bard summed it up for me;

Sure, here is a summary of the article "AI Will Eat Itself? This AI Paper Introduces a Phenomenon Called Model Collapse That Refers to a Degenerative Learning Process Where Models Start Forgetting Improbable Events Over Time":

• Model collapse is a phenomenon where large language models (LLMs) start to forget improbable events over time.
• This can happen when LLMs are trained on massive datasets of text and code, which often contain a lot of improbable events.
• As LLMs learn to predict the most likely next word or phrase, they may start to ignore improbable events, which can lead to them becoming less accurate over time.
• The authors of the paper propose a number of methods to prevent model collapse, including using a more diverse dataset and using a different training objective.

Conclusion:

Model collapse is a potential problem for LLMs, but it can be prevented by using a more diverse dataset and a different training objective.

Here are some additional points from the article:

• The authors of the paper believe that model collapse is a serious problem that could limit the usefulness of LLMs.
• They argue that LLMs need to be able to handle improbable events in order to be truly useful.
• They propose a number of methods to prevent model collapse, but they acknowledge that these methods are not perfect.