I know this is not to be taken too seriously but I don't even buy the premise: Why would those metrics predict the singularity?
Tokens will not suddenly become free (as in beer) – AI will still consume energy –, there will still be a finite amount of arXiv papers mentioning emergence (for not just one reason), release frequency of frontier models is not necessarily related to breakthroughs (model capabilities might also approach an asymptote), Copilot code share increasing will likely happen over time anyway even if all AI research stopped today. Etc.
Tokens will not suddenly become free (as in beer) – AI will still consume energy –, there will still be a finite amount of arXiv papers mentioning emergence (for not just one reason), release frequency of frontier models is not necessarily related to breakthroughs (model capabilities might also approach an asymptote), Copilot code share increasing will likely happen over time anyway even if all AI research stopped today. Etc.