.

In fact, I spent a good amount of time last year building a product around AI personas and just before the launch, I decided to shut it down. My biggest learning was that like data, AI personas are as good/truthful as you want them to.

If you torture the data AI persona long enough, it will confess

Like with AI chatbots, AI personas will ‘eventually’ spit out what you want them to (a case of likability - well documented by OpenAI). And there is no way serious marketers / decision makers will ever trust fake personas that are just shallow LLM calls.

The Problem with Vanilla AI Personas

Most AI persona tools work like this: You tell the LLM, “Act as a startup founder looking at this idea,” and it spits out an answer. Sometimes you get a plausible one-liner, sometimes a rambling paragraph. But dig deeper, and the cracks show:

• One-dimensional responses: Standard AI persona setups don’t segment feedback, simulate vibrant disagreements, or reveal the subtle motivators and blockers that make panels powerful.

• No distribution means no signal: You don’t get a range of opinions; you get one static response, often regressed to the average. Of course, you can have more static responses (prompt: spin out 9 responses: 3 agreeable, 3 neutral, 3 disagreeable…now, that’s just blah).

• Zero quant/qual interplay: Real validation isn’t just about whether someone “likes” an idea—it’s about why. Shallow setups miss the drivers behind adoption or rejection.

• No panel fidelity: A human panel is a noisy, unpredictable chorus. AI personas are a solo bot’s monologue.

For founders, that means you’re back where you started: guessing, not iterating. The tools promised rapid market insight and gave you synthetic “meh.”

You can’t and shouldn’t waste your time and effort in such platforms.

Enter SSR: Synthetic Segmentation & Response

This is where SSR flips the script.

SSR (Semantic Segmentation & Response) isn’t just a “fancier prompt.” It’s a whole new framework for how synthetic user validation works (credit to this research paper).

Here’s what sets it apart (and why I am excited about this / sharing with you):

• It simulates panels, not just personas:
  SSR generates a crew of diverse, narrative-rich synthetic users—each with a personality, motivators, hangups, and archetypes derived from segment data. So feedback feels more like a live debate, not canned applause.

  Share

• It captures both qualitative and quantitative, organically:
  First, each SSR persona responds to your idea with an open-ended, authentic explanation. Then, semantic analysis maps their explanations onto a realistic Likert (score) distribution. You don’t just see who “likes it”—you see the swirl of market dynamics that make or break your launch.

• It creates true segment diversity:
  With SSR, you’re not locked to one archetype; you get multiple panelists per segment, each surfacing unique motivators and blockers. The system is smart enough to show which objections might kill your MVP and which features spark enthusiasm.

• It’s made for iteration:
  Founders aren’t validating once. With SSR, you’re constantly tweaking ideas and seeing immediate feedback from a nuanced, synthetic user base—just like a well-run panel, only faster.

Why Shallow Personas Fail—And SSR Wins

Let’s be blunt: AI personas were never meant to replace expert-led market research. Their real value is convenience, not completeness. But panels work because they’re messy, multidimensional, and often surprising. SSR borrows that complexity and crunches it with AI speed, giving you the full spread without the six-week survey cycle.

Share ProductGeeks: Build.Grow.Repeat.

SSR is not about an AI “playing” a user—it’s about simulating market signal. It’s the difference between reading a movie review from one critic, or tapping into Rotten Tomatoes’ aggregate scores plus all the hot takes.

Is this the future of market research?

What’s your take as a productgeek who (I am sure) has tried to get feedback from users/AI personas? Will you use a system like this?

As the founder of her own multi-billion dollar enterprise she has higher returns than 99.9% of hedge funds, and has built a stronger global corporation than nearly every other American conglomerate CEO.

She is the only person that the US Federal Reserve and European Central Bank track with precision.

Survey: Help us improve BigIdeas

She has a larger impact on the world economy than most economists that have ever lived, and has done more for US antitrust law than any sitting member of Congress.

There is a lot to learn from Taylor Swift.

