The Frictionless Effect
Why Brave, Cursor, and other “Baked-In” tools keep winning
There’s a quiet pattern I’ve been noticing across the tools I actually like using, not just tolerate. And it feels almost obvious in hindsight.
Take web browsers. I used Google Chrome for years. Installed uBlock Origin, added Privacy Badger, tweaked things. I had an ad-free experience. It worked. I could actually read a newspaper website!
But the moment I tried Brave, I never looked back. Why?
Because the value was already baked in. No extensions. No configuration. Just open it and go. Ads blocked. Updates don’t break things. No fiddling with extension settings. That small thing? It’s actually huge. Brave didn’t offer me features. It offered me a done deal.
The experience was frictionless.
And it turns out this pattern shows up everywhere. The tools that win aren’t just powerful, they’re pre-decided in their value. They don’t ask me to “set them up.” They just are.
Cursor vs Copilot: The same code, but baked-In
I remember when GitHub Copilot dropped. I jumped in fast and got on the annual plan, even. The autocomplete suggestions were magical. The chat was nice, once it came. It was a game-changer fasho.
But still, it was just an extension inside VS Code, sitting alongside everything else. Useful, yes. But it still felt optional. You had to summon it.
Then along came Cursor, a fork of VS Code where AI wasn’t just added, it was hard-coded into the experience. Cursor wasn’t asking “do you want to use AI?” It just assumed.
It didn’t just embed Copilot’s features, it reimagined what a code editor is when AI is the default. Cursor had the same raw capability, but a wildly different feel. You didn’t need to justify its value. You just felt it.
Users don’t want features. They want outcomes.
This is the core idea.
Tools often brag about their features, extensions, customizations, power-user toggles. But what most users really want is simple:
Block the ads.
Write better code.
Don’t make me do extra steps.
And most importantly:
Don’t make me think about how to get the value.
The Value Activation framework
Here’s a mental checklist I’ve been using to make sense of this:
Principle: Zero Config
Question: Can I use the value immediately?
Winner: Brave
Principle: Integrated UX
Question: Does the feature feel native, not bolted on?
Winner: Cursor
Principle: Always On
Question: Is it persistent across updates?
Winner: Brave
Principle: Default On
Question: Do I need to opt out instead of in?
Winner: Cursor
Principle: No Maintenance
Question: Does it “just work” without upkeep?
Winner: Brave
If your product fails 3 or more of these, it’s probably losing users who’d otherwise love it.
Forking as a Strategy (FaaS?)
Cursor didn’t just build an extension. They forked the entire editor. They said:
“Let’s take a familiar tool, remove the distractions, and hardwire the magic.”
This is the same strategy used by:
Obsidian, with Markdown. It’s a clean-slate reimagining of what knowledge tools should feel like before the internet touches them.
Raycast, with Spotlight. Raycast isn’t a skin on top of Spotlight. It’s what Spotlight would be if Apple rebuilt it for developers in 2025.
Linear, with issue tracking. Linear isn’t a tuned-up Jira. It followed the same design-first, rethink-from-zero philosophy
They just committed to opinionated defaults. They stopped asking for permission. They built the spoon with the ice cream (please don’t bring this to life).
The “Taste Test”
Channelling my inner Chapelle here, I am going to start with the punchline:
People justify value after they taste it.
If your tool needs a landing page, a product tour, and three docs before I feel the magic you’ve lost me.
But if you let me taste the outcome immediately? If the thing just works?
I’ll pay. I’ll tell friends. I’ll write posts like this.
For the gearheads, or anyone who’s sat in a car that just felt better…
I certainly no car guy, but I have some interest in cars and so this is my “Explain this in Car Guy terms”.
This frictionless effect is exactly why a Mercedes AMG C63 will always feel better than a fully tuned C300 with the same horsepower and torque, upgraded interior and all. You can upgrade a C300. Flash the ECU. Add performance parts. Hit the same 0 to 60/100 numbers.
But it won’t feel like an AMG.
Why?
