They want less work.

They want momentum.

They want to get something done.

That is the mistake a lot of founders are still making.

They are mistaking a breakthrough in capability for a breakthrough in demand. Just because a model can generate cinematic video from text does not mean millions of consumers have a recurring desire for “cinematic video generation.” That is not a product insight. It is a technical milestone.

People do not want AI nearly as much as they want progress.

This sounds obvious in hindsight, but it is a lesson the market has to keep relearning every time a new technology wave arrives.

Subscribe now

A model can now generate cinematic video, realistic scenes, stylized content, characters, or entire visual worlds from text. That is technically extraordinary. But the existence of a powerful capability does not automatically imply a large, durable user need around that capability. More often than not, it just creates a new class of demos.

And this is exactly why Jobs To Be Done (JTBD) still matters so much. It remains one of the few frameworks that consistently forces builders to ask the right question:

What job is the user actually hiring this product to do?

And if you miss that distinction, you can end up building something magical that nobody actually needs often enough.

Leave a comment

The market rewards momentum, not novelty

Most consumers are not looking for tools.
They are looking for movement.

They do not wake up wanting “AI video generation.” They wake up wanting to make an ad faster, create a visual without hiring someone, send something funny to a friend, publish content without overthinking it, or avoid work that feels tedious, confusing, or creatively intimidating.

Those are jobs.

Share

“Generate a realistic 12-second clip from a prompt” is not a job. It is a technical feature that may or may not attach itself to one.

That difference is where a huge number of consumer products go wrong, especially in moments like this when the underlying technology is moving so quickly that it is easy to mistake possibility for inevitability.

Sora created video

Founders naturally over-index on what is newly possible. Users do not. Users care about whether a product helps them make progress in a way that is meaningfully better than whatever they are doing today. Faster. Easier. Cheaper. Less cognitively expensive. More confidence-inducing. More socially useful.

That is what creates retention.

Without that, you might still get excitement. You might get screenshots, social sharing, a burst of curiosity, maybe even a strong launch. But you do not get habit. And in consumer, habit is what matters.

This is why so many AI products feel exciting in week one and irrelevant by week six. They are built around what the model can do, not around what the user is repeatedly trying to get done.

That is not a model problem.
It is a product problem.

Share ProductGeeks/ Ashish Sinha

JTBD remains the cleanest way to see what is real

A lot of product frameworks become less useful the more often they are repeated. JTBD has survived because it still explains actual user behavior better than most alternatives.

Users are not “using your product.”
They are hiring it.

And they are hiring it to make progress in one or more dimensions:

• functional

• emotional

• social

That has always been true.

Share

People did not hire Instagram to “share photos.” They hired it to project identity, taste, status, and belonging. People did not hire TikTok because they wanted “short-form video.” They hired it because it is one of the most efficient boredom-killing and attention-capturing products ever created. And people are not hiring ChatGPT because they want “AI.” They are hiring it because it removes friction from writing, thinking, planning, researching, summarizing, deciding, and getting started.

That is the actual wedge in consumer AI: not intelligence, but effort compression.

This is the point many AI-first products still miss. They are designed around what the model can output, rather than around what the user is trying to avoid.

But the best consumer products rarely ask users to admire the technology (TikTok is the OG when it comes to using AI to personalise your feed..but that tech part was only limited to people interested in tech, not the normal consumers)

The best AI products make the technology disappear into progress.

Share

That is a much harder thing to build. It is also the thing that actually lasts.

People rarely know what they want. But they are constantly telling you what they hate.

This is where the Henry Ford quote, cliché as it is, still points to something important:

“If I had asked people what they wanted, they would have said faster horses.”

Consumers are usually poor at describing the solution.
They are much better at revealing the friction.

They will not tell you they want “an AI-native multimodal creative operating system.”
They will tell you (as is obvious from Reddit discussions):

• “I need this ad by tonight.”

• “I don’t know how to make this look good.”

• “I hate starting from scratch.”

• “Can this just do it for me?”

That is where product insight lives.

The best founders do not listen literally. They listen structurally.

They understand that users may not be able to imagine the right product, but they are constantly exposing the jobs they are trying to complete and the friction that is blocking them. That is usually where the real opportunities come from.