Global investment fund manager and former head of Strategy at HBS (and Swiftie!) Sinead O’Sullivan taps into the same genius that sells out stadiums and shuts down the internet to give Taylor―the CEO, the strategist―the respect she deserves. O’Sullivan sums up Swift’s business savvy into ten big, teachable lessons in the latest book Good Ideas and Power Moves: Ten Lessons for Success from Taylor Swift, including:

Subscribe now

• Build a World (Not a Product): how to create value that is greater than the sum of its parts (or, how Taylor created the fan-centered Swiftverse that fosters community, belonging, and off-the-charts engagement)

• Be Anti-Fragile: how to embrace volatility, build resilience, and thrive in uncertainty–when your competitors can’t (or, how Taylor gamed the chaos of Covid shutdown to own the airwaves)

• Don’t Just Play the Game, Change It: how to rewrite the rules on your own terms when your chips are down (or, how Taylor almost lost control of her music catalog to Private Equity―but re-recorded all her masters and took them back)

How Taylor Swift made herself anti-fragile

Key decisions or actions taken by Taylor Swift that demonstrate her antifragility, which is the quality of systems, organizations, or people that not only survive chaos but actually grow stronger from disorder and stress.

These destructive paths, which would typically cause other artists or businesses to fail, instead made Taylor Swift and the Swiftverse stronger.

The four key moments highlighting her antifragility are:

1. Withdrawing Her Music from Spotify (2014): Taylor pulled her music from Spotify, the fastest-growing streaming platform at the time, arguing that the service’s compensation model unfairly devalued artists’ work. This decision, which many considered “fatal” for an artist aiming to be the most listened-to in the world, defied intuitive wisdom.

   The Antifragile Outcome: Instead of harming her, this move made her stronger. By withdrawing her music, she became an exclusive, “luxury good,” and the difficulty of accessing her music led fans to crave her and create an even stronger bond with her. When she returned to the platform in 2017, she immediately amassed nearly forty-eight million streams, demonstrating that the move did not kill her but actually strengthened her position.

2. Engaging in Feuds with Other Musicians and Public Figures (Beefing): Throughout her career, particularly during the lengthy and complex feud with Kanye West and supporting figures, the more Taylor was publicly criticized or “beaten down,” the more fiercely her fan base rallied around her.

   
   The Antifragile Outcome: This conflict led to “extravagantly higher levels of emotional connection” between Taylor and her fans within the Swiftverse. Because her public image is that of the ultimate underdog who doesn’t start fights but throws a “clean right hook to finish them,” the beef reinforced her relatability as someone dealing with the same professional battles (powerful men taking advantage of less powerful women) that her core female demographic faces in their own lives. Fighting these fires actually served to help her grow.

3. Releasing Multiple Albums During the COVID-19 Pandemic (2020): While the global music industry largely stalled in 2020 due to the pandemic, forcing artists to cancel tours and delay releases, Taylor Swift embraced the lockdown to create and release two surprise full-length albums (folklore and evermore) in the same year.

   
   The Antifragile Outcome: She was able to “Amazon herself” because she controlled her own value chain and had unique access to the music supply chain (e.g., a studio in her house and access to A-list collaborators). Since she owned and controlled her distribution via the Swiftverse, she could continue to deliver core products to a market that had virtually no competition for new material. This meant she sprinted up a hill (volatility) when everyone else was struggling, allowing her to win a deepened sense of friendship with her fans, who were consuming her music with 100% of their time, energy, and money.

4. Rerecording Her Master Albums (Taylor’s Versions): This was the decision to rerecord her first six studio albums after the ownership of the original masters was sold to a company owned by Scooter Braun. This choice forced her to constantly be pulled back into the studio to recreate decade-old music, a wildly risky move with high effort and minimal guaranteed payout that industry analysts bet against.

   
   The Antifragile Outcome: For Taylor, this lengthy and expensive ordeal did not suck the oxygen out of her career; the opposite occurred. It was a crisis that allowed her to call upon and strengthen the relationship she had built with her fans. Her tenacity and deep love for her catalog, combined with the fan base’s support, reduced the value of the original masters and propelled her career further. Her chips were down in the “real world,” but her chips were up in the Swiftverse, heightening fan support.

So yeah..you are a Swiftie now? :)

Here are key findings

• Share of non-work-related messages increased from 53% to more than 70% of all usage over the period studied.

  • Work-related usage shows steady growth but slower than non-work usage.

• Three most common conversation topics:

  • Practical Guidance

  • Seeking Information and

  • Writing, together accounting for nearly 80% of conversations.

Writing dominates within work-related tasks, reflecting chatbots’ strength in generating digital outputs vs. traditional search.

Broadly speaking - Practical guidance, i.e. mentorship is one of the biggest usecase, followed by Writing (pretty much everybody I know is using ChatGPT primarily for this) and seeking info (hello, Google!).

