Show IH: I thought building Mindfold was the hard part. Getting the first users is way harder.

The reason I don't finish saved learning material almost never comes down to time or forgetting, it's that I saved it before I had a real problem it solved. I have a backlog of tutorials on prompting and agent design I bookmarked because they looked useful in the abstract. The ones I actually finished were the three or four I opened while stuck on something specific building Alisio with Claude, mid-task, needing an answer right then. Progress tracking and quizzes assume the blocker is retention or structure. For me it was relevance timing: content saved for later competes with nothing urgent, so it loses by default. I'd test whether Mindfold can resurface a saved item at the moment someone's actually blocked on the related problem, not on a spaced-repetition schedule. That's a harder trigger to build than a quiz, but it's the difference between content you engage with and content you archive.

It sounds like you've put in a tremendous amount of effort into Mindfold, and now you're facing the challenges that come with scaling and getting users engaged. I can relate—when I was launching my own product in the content creation space, the technical build was just one part of the equation.

Once we had a solid MVP, the next phase was all about understanding user needs and refining the experience. I recommend conducting focused user interviews or surveys to gather feedback on specific pain points. This helped us validate our assumptions and pivot our features based on what real users wanted. For instance, we initially thought automation would be a top priority, but many users wanted more customization options instead.

Also, consider establishing a community around Mindfold. We found that fostering an engaged user base not only offered valuable feedback but also led to word-of-mouth growth. Implementing a referral system or offering early adopters exclusive access to new features can incentivize sharing.

You might also want to explore partnerships with educational institutions or corporate training programs. We did outreach in that area, and it opened doors for pilot programs, which provided not just revenue but also invaluable case studies and testimonials.

Lastly, if you haven’t already looked into content marketing for your SEO, it can amplify your reach dramatically. Sharing insights on how users can best leverage Mindfold for their learning goals could drive organic traffic. Just remember that your content should reflect real use cases and successfully showcase what sets Mindfold apart from other learning tools.

Keep pushing it’s a tough journey, but getting through these stages can lead to a product that truly resonates with its audience.

To answer your question honestly: for me it's almost never time, it's that saving something is a way of ending the thought instead of starting it. The save button feels like progress, so the itch goes away and I never reopen it. If I were you I'd test whether Mindfold can push the first tiny step at save time — one question or a 2-minute chunk generated right then, before the motivation dies — rather than waiting for someone to come back to a library. Watching those 10 people, I'd track one number above all: how many items get a second session.

The part about watching 10 different users apply the product is the most valuable step.

You could hypothesize about their needs until you are blue in the face but it will not be as valuable as observing people and what they do not know.

I also liked the question about building something people do not need versus something you think they might need.

It is a good question, but a hard one to answer.

Sometimes you need to see how an individual uses a product to see what needs to be improved upon.

I agree that building products based on assumptions can lead to failure, but I also know that not building products based on assumptions leads to stagnation.

Ten is enough if you choose people who already have a live problem, not prospective users. Ask them to show you the last thing they saved but did not return to, then make them try the smallest path to getting value. The gap between what they say they want and what they actually reopen will probably tell you more than the feature list.

The "save it and never come back" problem is real, and the useful thing about your ten-people plan is that you can watch whether they return without a reminder. If they only re-open it when nudged, the pile moved rather than got smaller. That is a harsh test but it is the one that tells you if the journey framing works.

Feeling this in real time. Building Sabriel AI took a year, and I assumed the hard part was behind me. Turns out getting anyone outside my own client base to even try it is a completely different skill, one I have zero experience in compared to the technical build. What's worked even a little for you so far, or is it still all trial and error at this stage?

Building on the point above that recruiting the 10 is the distribution work: the wrinkle nobody mentions is that most of the places those 10 hang out will not let you mention the product at all. We run a remote jobs index and learned that expensively, two posts removed before we had properly read the sidebars.

What worked instead was answering questions in those same communities with no link and no product mention. People click through from the profile. Slower, but it puts you in front of exactly the ten you want, and they turn up already trusting you a bit.

So read the rules of every community on your list before you write anything. Boring hour, saves you a removal or a flagged account, and the second one is much more expensive than it sounds.

The 10-user approach makes a lot of sense. Real usage will reveal much more than chasing traffic—especially where people get stuck and what actually brings them back. The “save it for later” problem is very relatable, so I’m curious to see what you discover.

Answering your actual question: I don't finish saved things because nothing ever asks me to. There's no deadline, no other person, no cost to skipping it - so it loses to whatever's on fire that day. That's the risk with "turn your pile into structured learning material": generating the course is the easy half, and the reason the pile exists isn't that it was unstructured, it's that it was optional. Watch for that in your 10 - if someone gladly imports 15 links and then never opens day 2, more AI structure won't fix it, a reason to show up will. One suggestion for the interviews: ask each person to open their actual saved-links graveyard on screen and talk you through the last five. What they say about why they saved it, and their face when they see it again, is worth more than any answer to "would you use this".

Honestly answering your question first: for me the pile never gets finished because saving something is the moment it feels handled. The bookmark is a little hit of "I've dealt with you." Opening it later is homework with no deadline, so it loses to anything urgent. The learning tools that stuck for me shrank the unit of progress to something absurdly small — one card, one question — so starting was cheaper than avoiding.

On the 10-learners plan: I'm two weeks into the same experiment (10 pilot families for a pet-care tool instead of learners) and the part nobody warned me about is that recruiting the 10 IS the distribution work. The channels where they hang out, the pitch, the follow-up — that's not preparation for the test, that is the test. If you can't find 10 people to watch use it, you have your answer before anyone touches the product.

One thing that's been working: don't ask "would you use this" — ask them to show you the last time they actually had the problem. People's screenshots and workarounds tell you more than any survey answer ever will.

The strongest signal is the willingness to stop building and watch real users instead. The founder is explicitly questioning whether the product solves a problem beyond their own experience.