Osman R. (@UsmanReads) header
Osman R. (@UsmanReads) profile picture

I think I know, but I really don't. AI and Tech with 15 years in Industry. prev @groupon and @toptal

Universal
Joined March 2023
Photo by Osman R. (@UsmanReads)Photo by Osman R. (@UsmanReads)Photo by Osman R. (@UsmanReads)Photo by Osman R. (@UsmanReads)Photo by Osman R. (@UsmanReads)Photo by Osman R. (@UsmanReads)Photo by Osman R. (@UsmanReads)Photo by Osman R. (@UsmanReads)Photo by Osman R. (@UsmanReads)Photo by Osman R. (@UsmanReads)
The internet didn’t make everyone a creator. It made everyone an audience, and a few people with cameras the creators. We built ContentRoof for the other 99%. Your words in, a finished film out, nine languages, minutes. This is our film about why. Visit contentroof.com
1
1
1
122
AI video generation just got 5.1× faster. At ContentRoof, we optimized our rendering engine to deliver: • 5.1× faster rendering • 80% lower costs • Same visual quality Less waiting. More creating. See how we did it: contentroof.com/lab/the-soft… #generativeai #ai #Startups
2
1
3
42
Cloudflare is down!!
37
help me @amasad - how did you do it with @Replit ? hey @ReplitSupport - can you ask him for me?
seeking advise from founders - my tool is getting traction -50 sign ups a day -200 visits - conversion rate is good -60 videos a day but 0 paying customer, how do I get them convert?
1
38
seeking advise from founders - my tool is getting traction -50 sign ups a day -200 visits - conversion rate is good -60 videos a day but 0 paying customer, how do I get them convert?
4
1
1
108
I’ve spent the last few weeks building an app that turns a script into a narrated, animated short video, in 9 languages. Most of what I believed going in turned out wrong. Here’s what actually worked, what flopped, and the numbers behind it 🧵
1
58
Rendering video in Chrome: 80% of each frame was the screenshot, not my React code. My “optimization” made it slower. So I wrote my own Skia renderer. A 55s video: 218s → 90s. On an M4 GPU: 67s → 13s. Profile before you optimize.
1
9
The pattern behind all of this: whenever I reached for a bigger model, a measurement, a filter or a tiny trained model won. Build the eval first. I’m building this in public, so follow along. Make your first short free: contentroof.com
3