Look, we’ve all been there. For years, coders like us dreamed big: Xcode on the iPad. Coding on the go, tapping away on that gorgeous glass screen, free as a bird. It was a killer vision. But in 2025? That dream’s deader than disco. And you know what? We don’t need it. We’ve built something way better ourselves.
The game didn’t just change – it got blown up. While we were all staring at our iPads, begging Apple to let us code, the real action happened with Claude Code. You don’t need a monster machine in your bag anymore. All you need is a terminal. Boom. Done.
Claude Code is the real deal now. It’s like having your entire dev environment in your terminal, no matter what device you’re on. Your iPad, phone, or laptop – it’s all just a portal to your codebase. Claude Code handles the heavy lifting, understands your code, fixes bugs, manages git, all with natural language commands. It’s coding supercharged. Tools like Tailscale or Cloudflare make it stupid easy to connect securely to your projects. Set it up once, and you’re golden. No more dependency nightmares or local setup headaches. Just pure, fast coding.
And here’s the kicker: we’re in the age of Vibe Coding. AI genius Andrej Karpathy nailed it. It’s not about sweating the syntax. It’s about the vibe, the intent. You tell Claude Code what you want, tweak it, iterate, and boom – it’s alive. All from a terminal, anywhere. Your iPad? Unless you’re sketching in Procreate, it’s just a fancy paperweight.
Apple? They’re stuck in the past. WWDC25 was a snooze – shiny “Liquid Glass” menus for iOS 26 and macOS 26. Looks nice, sure, but it’s lipstick on a pig. They’re obsessing over pretty interfaces while the world’s moving to AI-driven coding. Sam Altman’s calling it the third HCI revolution: first keyboards, then touch, now AI that just gets you. Apple’s still fiddling with Swift, a language that’s elegant until you hit its buggy frameworks. Good luck building anything complex without losing your mind.
The Xcode-on-iPad dream was old-school thinking – shoehorning yesterday’s tools onto a new toy. Forget that. The future’s about working smarter, not harder. A terminal, a secure connection, and Claude Code’s AI brain – that’s the winning combo. Everything else? Just noise.
I’ve been developing for Apple platforms since 2011. I lived through the last major UI refresh—flat design in iOS 7. People complained then too.
But this time feels different. Flat design actually simplified things. It stripped away the cruft and let users focus on what mattered. Liquid glass UI does the opposite. It buries utility behind flashy effects and makes users work harder, not easier. That’s not progress, that’s regression disguised as innovation.
I was shocked to see this on my iPad.
The Control Center is basically unusable now. I worked on Outlook Mobile, where we obsessed over contrast ratios daily. This iOS 26 beta wouldn’t pass our internal review. It wouldn’t even get close.
This isn’t just bad design. It’s breaking accessibility for people who actually need to read their screens. We’d never ship something this low-contrast because it would violate our customer commitments.
Apple chose pretty over functional. That’s not progress.
AI Era: UI is Legacy?
The way humans interact with computers is changing. UI matters less when everything has chat or voice interfaces. You don’t need to tap buttons when you can just talk.
In 2018, Chinese startup Smartisan built the TNT Station—a workstation controlled entirely by voice.
However, Smartisan was too early. The company died, even after ByteDance acquired them.
Now it’s 2025, and I’m impatient with traditional interfaces. Keyboards and mice feel slow. My programming workflow is basically prompt engineering. I use voice input with GPT transcription instead of typing—it’s more accurate than Apple’s dictation, especially for me who’s not a native English speaker. Even this blog post was voice-written.
UI matters less than it did in 2013 when iOS 7 launched. The mobile market isn’t exploding anymore.
With slower growth, companies can’t afford teams of designers rebuilding their systems for liquid glass effects. Even as a designer, I think refreshing UI is the wrong priority.
We should build AI-native interfaces, not fancy UI with worse accessibility.
Let’s talk about Sahil Lavingia, the author of *The Minimalist Entrepreneur*.
About three to five years ago, I finished reading this book. To be honest, I didn’t consciously apply its methods when tackling tasks afterward. However, Ray frequently reminded me: when it comes to purchasing tools and resources, you need to be willing to spend money—on things like useful software or outsourcing accounting services. That’s how you free up your energy to focus on what you’re best at, your core strengths, and create maximum value.
This principle sounds straightforward when you hear it, but you only truly come to appreciate it after experiencing the pitfalls yourself.
My “Minimalist” Detour
Over the past six months, I left Microsoft to work on my own product. To put it simply, it was pretty much a failure. My initial approach was to stubbornly stick to one direction and adhere to “extreme frugality,” unwilling to spend money on anything.
Looking back now, a few mistakes were made:
Putting all your eggs in one basket: It’s just like investing—going all-in on a single direction, like putting everything into one stock, maximizes your risk. But if you’re buying an index fund that covers hundreds of companies, the risk is much lower. When building products, especially in the early stages, you also need to try different approaches and iterate quickly.
Time > Small Money: This means not understanding the true value of time. Often, spending a bit of money on a service or a tool can save you time—and that saved time is worth far more than the money spent, especially when compared to your previous hourly rate (like when you worked at Microsoft, for example). Many people, including my past self, have a hard time making this mental shift.
