the holy grail is to not have a separate train and inference steps. when the model can be updated while it is inferencing is where we're headed. deepseek only accelerates the need for more compute, not less
THIS is the only correct statement in all of this.
The goal for AGI and ASI MUST BE to train, inference, train, inference and so on and that all on the fly in fractions of a second from every token produced.
Now good luck calculating the compute and hard work in algorithms to get there.
Not possible? Then AGI won't ever work because how can AGI beat a human if it can't learn on the fly? Not to mention ASI lol.
The goal for AGI and ASI MUST BE to train, inference, train, inference and so on and that all on the fly in fractions of a second from every token produced.
Now good luck calculating the compute and hard work in algorithms to get there.
Not possible? Then AGI won't ever work because how can AGI beat a human if it can't learn on the fly? Not to mention ASI lol.