Today was one of those genuinely fun build days.

I’ve been curious about video for a long time, but also a bit hesitant. Video feels expensive in the AI world, both in terms of tools and expectations. There’s also an aesthetic bar that’s hard to clear if you want to produce truly world-class AI video content and turn it into a sustainable business. At least that’s my current assumption — and to be fair, my knowledge of AI video is still very surface-level.

That’s exactly why today felt like a win.

Instead of going all in on video production, I set up my first simple, practical flow that turns an existing blog post into a YouTube video. No overthinking. No perfection. Just shipping something real.

🎥 Prefer video? Watch the narrated Build in Public log on YouTube.

The setup

To make this work, I needed a YouTube account for the project. I initially assumed that meant yet another Gmail account… but Google stopped me.

After some back-and-forth (and help from ChatGPT), I realized I could create Brand Accounts inside YouTube, all under a single Google account. That small discovery made me question whether creating a separate Gmail for every project in the past was actually the right move.

With the channel in place, the flow came together:

• ChatGPT generates a video script from a suitable blog post (we started with post #3).
• ElevenLabs handles text-to-speech. I picked a voice, and ChatGPT helped fine-tune the settings for a calm, neutral narration.
• The audio goes straight to YouTube as a long-form video.
• Finally, I asked Claude to add proper YouTube support back into the blog, so the content is clearly connected across platforms.

The result isn’t flashy — and that’s the point.

It’s a lightweight, repeatable system that lets one piece of content live in more than one place, without dramatically increasing cost or complexity. For now, that feels like the right balance.

And yes…
nano banano — even the infographic generation worked.