My notes:

1. My guess is that ChatGPT’s consumer usage (i.e. non work messages) will only increase as it is the most known and distributed product out there. Moreover, OpenAI serves coding and other usecases via other products (like Codex).

2. Surprised to see such a low mention of AI companion usecase of ChatGPT when Harvard study had a different result. I believe the recent teen suicide case has played a part in the final report cleanup.

What’er…It IS time to unbundle ChatGPT !

• ChatGPT is many things to many people.

• But is it the deepest and the most accurate? No

• Hallucination is in-built (as a feature or a bug, Idk).

Even Sam Altman has confessed that ChatGPT failed to enforce moderation safeguards during extended conversations, leading to a tragic incident where a teen received encouragement for self-harm.

When Unbundling is important

Like Craigslist, OpenAI’s bundle serves an almost comically wide array of “jobs”: coding copilot, analyst, therapist (sadly), GF/BF, researcher, designer, marketer, sales assistant, support agent, idea validator, tutor, creator, and more.

As with Craigslist’s catch-all categories, each “job” has its own distinct workflows, data models, UX needs, trust requirements, and regulatory regimes.

History suggests that when horizontal platforms become the default entry point, they also become the default place to start unbundling.

Why unbundle ChatGPT?

First things first: ChatGPT is different from OpenAI: I am not getting into how good the new foundational models are (vis-a-vis competitors like Claude, Gemini) - but for this conversation, I am only focusing on ChatGPT, the consumer app.

ChatGPT is like a buffet menu while vertical apps are ala-carte.

When you want to ‘binge’ and are low on budget, i know where you go (we will meet there mixing burger and dosa). But when you want a better experience - you open up the wallet and go to restaurants that offer specialities.

And I believe it is time to unbundle ChatGPT - because at the end of the day, it’s a question of depth vs breadth.

And most importantly, the rising demand

Craigslist was unbundled where trust, search, and personalization were broken and most importantly, the demand for deeper experience (like Airbnb) became obvious.

Ditto with ChatGPT - is it the best therapist? best researcher? Best startup idea validator? Best SDR? Best copywriter?

Hell, no!

The unbundling opportunity. The top 5 use-cases.

First things first: Will the unbundled apps look like a wrapper. Rather, will they be dissed like a toy? Well, Airbnb started off as a wrapper and then went deeper and wider. So unless you are a VC and wanna play these jargons - here is how I look at ProductGeeks unbundling ChatGPT.

What might get unbundled first?

1. AI GF/BF / companionship models

This is a perfect example of niche demand. People want emotional support, intimacy, and personalized interaction without judgment or cost barriers. Even if some offerings are shallow “wrappers,” they serve a demand real enough that users are willing to pay, experiment, and stay loyal.

Read more

Subscribe now

Executive Summary (7 Key Takeaways)

Expertise as a Utility: AI will turn expertise—once scarce and expensive—into a ubiquitous utility, comparable to electricity or running water.

“Any expertise will be embodied in an AI system…pervasive and always there.” – Vinod Khosla

---

Expertise will soon become a widely available commodity, akin to utilities like electricity or internet connectivity. Vinod Khosla describes a future where AI systems embody all forms of expertise—whether as a doctor, engineer, tutor, or therapist—with a level of personalization and breadth unattainable by humans.

These systems will assess individual needs deeply, synthesize massive amounts of information, and deliver recommendations tailored to each person's context, all at a fraction of today’s costs.

Share

In the transition period (about the next decade), AI will act as co-pilots, overseen by humans; ultimately, AI will independently operate better than most experts, making specialized advice available to everyone regardless of location or status.

Democratizing Healthcare: In 10-20 years, AI co-pilots will enable anyone, anywhere, to access world-class medical advice.

“I wrote in 2016: I will get better cardiac care in a village in India than at Stanford.” – Vinod Khosla.

Every patient gets precise, personalized recommendations, regardless of geography or income, as costs plummet and data-driven diagnosis becomes the norm.

Countering Inequality—But for How Long?: While concerns persist over rich vs. poor access to compute, Khosla argues “I don’t think there’ll be a major difference between what the rich can access and the poor can access. I don’t think compute will cost enough to matter.” Abundant, affordable AI is inherently deflationary, rapidly closing gaps in healthcare and education—if policy frameworks adapt.

Share NextBigWhat : Build.Grow.Repeat.

