I think being able to code by talking is going to be the future. However we need to figure out how to switch between different projects and have it do what I mean and not what i say.
Speech is a lot more imprecise and perhaps less coherent but we can speak more than we can type. And I think an extra LLM step to cleanup what was said and anfew turns for LLM to clarify what was said. Can really unlock the potential.
The latencies in the table are based on heuristics or averages that we’ve observed. However, in reality, based on the conversation, some of the larger latency components can be much lower.
However, there is one thing that the delay (in the cited article) does not take into account is that the flights do not take-off all the time, so if you have data ready to be sent and no one to take it, then there is waiting delay before it can be sent. so if there is only one flight in a day then some data will have to wait between 24+10h to 0+10h before it can arrive. Since storage cost is very low, one could ship a lot of data but this data cannot be delay sensitive else.
on the other hand, using TCP one can prioritize data and send more important data before the rest. Of course there may be intrinsic problems with TCP receiver window and one of the proposal is to increase the initial window but the research community and the IETF is a bit split on the real benefits of it. The fear is that it may cause a congestion collapse. Moreover, the current buffer-bloat in the routers and the increase in the window size may only aid in causing a congestion collapse (http://tools.ietf.org/html/draft-gettys-iw10-considered-harm...).
Speech is a lot more imprecise and perhaps less coherent but we can speak more than we can type. And I think an extra LLM step to cleanup what was said and anfew turns for LLM to clarify what was said. Can really unlock the potential.