Author here. Built this solo with Replit Agent + Perplexity over 15 days—echoing those recent HN threads on agents in production. Happy to dive into RLS bypasses, OAuth double-encryption, or background-job tenant leaks if folks are building similar stacks.

I do think the narrative is propagating a false sense of easy to achieve story, considering the founder wrote a lengthy note of self-reflection. Very sad to see this narrative this removes the faith in the founder's earlier reflection. Creating a successful business is never easy, you could get lucky, but you rarely get that lucky. This kind of article makes it seems all rosy.

Tried the chrome extension - doesn't seem to be working. Can you guys look into this?

Startup school's valuable first lesson in customer communication and PR is right here. Well, software screwed up the experience and crushed the expectations of so many, including me. Moving on :)...

You're right - it was a huge lesson for us that we need to have way better protections in place to prevent something like this from happening.

Acknowledge the effort you're taking to correct this snafu.

Yes the lessons learned here were translated. As of now the product hunt for marketing is a place for discovery and we realize that our new model will build a community and we have to build active monetization around value-based product around the same premise - yes product hunt for marketing won't make money by itself.

Well, the example is a very simplistic one for the purpose of this interview. The vision, if you could call it that, was to package other signals e.g. weather, geography, and demography into personalizing a message. Having come from a background of e-commerce I definitely know personalization has a positive impact on cart values, engagement, and conversions. Considering all of that our hope was that we would be that kind of a personalization software for mobile apps.

But how does that work in practice?

Does their system determine those signals or are they fed to it as input?

How can it possibly provide a generic solution to "personalizing a message" that works across all apps and is not just something trivial like the example in the interview?

On the web the kinds of "personalization" I can think of are public account profiles, targeted advertisements, recommendation engines ("you may also like...", "products like this...", etc.), algorithmic newsfeeds and history-sensitive search. If their software does any of those, maybe they should just say what it does.

You bring up a good question. The $ amount while painful was spent over getting some design work,domain name etc... The value we thought the product would add was to make apps more engaging. Personalization is a big deal in e-commerce world and I naturally translated that as a consequence to mobile app - I do realize that was a mistake. So yeah, we definitely expected to get paid for such a service.

> Personalization is a big deal in e-commerce

That's a claim that is not backed up by anything. Most personalized service one can get is in boutique non-ecommerce companies. Those companies are struggling.

That in turn means that personalization at this time is a gospel.

Yes, hard lessons and the funny thing is we were oblivious to that name and paid a lot to acquire the domain name.

Absolutely the blood thinner won and it's a hard lesson that we learned.

My point was not so much about losing a battle over name recognition, that's something that happens relatively all the time. I mean specifically losing it, even to the general public that isn't on blood thinners (like myself), to a blood thinner. In a lot of developed countries, where direct to consumer pharmaceutical advertising isn't legal this obviously would not have happened...