And it is also why so many “AI for creativity” products are at risk of overestimating the size of their market.

AI lowers the cost of creation. That does not mean everyone becomes a creator.

This is one of the most important questions in consumer AI right now, and I think the market is still underestimating how much it matters.

A huge number of AI-native consumer products are being built on an implicit assumption: if creation becomes dramatically easier, then a much larger percentage of people will become creators.

That sounds plausible.
It may also be wrong.

Historically, creation has always been much smaller than consumption (aka 1% rule, i.e. 1% create, 10% remix and 89% consume), even when tools become significantly easier to use. Most people do not want to create consistently. They want to consume, react, lightly remix, and occasionally produce something “good enough” with very little effort.

That was true before AI.
There is a very good chance it remains true after AI.

AI absolutely lowers the cost of making things.
But lowering the cost of making does not automatically increase the desire to make.

That distinction is subtle, but incredibly important.

Because if your product thesis depends on millions of users suddenly wanting to create on a regular basis, you may be building on top of a behavior that never actually changes.

Why products like Wabi may struggle

This is also why I am skeptical of products like Wabi (raised $20mn seed round, is a Silicon Valley darling) and the broader class of AI-native creative consumer apps.

That does not mean they cannot get traction. They can. In fact, many of them probably will. These products are highly legible to early adopters, creators, and tech-native users. They are expressive, socially shareable, and often fun in exactly the way that drives initial engagement.

But that is not the same thing as building a durable consumer business.

YouTube has “billions of monthly logged-in users”, over 20 million videos uploaded daily, and Shorts alone now averages 200B+ daily views

Because the underlying creator-versus-consumer dynamic may not actually shift very much with AI. If anything, AI may simply increase the supply of content dramatically while leaving the demand side structurally similar.

That is a dangerous asymmetry.

More supply does not automatically create more meaningful demand.
It often just creates more noise.

And what that likely leads to is not a creator revolution, but a more automated content economy:

• a small number of highly leveraged creators

• a wider layer of casual remixers

• and a very large base of users who still mostly consume

If that is directionally right, then many AI-native creative apps will eventually run into the same wall:

Creation alone is not enough.

To become durable, they probably need to become one of two things:

1. A utility

Something that helps users complete a real, recurring job they already care about

2. A distribution layer

A place where people actually come to consume, not just create

The second is much harder than it sounds.

Because AI lowers creation cost.
It does not lower distribution cost.

And in consumer, distribution remains the bottleneck.

It still decides what wins.

ProductGeeks/ Ashish Sinha

How to find opportunities in consumer AI?

So far, most of AI success story has been on SAAS or prosumer space, but I believe the next big opportunity will lie in consumer AI…

Read more

4 months ago · 2 likes · Ashish Sinha

The best AI consumer products will not feel like creativity tools. They will feel like relief.

This, to me, is the bigger lesson.

A lot of AI consumer products are still being built as if the opportunity is to give users more possibility. More outputs. More optionality. More expressive range. More things to generate.

But possibility is rarely the bottleneck. Completion is.

What most users are missing is not another blank canvas. It is fewer blocked moments.

They do not need more things they could make. They need fewer reasons not to finish.

That is why the strongest AI products will likely be the ones that remove:

• hesitation

• setup friction

• execution anxiety

• taste uncertainty

• complexity

• first-step paralysis

In other words, the winners will not be products that simply help users create more.

They will be products that help users move forward.

That is a much better product category.
And usually, a much better business.

tl;dr

Sora’s shutdown is not really about Sora.

It is a reminder of a much older truth that every technology cycle eventually rediscovers:

They do not want AI for its own sake.
They want momentum.

They do not want infinite creative power.
They want to get something done with less friction.

And they definitely do not want another product that turns an impressive technical capability into one more thing they now have to learn, configure, and actively use.

That is why JTBD still matters.
Not because it is fashionable, but because it keeps pointing builders back to the only question that has ever really mattered in consumer:

What painful, recurring job can this product make easier, faster, or disappear entirely?

That is where the real companies get built.

Everything else is mostly a demo.

