4 ms·
Hey HN! I'm very proud to release the deepest interview deep dive into the SAM model I can find on the Internet (seriously - i looked on youtube and listennotes
by swyx 4y ago
Hey HN! I'm very proud to release the deepest interview deep dive into the SAM model I can find on the Internet (seriously - i looked on youtube and listennotes and all of them were pretty superficial). the Roboflow team has spent the past week hacking on and building with SAM and I ran into Joseph Nelson this weekend and realized he might be the perfect non-Meta-AI person to discuss what it means for developers building with SAM.
so.. enjoy! worked really hard on the prep and editing, any feedback and suggestions/recommendations welcome. still new to AI and new to the podcast game.
edit: Video demo is here in case people miss it
https://youtu.be/SZQSF-A-WkA https://youtu.be/SZQSF-A-WkA
- re5i5tor 4y agoReally great, thank you.
- swyx 4y agoany requests for under-covered topics? i felt like this one resonated because somehow the other podcasters/youtubers seemed to miss how big of a deal it was. hungry for more.
- re5i5tor 4y agoYeah, the podcast + demo really got to the heart of what was game changing, explained very well. No additional requests yet!
- CyberDildonics 4y agoIt's easy to have someone lay down 30 points for a simple banana shaped outline and compare segmentation to that, but how does this compare to other automatic techniques like spectral matting (which is now 16 years old) ? http://people.csail.mit.edu/alevin/papers/spectral-matting-levin-etal-cvpr07.pdf http://people.csail.mit.edu/alevin/papers/spectral-matting-l...
- teruakohatu 4y agoDeep methods are a vast improvement over classical computer vision techniques. Classical techniques can be thought of as a function of the raw pixel data. Deep learning techniques understands the context and are more likely to get segment how a human might segment. spectral matting, as I understand it, is used for subject/foreground and background separation.
- CyberDildonics 4y agoIt's easy to say something is better, but good computer graphics and computer vision papers compare themselves to the state of the art. This is ignoring all that and comparing itself to the most manual method possible. This also happens sometimes and is more of a marketing stunt to people who don't follow current research.
- Eisenstein 4y agoWhy is there no volume control on your podcast page player?
- HanClinto 4y agoExcellent episode, I really enjoyed this. Thank you!
- swyx 4y agothanks for listening! feedback welcome, seriously dont be shy