Around 5 years ago we were very proud of these state of the art results in image generation, trained on 32×32 "images" of CIFAR-10. You can kind of make out little wheel shapes, car/plane parts, and organic structures and textures. Pretty cool right
INNOVATION
-
Did OpenAI Follow Lean Startup Methodology Principles?
By
–
For example, maybe it’s an exception but do you think OAI followed the lean startup methodology?
-
Long Tail Applications Drive AI Needs and Cost Reduction
By
–
Needs are driven by how long tail the end application is. And costs will go down over time too.
-
AI Models Focus Shifts to Reliability Over New Capabilities
By
–
It's not just about what you couldn't have done with prior models. It's about doing what you can already do, but a lot more reliabily.
-
Training Models: Foundation for Effective Testing Processes
By
–
Because if you don’t train it, then you have nothing to test?
-
Karpathy Building JARVIS Follows Nakajima at OpenAI
By
–
@karpathy (Building JARVIS @OpenAI) is now following @yoheinakajima
-
Building from Scratch vs Customizing Libraries in Development
By
–
I seem I have an easier time building from ground up, vs customizing a library – felt like my code was becoming increasingly hard to read and inefficient.
-
Alternatives to Langchain and ZapierNLA for AI Integration
By
–
I’ve tried Langchain and ZapierNLA, any other suggestions?
-
AI Video Creation Engine Transforms Ideas Into Content
By
–
Have content and ideas, but no video?
— Yohei (@yoheinakajima) 2 avril 2023
Head to this AI video creation engine! By default it uses existing media from across the web, but you can generate them on the fly as well.
Idea to video in 1 min🎥 https://t.co/z9cFbeIylVHave content and ideas, but no video? Head to this AI video creation engine! By default it uses existing media from across the web, but you can generate them on the fly as well. Idea to video in 1 min
-
Generated Code Replacement and Program Decomposition Strategy
By
–
Aha ok. As far as I know we are replacing the whole generated code, not parts of it. Decomposition of the program is one of our next steps, but I also think it’s much harder.