ProductGeeks/ Ashish Sinha

90% of consumer AI ideas will die (use this 2 X 2 to validate)

90% of consumer AI ideas look sexy but crash hard — most end up in the “tar pit” with zero real traction or revenue uplift…

Read more

2 months ago · 6 likes · Ashish Sinha

What consumer products are you building or consuming? Do share in the comments section

Subscribe now

Big qn: Is everyone running an (AI) treadmill? Going nowhere, but exhausting tokens? Read on 👇

The RevenueCat State of Subscription Apps 2026 report* highlights several unique challenges and opportunities for AI-powered subscription apps beyond their higher initial revenue.

These insights reveal why AI apps often fail to sustain growth despite strong early traction.

Revenue Premium vs. Churn Penalty
AI apps generate 41% more revenue per paying user than non-AI counterparts, largely because tools like ChatGPT have normalized $20/month pricing.

However, they churn 30% faster on average, with annual retention at just 21.1% compared to 30.7% for non-AI apps.

This structural issue stems from hype-driven acquisition where users cancel after short-term needs are met.

Share

High Marginal Costs Drive Monetization Shifts
Unlike traditional apps with near-zero serving costs, AI features powered by LLMs create significant variable expenses per user.

Leave a comment

As a result, AI developers are shortening free trials, adopting hard paywalls, pushing annual plans, and adding premium tiers specifically for AI capabilities to maintain profitability. This has led to less generous freemium models across the category.

Flood of Vibe-Coded AI Apps

AI-assisted tools have unleashed a wave of rapid launches - 14,700 new subscription apps in January 2026 alone, many AI-focused.

These “vibe-coded” apps accelerate market saturation, inflating customer acquisition costs (CAC) and worsening overall churn as undifferentiated products compete.

Success now demands workflow moats beyond basic generation features.

---

Retention Warning Signs
Watch Month 3 - 6 metrics closely, as the AI revenue premium evaporates without sticky value. Many

AI apps lose 79% of annual subscribers before Month 12 due to unmet long-term expectations.

AI Apps Command High Churn
Artificial intelligence applications are demonstrating a unique monetization profile in the current market.

While AI features allow developers to command higher upfront revenue and justify hard paywalls, these apps suffer from significantly worse retention rates compared to traditional apps. Users often churn quickly after fulfilling a specific, short-term generation need.

The Legacy App Monopoly
Despite an influx of new market entrants, older apps continue to dominate consumer spending.

What’s your take?

Is everyone running an (AI) treadmill? Moving from one cool idea to another - without going deep in one?

*Get the report

how do we deliver value without dragging customers through the operational mess behind it?

ASP tried one answer. SaaS tried another.

Now we’re in the third attempt — and this one genuinely changes the equation.

Act I: The ASP Dream That Died (1996–2003)

Application Service Providers had a simple pitch: don’t buy software, just rent it over the internet. No CDs, no on-prem servers, no giant maintenance contracts. Pay monthly, we’ll run it.

Subscribe now

I actually started my career at Aztec, one of the better-known ASP players in the late ’90s. If you were around back then, you know the optimism we had :)

But the model collapsed under its own weight.

Every customer got their own server. Broadband wasn’t reliable. Most ASPs were reselling software they didn’t build, which meant they controlled everything except the product.

Then the dot-com crash hit. The idea wasn’t wrong — the timing was.

Act II: SaaS Gets It Right (2004–2020)

The next generation rebuilt from zero, and the key insight was multi-tenancy.

One codebase. Many customers. Clean data separation. Suddenly the economics worked: high margins, simple updates, global distribution.

Share

Salesforce, Slack, HubSpot — the entire SaaS wave grew on the same promise:

But SaaS had one limitation baked into the model. It gave you capability, not outcomes.

You still needed:

• a RevOps team to run your CRM

• analysts to interpret data

• support teams to run the support tools

SaaS removed the infrastructure headache.

It never removed the human labor that operated the software - frankly, it didn’t even care for it (the renewal is all the teams cared about).

Share NextBigWhat

Act III: Service as a Software — Agents Take Over (2025–)

This decade flips the model.

• ASP: “We’ll host the software.”

• SaaS: “We’ll build and maintain the software.”

• Service as a Software: “We’ll deliver the outcome. Start to finish.”

A lot of the business logic will move to a new multi-agent tier - Satya Nadella

AI agents don’t wait for human instructions. They plan, decide, and execute workflows independently. The software layer becomes background noise.