Because a tuned C300 is retrofitted. The C63 is engineered. You don’t need to explain it to anyone. They sit in it, they know.
Same thing in software:
Brave is the AMG.
Cursor is the AMG.
Your product? Might still be a C300 with after market exhausts.
And that’s fine, but only if you’re honest about the gap.
In car culture, the difference between tuning and engineering is respected and monetized.
Who else knows about Forking as a Strategy (FaaS)?
Kourosh Mansory. You can look him up if you aren’t familiar with his art.
Mansory doesn’t just tune a car. They rebuild identity. They don’t slap on a spoiler and slap on a badge like “Tuned by TJ”, they start from the AMG’s core and go further.
You see it, and you already know this didn’t come from a garage. This came from a different philosophy.
Even website copy is subject to this Law
The frictionless principle isn’t just for product architecture. It shows up in UX writing and interface design, especially when it comes to calls to action (CTAs).
Bad button copy hides the value:
[Learn More]
[Click Here]
[Submit]
These phrases ask the user to guess. To click based on curiosity instead of clarity. It’s like putting the good stuff behind a fogged-up window.
Better? Make the outcome obvious:
[See Why Developers Love It]
[Block Ads Instantly]
[Compare to Copilot]
Don’t make them search for value. Let them feel it, before they commit.
If your button says Submit, you’re asking the user to do work. If it says Get the Report, you’re delivering value.
Real-World case: How Pieter Levels weaponized Frictionless value
You may know Pieter Levels (@levelsio on Twitter/X). He is a machine when it comes to app development. But the interesting bit is that he builds with minimal tech stacks. No teams. No dashboards of dashboards. Just pure ship-it-now energy.
And somehow, his apps like:
…get traffic, get users, and print money.
Meanwhile, nerds on Twitter are in his mentions every week like:
“This would be so much better with NextJS or Astro or TypeScript strict mode or shadcn/ui (shadcn/ui enjoyer btw) or or or …”
But they never stop to ask:
Why does it work so well without all of that?
Because Pieter bakes in value.
You land on the site and it does the thing.
It looks good enough.
It works immediately.
No empty state. No onboarding friction.
And that’s what most users want. Not a 99 Lighthouse score. Not bleeding-edge frameworks. Just:
“Can I do the thing I came here to do, right now?”
Sidebar: Why Microsoft Might Be the Only Real Winner in the AI Pivot
Everyone is scrambling for an AI strategy. Startups are shipping wrappers. Big players are announcing labs, Airtable has an AI app builder now, or as they call themselves, “AI App Building for Enterprise”. But the ones who’ll see real growth aren’t the ones with the most models but the ones who bake AI directly into the workflows people already trust.
Microsoft is uniquely positioned for this.
They own the IDE (VS Code). They own the LLM (OpenAI). They own the cloud (Azure). And they’re already deploying Copilot across Office, GitHub, Teams, and more.
But there’s a deeper problem here:
VS Code wasn’t designed for AI. Cursor proves that.
Microsoft can’t just keep duct-taping Copilot into the sidebar.
If they want to truly compete in a Cursor world, they’ll need to do what Cursor did, but bigger:
Rebuild VS Code. From scratch. For AI-native workflows.
No toggles. No extensions. No opt-ins.
A completely new editor with AI as the foundation, not the add-on.
And if they don’t?
Cursor and its clones will keep eating their lunch, one seamless and frictionless experience at a time.
Closing thought
What we often call “product-market fit” might just be frictionless value delivery. Not faster. Not cheaper. Just less thinking required to get the good stuff.
So next time you’re shipping a tool or choosing one, ask yourself:
“Does this feel like an AMG…
…or like something tuned to sound like one?”
And while you’re at it,
“Does that button say what the user is actually getting?”
And maybe also,
“Would this make Pieter Levels roll his eyes… or hit deploy?”
If your answers are honest, the product might just ship itself.
What about power users?
Yes, some users want knobs and toggles. But even those users want the default to be great. That’s the bar now.