Both big AI labs quietly switched on watermarking this month Anthropic marks every text Claude touches, OpenAI marks GPT-Live voice audio (EU AI Act, Article 50).

The catch: a detected mark means the model processed the text, not that it wrote it. Fixing grammar leaves the same mark as full generation. And no mark proves nothing either edits, translation or transcription wash it out.

Last week I changed my free tier. Free exports used to be watermarked. Now they are clean, and the watermark only appears for visitors who have not signed up at all.

My reasoning: the watermark was supposed to push people to upgrade, but it was mostly pushing them to leave. A clean free export gets used, shared and seen. The signup is the new gate, and the paid plans sell on product limits and extra formats instead.

Let s be honest most AI video tools are still too complicated or just don t live up to the hype. Why is it still so tough for creators and marketers to make quick, simple edits? Most tools are built for pros, and AI videos often feel a bit off. We kept hearing: Why can t I just make a fast explainer or a talking avatar without a studio?

That s what we want to change:

• Instantly create videos from text prompts

• Bring blogs, courses, and social posts to life with talking avatars

• Experiment and storyboard easily not just edit

Posted daily progress updates, screenshots, behind-the-scenes stuff. Got likes, some nice comments, almost zero actual signups. Made me wonder if "build in public" is overrated for early traction, or if I was just doing it wrong.

For those who got real early users this way, what made the difference? Was it a specific platform, a different kind of content, or just consistency over a longer time than 3 weeks?

I recently caught myself explaining a product flow to someone before they had even used it.

That made me realize something uncomfortable:

I got burned by this. My site says things like "152 label formats" and "114 countries" on hundreds of pages. When the real count changed, prose across dozens of pages kept the old number for weeks.

What I do now:

AI has made it easier than ever to build and ship with a small team and no funding- but is that a better playbook than raising VC?

I see it from both sides:

• Getting to production, ramen profitability, etc. seems easier than ever. Yay bootstrapping.

• Because it's so easy, getting blown out of the water by a VC-funded competitor also seems easier than ever.

I have been commenting on launches here most days since the end of July. One argument has come up in about a dozen of them without me planning it, and twice the maker went and found the same defect in their own code afterwards. That seems worth writing down properly rather than repeating in other people's threads.

The shape is always the same. A system has a value that means healthy. It usually does not have a separate value for "I could not tell". So the second collapses into the first, and it collapses in the direction that looks fine.

Something we keep noticing: even when speaking a prompt would be faster, we still type it.

Voice feels great for quick asks like "set a timer" or "what's the weather" and awkward for anything that needs revising mid-sentence, or reading back a complex answer.