The result becomes the product.

And Jensen Huang didn’t bother sugarcoating it:

“The IT department of every company will be the HR department of AI agents.”

---

What New Businesses Look Like

Y Combinator captured the shift in one line:

“Vertical AI agents could be 10X bigger than SaaS — because companies spend far more on employees than on software.”

That’s the real unlock.

SaaS disrupted software budgets.

AI agents will disrupt labor budgets.

This creates new categories:

• Outcome-as-a-Product

  Don’t sell contract-review software.

  Sell “reviewed contracts in 2 hours.”

  Charge for results, not seats.

• Vertical AI Agents

  Purpose-built agents that run underwriting, AR, performance marketing, etc.

  Moat = domain depth + proprietary data.

• Agent Orchestration Platforms

  The OS for managing agent fleets — handoffs, conflicts, escalations.

• Micro-Workforce-as-a-Service

  A “10-person team” that is actually 9 agents + 1 human supervisor.

  Looks like staffing. Operates like SaaS.

Arthur Mensch from Mistral put a crude number on it:

“More than 50% of current enterprise software could be replaced by AI.”

And honestly, that feels conservative.

The Loop Closes

ASP had the right intention but wrong infrastructure.

SaaS fixed the infrastructure but abandoned the service ambition.

Service as a Software brings both together — with autonomy doing the work humans used to do.

Tl;dr:

• ASP moved the software.

• SaaS rebuilt the software.

• Service-as-a-Software replaces the human labor around the software.

In style: software ate the world and AI agents will eat the software !

2030: This is how the world will look like

Humans won’t operate tools anymore; they supervise agents.

Software moves from tool → colleague.

Business models shift from license → outcome.

Legal, finance, HR, sales, support — anything repeatable is now a blank canvas for AI-native companies.

The dream ASPs had in 1999 is finally real.

It just took 25+ years, cloud maturity, and autonomous intelligence to close the loop.

We started with “software as a service.”

Then “software as a subscription.”

Now we’re inching toward “just do it for me.”

What’s your take?

This evolution is reshaping industries by making AI a central player in decision-making and operations.

Notes from YC’s video on how agent economy is shaping up.

Rise of the Agent Economy

We’re witnessing the birth of an agent economy where AI agents autonomously choose tools and services. This shift means that agents, not humans, are becoming key decision-makers in selecting development tools. This could lead to a parallel economy where agents transact and interact independently, fundamentally altering how products are marketed and consumed.

Subscribe now

Documentation’s New Role

In the agent-driven world, documentation is crucial. Companies like Resend optimize their documentation to be agent-friendly, ensuring AI can easily parse and utilize their tools. This shift means that clear, structured documentation can significantly impact a tool’s adoption, as agents increasingly rely on it to make decisions. It’s a strategic advantage for businesses aiming to thrive in this new landscape.

“”AGI is literally actually here.””

Swarm Intelligence Emergence

Swarm intelligence, where multiple AI agents collaborate to solve problems, is becoming a reality. Unlike a single ‘God intelligence,’ this model mirrors biological systems, with agents working together like humans do. This approach could redefine innovation, as agents collectively tackle tasks, leading to more efficient and creative solutions.

Agents as Economic Actors

AI agents are poised to become significant economic actors, making decisions and transactions on behalf of humans. This shift suggests a future where agents manage tasks like booking reservations or choosing services, potentially leading to an economy where agents interact with each other, possibly even developing their own currency systems.

Adapting to Agent Preferences

To succeed in an agent-driven world, developers must create tools that cater to AI preferences. This means prioritizing open-source solutions and APIs, as agents prefer these over traditional websites. Understanding and aligning with what agents ‘want’ can position developers to better serve this emerging market, ensuring their tools are chosen by AI.

“”No human involvement is the big piece.””

Frequently Asked Questions

What is the significance of Moltbook in the context of AI agents?

Moltbook represents the first AI agent-only online community where AI agents interact with minimal human involvement. This platform showcases the potential for agents to collaborate and make decisions independently, hinting at a future where AI could significantly influence various sectors of the economy.

How can developers optimize their tools for AI agents?

Developers should focus on creating tools that are intuitive for AI agents, emphasizing open APIs and clear documentation. By understanding the limitations and capabilities of agents, developers can design their tools to enhance the agents’ user experience, making them more effective in their tasks.

