It’s like watching a disaster about to happen, but not really soon. Later in, when all the actually programmers move on or retire and the generated code slop needs experienced intervention.
AI will improve more quickly than senior engineers will age.
Maybe, maybe not. There seems to be polarization on this topic not unlike most other topics.
People who believe hyper-advanced AI and AGI are inevitable within years.
People who believe both are attainable, but current models are peaking and will hit terminal progress soon like a wall (due to physical or more likely ethical reasons)
People who couldn’t care either way.
Reality is likely to be closer to the 2nd option unless the handlers are very short sighted.
Any AI that is allowed to get advanced enough to:
- formulate its own goals
- fabricate the means to attain those goals
- learn and improve the methods used
Will have to be abruptly terminated or humanity will be and none of this matters. For an entity capable of such things, we would inevitably phased out as an inefficiency.