Of course, the inability to pivot might not just be a cognitive issue; it could also be due to practical constraints—like being in debt, having high living expenses, or simply not having the spare cash, resources, or even time to make these “optimizations.” This is one of the reasons why many people struggle to succeed.
Visit Sahil in Person
Last December, I made a trip to Brooklyn, New York, and while I was there, I paid a visit to Sahil Lavingia himself at his company, Gumroad.
Gumroad, the company that serves as a platform for trading virtual goods such as courses and other digital resources, has quite impressive revenue streams (it’s said their total revenue this year is between $100 million and $1 billion? The exact figure is uncertain, but the transaction volume is undoubtedly significant). However, contrary to expectations, their office is surprisingly “minimalist” — just a small room with everyone squeezed together. When I visited, they were preparing for a Christmas party, and Sahil himself was there frantically unboxing packages, with absolutely no airs of being the boss.
Quite interesting. We sat down and talked for a bit, and he gave me a piece of advice that left a deep impression: try using Devin AI, a fully automated programming tool that costs $500 a month. “Definitely worth a try,” he said. “Compared to hiring an engineer in New York with an annual salary of two to three hundred thousand dollars, it’s incredibly cheap.”
New York Insights & Core Observations
We talked about a few other things as well. For instance, the renowned open-source company in the AI community, Hugging Face, is located right next door to him. I even took the time to go over and snap a photo; their office is also quite small.
This trip to New York, combined with my own experiences of failure, made me realize something even more deeply: whether it’s a small entrepreneurial effort or a larger startup, the most critical thing is always finding Product-Market Fit (PMF); that is, your product must have a market and customers willing to pay for it. Take our community platform as an example—its technology is actually quite poor (typing out so many words right now is really unpleasant, the experience is subpar), but it tackles many problems like live streaming, community engagement, etc. This is PMF.
If you don’t know how to find PMF, there’s no shortcut, the only way is to “do more”:
Build more products / try more things
Meet more people, especially potential customers
Understand their biggest pain points
Find ways to help them—either by saving them time or helping them make more money
If you can solve these problems for your customers, you can naturally get a piece of the pie. Sahil’s most profitable product, Gumroad, is exactly doing this.
The Changes and Constants in Sahil
Sahil is quite an interesting person. I asked him, “We all think your book, *The Minimalist Entrepreneur*, is excellent. Do you have plans for a second edition?”
His response was very Sahil: “By the time I leisurely finish writing it and go through the publishing process, the ideas in the book would likely be outdated. The times change too fast.” He believes that instead of writing a book, it’s better to share instant thoughts directly on Twitter (X) or do some live streams so people can follow him directly. It’s more efficient, and the information is fresher.
He is also quite interesting. Although he is an entrepreneur, he is politically a typical liberal of the American Democratic Party. I talked to him a lot and watched many of his technical shares, and he indeed has substantial insights. Last summer, he pointed out early on that the tool Cursor could perform a large amount of AI semi-automated programming and held four consecutive live streams about it, all of which I participated in. On Google Meet, we could interact freely, making the discussions very enjoyable and profoundly enlightening for me.
Founder Mode, God Mode & Open Source
Additionally, about six months ago, it became quite popular in the community to discuss Founder Mode vs Manager Mode. Sahil has an even more extreme take on this, mentioning on Twitter the concept of “God Mode.”
What does it mean? It means that a founder must have the courage to go beyond the conventional and even oppose internal consensus to make decisions based on deep insights and convictions.
He cited an example from his own company: Gumroad once adjusted its fee model, seemingly moving from a tiered system to a flat rate, which was considerably higher (I can’t recall if it was from 5% to 10% or 10% to 30%). At that time, almost everyone within the company was against it, fearing that customers would leave. However, Sahil stood firm in his belief that as long as Gumroad provided enough value, customers would not be lost.
So what happened? About a year later, it turned out he was right. At that time, quite a few people on Twitter (including some well-known figures in the Chinese community, like Ding Yi) were saying that Gumroad was just trying to “cash in,” and that users would definitely flee. But in reality, they didn’t.
Certainly! Here’s the translated section:
What’s even more intriguing is that Sahil believes Gumroad’s core competitive advantage doesn’t lie in its fees, or even in the code itself. Therefore, he even plans to open-source Gumroad’s code (on April 4, 2025)! Isn’t that quite surprising?
Flexible ‘Growth Mindset’
Sahil is incredibly adaptable in his approach, embodying the classic “Growth Mindset,” which essentially means being flexible.
The best example is his perspective on remote work:
In 2023, he was still “calling out” to YC founder Paul Graham about remote work, believing that if remote employees weren’t managed well, it was a manager issue, not a problem with the remote work model itself. (link here)
A year later (2024), he openly admitted that, after practical experience, he found that working together in person was actually more effective, even if it wasn’t every day in the office. (link here)
This attitude of being willing to revise one’s views based on practical experiences might also be key to his continued success.
In summary, Sahil Lavingia and his “minimalist entrepreneurship” go far beyond just saving money. It’s more about focus, leverage, rapid iteration, and value-based decision-making. This experience has given me a deeper understanding of these concepts.