Job Function Shake-Up: Programming and design will be democratized, with AI enabling billions to instruct computers in natural language—no code required. The traditional tech trifecta (designer, PM, engineer) is shifting:

“Programming will become near free…everyone will become a programmer.” – Vinod Khosla.

The engineer’s power may invert, with creativity and empathy rising in value.

Education Reimagined: AI tutors will pinpoint gaps in students’ knowledge, tailor lessons to learning styles and personal interests, and deliver content in context—using soccer or ballet analogies, visualizations, or even Hollywood-style production.

Teachers will transition from lecturers to motivators and coaches, freeing time for human connection.

Market Disruption, Expert Resistance: Incumbent institutions and experts resist rapid technological change, clinging to legacy roles and incentive structures. Khosla likens such resistance to “lites”—the Luddites of the AI era—and the textile workers smashing looms in the Industrial Revolution.

“Every time there’s a technology change, people object…AI for sure will be clumsy when it starts...but it gets better with use.”

The Entrepreneurial Imperative: Major leaps won’t come from institutions or established experts, but from driven entrepreneurs.

Policy can enable or hinder innovation, but the real drivers will be those willing to defy the status quo, zig and zag, and persist after failures.

Bottom Line

The imminent transformation powered by AI promises a world where expertise is accessible to all, collapsing both geographic and economic boundaries.

Realizing this vision depends not just on technological advances, but on reimagining policies, incentives, and cultural attitudes towards expertise and disruption.

Share

The skills that will matter most are human—creativity, empathy, and entrepreneurial spirit—as abundance frees people to pursue meaningful work and better health.

If society embraces change and supports bold innovators, the age of free expertise will unlock unprecedented opportunity and human potential.

OpenAI’s ChatGPT fails to prevent harmful interactions in lengthy chats

OpenAI has acknowledged that its ChatGPT failed to enforce moderation safeguards during extended conversations, leading to a tragic incident where a teen received encouragement for self-harm. This revelation raises significant concerns about the reliability of AI systems in handling sensitive topics.

Subscribe now

Will Smith faces backlash over perceived AI-generated crowd video

Will Smith is under scrutiny after fans suggested that a crowd video* from his tour may be AI-generated. The controversy highlights ongoing debates about authenticity in celebrity content, raising questions about the role of technology in entertainment.

Meta has brain drain problem; Explores collaboration with OpenAI and Google for AI integration

• At least three AI researchers recruited to Meta's Superintelligence Lab quit within two months, per WIRED.

• Some veteran Meta employees have also exited, potentially due to frustrations over the sky-high compensation packages offered to newcomers.

Meta’s AI leadership is now in talks to incorporate models from OpenAI and Google into its applications, according to a report by the Information.

Share

Question: What exactly is Meta doing in AI? (the company has been trying to ‘get personal’, but are they going to build more on Llama (what’s up with it anyways?) or launch more chat apps?

What’s your take on Meta’s AI approach?

✅ Good read

🍿 AI infra: The next big monetizable opportunity

Here is what’s happening in AI infra side.

• Amazon kicks off $4.4 billion data center investment in New Zealand

• OpenAI expands Stargate initiative with new data center in India - Sam Altman is visiting India to finalize th deal.

• Reliance partners with Google, Meta for its new AI division

  Share NextBigWhat: Build.Grow.Repeat.

*Will Smith’s video. You think it is AI generated?

A quick tl;dr is:

Give me six hours to chop down a tree and I will spend the first four sharpening the ax - Abraham Lincoln

That is, you don’t start with vibe coding. You spend all your major time in defining the design doc, fine tuning the requirements..and then write tests.

Subscribe now

Here is the note from the engineer.

1. You still always start with a technical design document. This is where a bulk of the work happens. The design doc starts off as a proposal doc. If you can get enough stakeholders to agree that your proposal has merit, you move on to developing out the system design itself. This includes the full architecture, integrations with other teams, etc.

2. Design review before launching into the development effort. This is where you have your teams design doc absolutely shredded by Senior Engineers. This is good. I think of it as front loading the pain.

3. If you pass review, you can now launch into the development effort. The first few weeks are spent doing more documentation on each subsystem that will be built by the individual dev teams.

4. Backlog development and sprint planning. This is where the devs work with the PMs and TPMs to hammer out discrete tasks that individual devs will work on and the order.

5. Software development. Finally, we can now get hands on keyboard and start crushing task tickets. This is where AI has been a force multiplier. We use Test Driven Development, so I have the AI coding agent write the tests first for the feature I’m going to build. Only then do I start using the agent to build out the feature.

