As a technical founder I empathize with the authors situation and approach, but honestly spending that much time and energy on a product before even getting one sign up is a known failure mode.
In his own words he all-ready got early feedback from his family.
"I tried the local Iranian market. I showed it to friends, family, and potential clients. Their response: "Nobody in Iran will pay $500/month for this. The Persian language quality isn't perfect. We'll use free ChatGPT instead.""
Which should of been free feed-back on the risk vs reward.
scan the QR code with your phone, and your phone becomes the microphone. The avatar responds on screen.
Known limitation: Running on my GTX 1060 in Iran with 4G internet, so international latency might be rough. The code itself handles 100-300ms on proper infrastructure.
I get the scam concern given the circumstances. Happy to do a live Zoom walkthrough if you're seriously interested in verifying this is real.
I don't think I need to tell hacker news this, but don't scan that qr code with phone.
The only way HN is going to do that is by opening that link with either a js disabled browser and then opening the qr with a sandboxed android emulator.
you can also open a URL with another tab
press F12 and open the STT page URL and you have both the avatar and the STT page (where you use it as microphone) on the same computer without the need of phone, this design was to eliminate the need of microphone and setup user base sessions and control it in kiosk like environments
I should have validated with 5-10 customers before building the full enterprise stack. The monitoring, analytics, multi-tenant architecture - all premature.
My reasoning (wrong in hindsight): I thought the barrier was "enterprise features" not "can anyone actually sell this." I'm a better builder than salesperson, and it shows.
If I could redo it: MVP in 3 months, get 10 paying customers, THEN build the infrastructure.
not to mention that dopamine is generally associated with anticipation and searching + reinforcing behaviours, whereas pleasure and satisfaction is associated with the opiate system
I remember spending hours and hours on pivot and forums like droidz.org which used to host animations, models, and forums. I even remember learning about "easing" and different "levels" of animation and collabs as showcased on darkdemon.org[1]
Also want to shout out ts-rest. We have a typescript monorepo where the backend and frontend import the api contract from a shared package, making frontend integration both type-safe and dead simple.
System prompts themselves have many contradictions. I remember hearing an Anthropic engineer (possibly Lex Fridman's interview with Amanda Askell) talking about using exaggerated language like "NEVER" just to steer Claude to rarely do something.
