/post#cm8mon9g

AI Model History is Being Lost

I fear that we may lose – and are already losing – the modern history of AI, particularly with regard to the ability to use models that are/were only available in hosted capacities. As AI models are retired and become inaccessible, we’re losing crucial historical context about AI development that we will be unable to go back and reference in the future.

This essay examines the disappearance of proprietary hosted AI models and the implications of that on research, accountability, and overall historical documentation and preservation.

...

https://vale.rocks/posts/ai-model-history-is-being-lost

10d|

Vale!089a4b5364

|reply||

1 comment

Wouldn't it be enough to preserve outputs? Maybe a standard test for all the ephemeral models that will be created and destroyed. Sure, we won't have an accurate reproduction down the road when we retrain on output, but distillation is a thing and worked out ok for deepseek. If gpt dies out, we'll still have whatever synthetic datasets have been trained on.

10d|

Anonymous

||

Add a comment