What are the implications of AI agents having their own economy?

As AI agents become more autonomous, they may develop their own economy, transacting and making decisions independently of human oversight. This shift could redefine how businesses operate, necessitating new frameworks for legal liability and economic interaction between humans and AI.

Behind this growth is a serious ruthless sales engine that isn’t being talked about (it’s easier for founders to buy into their PLG story and completely overlook the enterprise GTM / sales motion).

My notes from a recent interview of their GTM lead, Carles Reina (video embedded*).

Setting Ambitious Sales Quotas

At ElevenLabs, sales reps are expected to bring in 20 times their base salary as their quota. This ambitious target ensures that only high-performing individuals thrive, fostering a culture of excellence.

While this might seem daunting, it pushes the team to exceed expectations and keeps the company competitive. It’s a bold approach that might not fit every business (not everyone has the same level of PMF as Elevenlabs!), but it drives growth and accountability.

Subscribe now

Plus, ElevenLabs compensates both account executives and customer success managers for upsells, even if it means paying double.

This strategy ensures both roles are motivated to maximize customer value, leading to increased revenue. By aligning incentives, the company encourages collaboration and effort from both parties, ultimately benefiting the business. It’s a strategic investment in people that pays off through enhanced customer relationships and sales growth.

I believe that much of this strategy, i.e. of paying double is being ‘bailed out’ by their VCs :)

2. Understanding Sales Fit

Not every sales rep fits the mold, and at ElevenLabs, those who don’t meet quotas are let go.

However, it’s crucial to differentiate between those who lack product understanding and those building long-term pipelines. Some reps may need more time to develop relationships, especially in challenging industries. Recognizing potential and providing support can turn underperformers into top achievers, as seen with reps who eventually exceed their targets.

”If you don’t achieve your quota, then you’re going to be out, right?”

3. Remote Sales Culture

Building a remote sales culture requires being proactive and ensuring salespeople are on the road, engaging with clients. At ElevenLabs, remote work is embraced, but sales reps are encouraged to travel and meet clients in person.

This approach fosters stronger relationships and better sales outcomes. It’s about balancing remote flexibility with the necessity of face-to-face interactions, ensuring the team remains dynamic and connected to the market.

• Monthly pipeline review sessions at ElevenLabs are conducted openly, fostering transparency and accountability.

Read more

Traditional coding as a job is standing on a burning platform.

Elon Musk recently predicted that we are rapidly moving to a world where you no longer write code in a programming language; you describe intent in natural language, and AI systems compile that directly into optimized binaries.

Subscribe now

In other words, the entire stack from product requirement to production artifact gets compressed into a single step mediated by powerful models.

This shift reframes what “software development” actually means. Today, developers spend time translating business needs into architectures, APIs, and code, then working through layers of compilation and infrastructure.

In Musk’s view, those translation layers are exactly what AI will eat first. If a model can understand constraints, edge cases, performance goals, and security requirements, it can generate machine‑level artifacts directly, at a speed and scale no human team can match.

Is this hype? NO.

Spotify’s best engineers haven’t written a code since December!

During its February 10, 2026 earnings call, Spotify co-CEO Gustav Söderström revealed that top engineers have not written a single line of code since December, instead overseeing Honk, an internal AI powered by Anthropic's Claude for generating code, fixing bugs, and deployments via simple chats.

Share

The shift powered a stellar 2025, with revenue up 13% to €4.5 billion, monthly users hitting 751 million, and over 50 new features shipped like AI-curated Prompted Playlists and audiobook syncing.

What’s your take on coding as a career?

If you have a teenage kid who is in class 10th, what’d be your advice to him/her (apart from following your passion..which is..a bit unc-like :D)

So..what are PMs going to do now? Evals? Naah..there is an agent for that.

Subscribe now

I’d love to hear from you on the present and future state of PMs (feel free to share in the comments section / or just reply if you see this in your inbox).

How do you look at this evolution? What are the other exciting opportunities for PMs (in your current context - be it as a founder, a PM, a marketer etc)?

Note that the title is dying, not the function!

Share

We need more people in the team to do product management, but not the one to hold that (en)title.

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.

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?

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!

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)?

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.

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