Certain scientific applications are one of the better uses of machine learning techniques. For example, an extremely long black hole jet was found by having machine learning go over a ton of data from radio telescopes.
It’s so long that the jet could not only hit another galaxy over, and not just a galaxy in a different cluster or even supercluster, but could go through the void in the filament structure of galaxies and hit another filament over.
LLM are not the only way and more people need to realise that. There are good uses for LLMs but most use cases want different ML approaches especially when we want numbers.
So… Just like general-purpose software development? I hope you get to work without LLMs.
Certain scientific applications are one of the better uses of machine learning techniques. For example, an extremely long black hole jet was found by having machine learning go over a ton of data from radio telescopes.
It’s so long that the jet could not only hit another galaxy over, and not just a galaxy in a different cluster or even supercluster, but could go through the void in the filament structure of galaxies and hit another filament over.
LLM are not the only way and more people need to realise that. There are good uses for LLMs but most use cases want different ML approaches especially when we want numbers.