gautam

building fakecardboard(dot)ai (chill its for my resume)

[gautam @gautxmmmm]

the bottleneck is this might not work with large files because of the time it's taking to analyze the files properly - hmm interesting

[gautam @gautxmmmm]

shot on RED in 8k yeah pack it up lil bro no one else is getting that job

[Vikrant Patankar @vikpat]

directed & produced my biggest brand film yet!!

for one of the best designers out there... who also happens to be my younger sister 💖

shot on RED in 8K, with a full crew and a sound stage turned into Barbieland.

presenting: DESIGN IS DEAD

[gautam @gautxmmmm]

done with with my week oncall - back to the grind

[gautam @gautxmmmm]

@ishandeveloper so cool 😮‍💨

[gautam @gautxmmmm]

@sxmawl @istgishaaaan crazyy

[gautam @gautxmmmm]

working on your personal projects when you are oncall alternate weeks is tough :(

[gautam @gautxmmmm]

@sxmawl instructions unclear - i have 5 tasks pending by EOW

[gautam @gautxmmmm]

deepseek 4.1 flash and opus 5.5 are incredibly close in accuracy - 5.5 is just fast af but does not justify the cost difference between them

[gautam @gautxmmmm]

@Joshua_WD thanks :)

[gautam @gautxmmmm]

a week of building an ai video editor at 2am. beat detection, 60x faster renders, music, intelligent cuts. not done. not close to done. not stopping. rendered video soon. fakecardboard[dot]ai is happening lol.

[gautam @gautxmmmm]

@sxmawl @usecardboard yes the changelogs were around 10 days apart in january - focused more on the changes not the dates after that haha:)

[gautam @gautxmmmm]

@usecardboard ships a changelog every 10 days. Let's see what I ship in 10 days on a 5-year-old laptop and free opencode models. Next 10 days are going to be interesting.

[gautam @gautxmmmm]

building my own @usecardboard as a personal project because why not 🤷‍♂️ stay tuned :)

[gautam @gautxmmmm]

@sxmawl hehe dw not close to cardboard, yet.

[gautam @gautxmmmm]

day 3

progress on the heuristic and LLM algos - want my fakecardboard to make a sick edit

[gautam @gautxmmmm]

i need to figure out how to make the edits LLM based instead of being heuristic for better results

[gautam @gautxmmmm]

Day 1/7

1.23 GB of raw footage in.

65 seconds of vlog out.

cuts on the silence boundaries, decent color grade, audio mixed underneath in 11 minutes.

I asked for "45-60 second cinematic vlog".

gonna sleep now. see ya day 2.

@usecardboard

[gautam @gautxmmmm]

@ishandeveloper @usecardboard let’s see what i ship till your next changelog then 👀

[gautam @gautxmmmm]

building my own @usecardboard as a personal project because why not 🤷‍♂️ stay tuned :)

[Naval @naval]

No one can fill the hole left by another human.

[Tanay Anand @anand_404]

After 2.5 years at Sarvam, I am starting something new!

I had a great time leading the RL team. We launched the 30B and 105B models back in Feb with a young team building everything from scratch. Excited to see the team keep growing and pushing the frontier. Grateful to @pratykumar for the early bet and trust.

As for what's next: I am convinced that there are a few critical parts of the stack that we still need to build to meaningfully accelerate science. One of them is verification.

Scientific discovery only scales with compute when proposing and verifying new ideas is fast and cheap. In math and code, we're starting to see what happens when models can generate ideas and reliably check them at scale. OpenAI's recent Navier-Stokes work is a glimpse of what that looks like.

Similar moments in chemistry, materials, biology and the rest of the physical sciences would require verification to get dramatically faster and cheaper. It cannot be bottlenecked by the throughput of physical labs

More soon. Very excited for what's next.