Jan Builds Digital

I’ve found that supervising the AI and making sure we’re aligned on the task design leads to better results for more complex tasks.

The biggest challenge here is the lack of trust, which leads to micromanaging any complex tasks.

If “vibe coding” means coding without supervision, what’s the technical term for using AI coding tools in a coordinated, supervised, human-in-the-loop workflow? What would you call that? #buildinpublic

Yes in this case, higher productivity means skipping some unnecessary tasks

Some decline in coding skills is inevitable. People tend to get lazy and stop thinking as hard when there’s an easier way to do things.

Yes, AI tools sped up the coding and prompting. But they had no real understanding of what the right workflow should actually look like.

There seems to be a tipping point in project complexity where pure AI-driven design starts to fall apart and stops adding any real value. (5/5)

In reality, building it meant talking to experts, implementing their advice, reaching out to potential customers, testing with real documents, and constantly refining the code and prompts based on what we learned. (4/5)

We recently built a fairly complex data-extraction workflow, and it’s funny to think I could’ve just asked AI to build it. (3/5)

But I’ve yet to see a tool that can deliver a moderately complex, full-stack product on its own without a human who understands high-level architecture, runs sprints, tests outcomes, handles debugging, and coordinates with other people. (2/5)

Generative AI is pushing developer productivity to the next level and is also fueling this new narrative that beginners can build anything with AI tools without ever looking at or touching the code. (1/5)

#buildinpublic

Building self-confidence and self-esteem naturally fuels motivation in business.

A confident founder realizes: I’m not fully using my skills, experience, or expertise right now. There’s a new market to grow, and I can excel there.

That’s a powerful source of motivation 🌊

#buildinpublic

Haha this is so accurate!

I’m sure Stripe works well for a lot of people, but the reviews suggest there’s another side to it. From what I understand, with Stripe you’re responsible for handling compliance yourself, which means there are more ways they can shut you down compared to using a merchant of record.

How am I supposed to partner with a company that can’t even respond?

From a customer’s perspective, there’s a huge gap in the market when it comes to payment processing and merchant-of-record services.

Instead, I went with Paddle, which has a solid 4.1 rating. Plus, using a Merchant of Record was supposed to make the integration easier. So what happened?

I’ve been waiting two weeks to get verified by Paddle. And the best part? No replies to my emails, and their in-app chat doesn’t work 😄.

I’ve been integrating a payment provider into my new product.

Based on Trustpilot reviews, I decided to skip Stripe. An average rating of 1.8 stars (yikes) and tons of scary reviews about nonexistent customer support didn’t inspire confidence.

#buildinpublic

Lately I’ve been thinking about this:

Being a founder can feel hypocritical in at least two ways:

• You hire people, but you don’t want to work for someone else yourself.
• You run paid ads and other marketing campaigns, even though you sometimes get annoyed by them yourself.

#buildinpublic

The two things you know about Ford’s business strategy are both wrong:

𝟭. “𝗜𝗳 𝗜 𝗵𝗮𝗱 𝗮𝘀𝗸𝗲𝗱 𝘄𝗵𝗮𝘁 𝗰𝘂𝘀𝘁𝗼𝗺𝗲𝗿𝘀 𝘄𝗮𝗻𝘁𝗲𝗱, 𝘁𝗵𝗲𝘆 𝘄𝗼𝘂𝗹𝗱 𝗵𝗮𝘃𝗲 𝘀𝗮𝗶𝗱 𝗮 𝗳𝗮𝘀𝘁𝗲𝗿 𝗵𝗼𝗿𝘀𝗲.”

First of all, he never said that.

(1/14)

In the age of LLM-powered search engines, GitHub hosts not just code, but also documentation tailored specifically for language models.

It seems like yet another attempt to game the algorithm that we keep seeing in SEO for years. LLMs will probably give scraped content less weight.

It does, and it’s usually safer to enter a market where competitors already exist rather than one where they don’t. It’s also difficult to break into a market/niche that has little to no competition.

I don’t think so. If anything, competition usually means there’s a real market.

At the very least, the number of competitors alone isn’t enough to judge our chances. We need deeper market insight to make an educated guess.

Do you take competition into account when choosing a market or niche?

Launching a new product means getting familiar with the market and competitors. The more I learn, the more existing solutions I discover, and it can feel emotionally crushing.

But putting emotions aside, does competition reduce your chances of launching a successful product? #buildinpublic

The first cookie banner I’ve come across that’s actually funny. 😀👍

It’s still just as hard to make a startup, despite vibe-coding and AI coding generally.

AI changes 𝘩𝘰𝘸 you write code, but hasn't changed anything else; you still need (a) to find people (b) who actually want to pay for this.

If this is shocking news, YNGMI.

What options do they have? They could detect real people and ban any ads that feature third parties. Or, better yet, they could authenticate third parties through a manual process or via technology like NFTs. Could this be a viable use case for blockchain?

Using fake likenesses of popular figures and celebrities in visual ads has surged in 2025 as generative AI has improved. This is making online ads less trustworthy and ultimately damaging conversion rates. I imagine this is a real pain point for platforms right now.

Would you add any other reasons?

Flexibility — solopreneurs can move faster and be more flexible with things like product development, adopting new tech, and turning user feedback into new features. Of course, that’s not always true if the founder is busy or tied up with other projects.

Customer support - being supported directly by the founder is a huge plus for users, and that usually disappears as teams get bigger. Employees are rarely as motivated as the owner.