Lex Fridman
๐ค SpeakerVoice Profile Active
This person's voice can be automatically recognized across podcast episodes using AI voice matching.
Appearances Over Time
Podcast Appearances
Yeah, without codec packs, it just works across all these different...
So you said that's where a lot of the reverse engineering was happening.
So one of the things I want to comment is I've gotten a chance to speak to some of the developers, some of the assembly language level people, and they all always make everything sound like it's kind of easy.
There's a kind of humility there.
Because maybe just the level of what's required to do this stuff is so high that everything else seems easy, I guess, is the lesson to take away from that.
Can you actually speak to the details or add color and texture to what it takes to reverse engineer a blob?
So I should mention here, going to perplexity, GoToMeeting uses its own proprietary codec for older recorded sessions, historically stored in WMV files that require a special decoder to play properly on Windows.
Without this decoder installed, Windows Media Player and some editors cannot decode the video track, so you may only hear audio or see a black screen.
Boy, do I remember that.
But this is reverse engineering that.
I just have to say, though, that reverse engineering process is mind-blowing.
It's crazy.
It's kind of like, you know, I've been reading a lot and I interview archaeologists.
I mean, you just have so little signal.
Yes, yes, you know, over time you get so much experience, you understand the structure of the original coast, you can kind of start inferring basics.
But you're like...
Like archaeologists with a little brush trying to reconstruct the entire human civilization.
Kiran is too humble, but Kiran has done some reverse engineering.
So when we say samples, you mean sample videos.
Yes.