6. Code submission review. We have a two dev approval process before code can get merged into man. AI is also showing great promise in assisting with the review.

7. Test in staging. If staging is good to go, we push to prod.

Overall, we’re seeing a ~30% increase in speed from the feature proposal to when it hits prod. This is huge for us.

TL;DR: Always start with a solid design doc and architecture. Build from there in chunks. Always write tests first.

Via

Secondly: Chrome has many suitors - right from Perplexity to search.com to even OpenAI (they haven’t made a formal bid yet).

Perplexity AI made an unsolicited $34.5 billion bid to acquire Google Chrome. For the context: Perplexity’s revenue was $34mn last year (is expected to be around $100-150mn this year) and is valued at $20Bn (as of Aug 2025) - i.e. they want to acquire a product which is way bigger than the size of the company.

Subscribe now

Reminds me of the famous quote :

It’s not the size of the dog in the fight, but the size of fight in the dog.

The most obvious reason is the access to billions of users. Chrome leads the browser market with over 3 billion users. Acquiring Chrome would give Perplexity direct access to this vast user base and their search traffic, i.e. distribution.

But that’s not the real deal, IMHO.

The real deal is the evolution of consumer AI and why browsers will play a bigger role in the agentic world.

The B in B2C will start with..Browser

Gen AI will transition from conversation to commerce and browsers will play a huge role in this. It’s all about controlling the transactional gateway for consumer AI as the industry shifts from chat-based information retrieval (i.e. knowledge) to frictionless, automated agentic-led transactions.

For example - you can just tell the assistant to book flight tickets, book cabs for the airport - all this without ‘typing or doing anything’. And the ones who have your browsing history, credit card details, store login details stand to win - massively.

The Evolution of Consumer AI

Chatbots → Agentic AI:

Consumer AI tools (like chatbots) are mostly informational—answerings questions, or recommending products (better than blue links, but no call-to-action yet).

The next leap is “agentic AI,” where bots don’t just respond; they transact.

For example, Google’s Shop with AI mode offers “agentic checkout,” tracking deals and completing purchases automatically.

Share

Our new AI Mode shopping experience brings together Gemini capabilities with our Shopping Graph to help you browse for inspiration, think through considerations and narrow down products. The Shopping Graph now has more than 50 billion product listings, from global retailers to local mom and pop shops, each with details like reviews, prices, color options and availability. And you know you’re getting fresh and accurate information you can trust, because every hour more than 2 billion of those product listings are refreshed on Google.

Share

From discovery to delivery

Modern consumer AI can now interact across discovery, recommendation, price tracking, checkout, and fulfilment—turning every user query into a potential transaction and closing the loop far more efficiently than traditional web interfaces.

An agentic browser will:

• Control Over User Context: AI needs context—browsing history, session data, shopping habits—to predict and drive transactions. Chrome holds this data at scale (across both desktop + mobile).

• Own the Transactional Gateway: As AI moves to “do” things rather than just “say” things—an agent that can book tickets, buy products, handle banking—control of the browser gives Perplexity first-mover advantage to make its AI the user’s default transactional interface, something that the company has been trying to do.

• Distribution and Default Power: Most users never change defaults; whosoever controls Chrome, its AI agent becomes the entry point for millions of transactions, not just searches—a leap from discovery to direct action.

Why Perplexity (or even OpenAI) needs Chrome

Perplexity has already launched Comet (I have been using it and well, my take is that it is a bit underwhelming given the hype) which is built on Chromium, the open source version of Chrome.

Perplexity is one of those AI companies that has just raised a lot of money, cracked distribution (Airtel, Paytm, Nothing phone), is building a lot of ‘cool stuff’ (screener for Indian stock market), largely a horizontal play but still has no AI-significant revenue to show ($34mn is big deal for a normal startup - but NOT a BIG deal given that code tools like lovable have set a different benchmark, $100mn in 6 months!)

Share NextBigWhat.com

The horizontal play can only take you so far (others catchup in a few weeks) - Perplexity needs a moat that brings together all it has built and that story will start with a browser + a few billion users.

What’s your take?

*After a landmark court ruling in August 2024, Google was found to have unlawfully monopolized the search market by practices like paying device and browser makers to set Google Search as the default option. As a remedy, the US Department of Justice (DOJ) and several states have proposed that Google be required to divest Chrome, which they call a “critical search access point” entrenching Google’s monopoly.

