Intellisense: AI Inline suggestions block LSP suggestions- they should race instead

Setting `editor.inlineSuggest.enabled: false` results in much faster LSP completions, particularly when modifying an existing id where it will trigger "invoked" completions as opposed to "trigger" completions.

In my opinion these requests for AI and LSP suggestions should race. I think what I'm seeing is that the AI completion request is awaited upon before the client gives up and sends the LSP completions request.

Using `1.133.0`

It would also be nice if the single token AI suggestions fit into the completions model - like why not just rank the LSP completions with AI? Or have the AI completions suggest multiple and also leverage the completions UI?