Can’t sleep late at night. Recently saw feedback from several potential clients and thought about some things I wanted to share casually. Of course, it’s still AI-related.
The Paradox of Cost Reduction and Efficiency Improvement
I’m increasingly feeling that from the perspective of human nature and overall development, emphasizing cost reduction and efficiency improvement or cost savings when making products and communicating with users might be the wrong approach.
Whose Costs Are Being Reduced, Whose Efficiency Is Being Improved?
When users in an enterprise start using your product, they will inevitably resist because it means their agency is being negated – their past work and skills might be replaced.
So I’m thinking, maybe that’s where the problem lies. We’ve been selling to enterprise decision-makers, telling them how much money they can save, how many fewer people they can hire.
But what about the people who actually have to use this thing?
Why would they cooperate with you in completing this process of “being optimized away”?
The Real Value: From Cannot Do to Can Do
The feedback from several clients these past few days made me realize that those who are truly using our product aren’t doing so because the company requires it, but because they’ve discovered something themselves – they can do things they couldn’t do before – the new narrative is the meta-narrative of David defeating Goliath.
A small team suddenly has the capabilities of a large team. One person can suddenly launch projects that previously required a bunch of resources.
I now believe that truly viable AI product narratives shouldn’t be about helping existing winners save costs, but about enabling people who didn’t have entry tickets to enter the game. It’s not about making the strong stronger, it’s about giving more people the qualification to participate in the game.
Insights from Independent Developers
An observable example now is independent development. The monthly recurring revenue of some developers’ applications I know of is already higher than some funded startups.
Why? Because AI writing code is a proven narrative right now.
But I think this is also why this is difficult to promote in other large companies – it doesn’t fit the meta-narrative.
Think about it: why would an engineer at a big tech company actively use AI to improve efficiency?
After improving efficiency, what then?
Complete KPIs faster? Then the company discovers one person can do the work of three?
What benefit does this narrative have for them?
But for independent developers, it’s completely different. AI is their co-founder, the only possibility for them to go from 0 to 1.
Supabase Is Also an Example
Tools like Supabase are tailor-made for independent developers and small SaaS businesses of this era. It’s not helping big companies “reduce costs and improve efficiency,” but rather making database architecture that previously required a complete backend team accessible to one person who can quickly get started.
Supabase’s value isn’t in helping big tech companies hire fewer database engineers, but in enabling a full-stack developer to build a backend system in a few hours that would have taken weeks or even months before. When AI helps you write frontend code and Supabase solves your backend data problems, an independent developer suddenly has the ability to compete with small teams.
This is the essential difference between empowerment tools and optimization tools – the former creates new possibilities, the latter compresses existing costs.
So the same technology, in different people’s hands, has completely opposite meanings:
One is optimization, the other is empowerment.
The users of the former are helping others save costs, while the users of the latter are creating possibilities for themselves.
More Possibilities
Then, I can now see other possibilities in things like Sora and Veo.
Not for designers to use, not for directors to use, but for those who have stories to tell but don’t have the budget or team to express them.
Letting a novelist create a trailer, letting a podcaster create visual content, letting a teacher create teaching videos that previously required an entire production team.
These people don’t have the fear of “being replaced” because they were never in that game to begin with.
They represent purely incremental demand.
David Defeating Goliath
So back to what I said at the beginning: David defeating Goliath. This isn’t just a marketing slogan, this is the only truly viable narrative structure.
Because in this narrative, the user is the beneficiary, and they have real motivation to learn, to tinker, to spread the word.
You don’t need to convince an independent developer “you should use AI” – they’ll stay up late researching on their own.
You don’t need to train a content creator who wants to create – they’ll figure out every feature themselves.
And those big companies? No matter how much you train, no matter how much you push adoption, employees know deep down this isn’t helping themselves.
So the real users of AI products might not be those already at the table from the start, but those who didn’t even have a chair to begin with.
Image caption – Tribute to Caravaggio’s new version of David defeating the giant
Model – OpenAI Image – 1
Prompt:
“oil painting in Caravaggio style, contemporary David wielding single beam of creative light standing over defeated corporate Goliath, dramatic tenebrism, intense contrast between light and shadow, oil painting texture with glazing technique, humility conquers pride, epic banner format –ar 21:9 –style raw”