3 ms·
How does that work for image generation AI? And even with LLMs the prompts are going to contain extremely limited information because the users are primarily re
by weare138 3y ago
How does that work for image generation AI? And even with LLMs the prompts are going to contain extremely limited information because the users are primarily requesting information, not providing it.
AI is only as good as its models. Now that we're flooding the internet with AI generated content with no way to determine if the content was also produced by AI while endlessly shoveling as much content as possible into these AI models to push up those parameter counts to attract investors, it's inevitable the current generation of popular AI systems will just eat themselves.
- visarga 3y agoLLM interactions can be more than a simple question & answer. Long chats allow the human ample opportunity to steer the model in a direction outside its normal scope. If they also use RAG or other tools like code execution there is even more ample chance to steer the model outside its original experience. Then we can use these sessions as training data to re-tune the base model. It's much cheaper than training a new model, can be done on normal computers. LLMs are not beholden to their training set. They can ingest new information any time, or relate two things that were previously separate to derive new insights.