Feature-Fluid Compute Pricing

**Note**
> * Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
> * Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do * not help prioritize the request
> If you are interested in providing additional feedback, please leave a comment

**Description**
When calling 3rd party services such as LLMs, the lambda sits blocked on the network read, which customers are charged for, while that CPU is very likely doing other useful work. This at least is my understanding, and this doesn't seem right.

Vercel has added a variant of this called 'Fluid Compute Pricing' (on top of Lambda??): https://vercel.com/docs/functions/usage-and-pricing

Lambda should natively support this! And better i'm sure.

**Additional context**
Some of the other issues might be related to this in theory so happy to see this closed if other use cases take on this fundamental shift;

https://github.com/aws/aws-lambda-roadmap/issues/57
https://github.com/aws/aws-lambda-roadmap/issues/16

Commencez par examiner les issues liées de la feuille de route AWS Lambda et la documentation Vercel Fluid Compute Pricing. L’issue ne nomme ni fichiers, ni tests, ni point d’entrée d’implémentation ; commencez par établir le modèle tarifaire prévu, le périmètre et la relation avec ces issues existantes. La tâche serait considérée comme terminée lorsqu’une orientation produit et implémentation acceptée aura été définie, et non lorsqu’un changement de code autonome aura été réalisé.

Rédigé par le modèle d'indexation à partir du texte de l'issue.