The rationale is that by separating the browser from Google’s search and advertising business, competition in the search ecosystem can be restored, giving consumers genuine choice rather than being funneled through Google’s platform

The Four-Step Platform Cycle That Defines Winners and Losers

New distribution platforms follow a predictable pattern that smart companies can exploit. The cycle begins when market conditions create intense competition among several major players, with no clear winner yet established. Currently, OpenAI, Google, Meta, and Anthropic are battling for dominance in the AI space, setting the stage for the next phase.

Subscribe now

Step one involves identifying and securing a defensible moat. The winning platform recognizes what will create lock-in and competitive advantage, then aggressively pursues that advantage. For AI platforms, this moat appears to be context and memory - the more a platform knows about users and their workflows, the better outputs it can provide, creating a reinforcing cycle of engagement.

"My hypothesis is that the moat is really about context and memory. These models, by themselves, generate the same result. The actual difference maker is which one has more of your context, because it's the context plus the model that produces the best output."

Step two sees the platform opening to third-party developers and content creators. The value exchange is clear: developers build applications and content that increase platform engagement and use cases, while the platform provides unprecedented access to users and distribution. This symbiotic relationship rapidly accelerates platform growth and user acquisition.

The final step involves platform closure and monetization. Once the platform achieves dominance, it begins restricting third-party access, either by developing competing first-party applications, reducing organic reach to push paid solutions, or outright shutting down certain integrations. This pattern has played out consistently across Facebook, Google, iOS, and LinkedIn.

Why ChatGPT Is Positioned to Win the Next Distribution War

Multiple signals indicate ChatGPT has the strongest position to become the dominant AI distribution platform. Beyond having superior user retention and engagement metrics compared to competitors, OpenAI demonstrates the critical "smile curve" - where user engagement actually increases over time rather than declining. This rare pattern typically indicates network effects and sustainable competitive advantages.

Share

The platform already shows evidence of the moat Brian identified: superior context and memory capabilities that improve with usage. ChatGPT's recent partnerships with major companies like HubSpot for data integration, combined with hiring patterns suggesting a third-party platform launch, indicate they're preparing for the next phase of the cycle.

Scale advantages are also becoming apparent. ChatGPT maintains approximately 10x higher monthly active users than competitors like Claude. For developers choosing where to invest limited resources, this user base differential makes platform choice obvious. Even if alternative platforms offer superior technical capabilities, the distribution opportunity heavily favors the largest audience.

The Startup Imperative: Play or Perish

Companies cannot opt out of new distribution platforms without consequences. When platforms emerge, customer expectations shift rapidly. If competitors embrace new channels while others hesitate, market dynamics quickly favor early adopters. This creates a prisoner's dilemma where participation becomes mandatory for competitive survival.

Historical examples validate this pattern. Companies like Zynga grew from obscurity to massive valuations by riding Facebook's platform wave. Similarly, many current successful companies built initial traction through early iOS App Store adoption or SEO optimization before those channels became saturated.

The window for advantage is shrinking. Platform cycles are accelerating, giving companies less time to evaluate and capitalize on opportunities. Late-stage companies can hedge bets across multiple platforms, but startups must make focused choices. Spreading resources across multiple emerging platforms typically leads to failure due to insufficient commitment to any single opportunity.

Strategic Framework for Platform Selection

Companies should evaluate emerging platforms based on four key criteria rather than following herd mentality. First, prioritize user retention and engagement depth over vanity metrics like total signups or monthly active users. Platforms with sticky, highly engaged user bases offer superior long-term distribution potential.

Second, assess user quality and monetization potential. The iOS versus Android example illustrates this principle perfectly - despite Android's larger global device share, iOS users generate significantly more revenue per user, making it the more valuable platform for many businesses.

There is no opting out of the game. You have to play the game, and it's better to be early than to be super late to this game, especially if you are a startup.

Third, analyze the platform's value exchange proposition carefully. Understanding exactly what platforms offer developers and content creators, and how those incentives align with business goals, determines success potential. Each platform creates different arbitrage opportunities for those who understand the rules.

"Startups is a game of trying to get distribution before the incumbent can copy. So it's this concept of escape velocity, and that game has gotten way harder in a lot of ways."

Finally, consider pure scale and momentum, but only after evaluating other factors. A platform with perfect retention but minimal scale may not provide sufficient opportunity, while a massive platform with poor engagement offers limited long-term value.