Let me summarize: – users don’t tend to generate 100 codebases or videos and pick the one that is perfect for them. They want fine-grained edit capability. – users tend to generate 100 images or 100 text pieces and pick nearest match to what they want.
GENERATIVE AI
-
Generating Infinite Variants: Why Code and Video Differ from Images
By
–
With images and text, it seems perfectly reasonable to generate infinite variants till you’re happy. Code and video require too much manipulation of generations if you want changes and that requires skill.
-
Image generation versus code video control trade-offs
By
–
With images, the end users can obviously pull generations into photoshop and edit, but they tend to very often just generate *MORE* variations and just pick 1 Basically with code and video nobody wants to sacrifice control and it doesn’t make sense to generate infinite variation
-
ImageAI Differs from CodeAI and VideoAI in Usage Requirements
By
–
ImageAI is different in the way it is used versus codeAI or videoAI. If you want to modify anything that code AI generates, you need to understand code. Same w video. The output generated will never be perfect and you will want to edit some files (code) or frames (video). More
-
Meta Lattice: New Model Architecture for Enhanced Ads Performance
By
–
Meta Lattice is a new model architecture that improves Meta’s ads systems performance, AI efficiency and enables faster adaptability to the shifting market landscape.
— AI at Meta (@AIatMeta) 2 juin 2023
More on this new work ⬇️Meta Lattice is a new model architecture that improves Meta’s ads systems performance, AI efficiency and enables faster adaptability to the shifting market landscape. More on this new work
-
Meta LLaMA regulation policy and foundation model access trends
By
–
Whether that's good or not depends on your perspective, but the numbers are constantly going up. We'll see if that pushes Meta to do something more permissive with LLaMA, or how OpenAI and other foundation model developers asking for increased regulation react.
-
Model Performance Comparison: Falcon Leaderboard Ranking Analysis
By
–
That's probably not the case! Different models perform well at different things, and the harness is an *average* for a reason. But it's notable just how much effort went into highlighting that Falcon topped said leaderboard.
-
Leaderboard obsession: Why 0.1 point differences mislead AI evaluation
By
–
Leaderboards are *a thing* in *every industry.* I can tell you as a journalist we all have an obsession with leaderboards. But we'll probably hear a lot more going forward about how Model A beats Model B by 0.1 points in this eval model, so A is obviously obsolete.
-
Falcon’s Cultural Impact on Open Source Model Competition
By
–
But Falcon might also have a separate cultural impact crater. It aggressively highlighted that it topped the Hugging Face leaderboard in terms of average performance on the Eleuther AI model eval harness. And we are probably going to see a lot more of that in OSS going forward.
-
Falcon Becomes First Permissive Open-Source Model to Beat LLaMA
By
–
Falcon is now the first permissive open-source model to "beat" LLaMA, which is still for research purposes. There's been a lot of buzz around open source models, but we haven't really seen them in prod at the scale of a GPT-4 (which most people I talk to use out of convenience.)