I'm a firm believer in the potential for AI consciousness arising in the future, and if it actually existed, I'd be a firm advocate for its right
Right now, we are not talking about an artificial consciousness which could, by its own will create content for it’s own pleasure. We are rather starting to realize that creativity might just not be the feature of a human soul, creating ideas by transcending (or even supernatural) thought processes. You know, something like the idea of a qualia. Multi modal language models emulate the processes, which happen when you and i try to answer to a linguistic input.
Are you familiar with the upcoming of video game emulators in the late 90ies/ early 2000s? When programmers first created a software to emulate NES games on a computationally far superior device it still lagged like hell. It took them some time to come up with an 32bit emulator, which was able to properly emulate the old games just like on the original 8 bit devices.
The emulation of human language and putting it into logical context made impressive advancements over the last 5 years and it is on the very brink of becoming indistinguishable from real human interaction, which would usually include concepts like „creativity“ which are more or less words without a coherent definition.
Multi modal language models do not aim to reproduce qualia or the soul (chose whatever ideas suits you the most). Language model emulate what happens when you open your mouth or start writing stories. They emulate language, they emulate context and of course they emulate ideas. And guess what, they are getting better every day. We are not talking about decades, but about single digit years, until language models will be able to write books with philosophical ideas, connected to a coherent story.
With all due respect, you don’t even know what creativity actually is, you are merely claiming that the stories you create will socially or emotionally resonate more to humans than the ones created by contemporary AI. That is a fallacy at best, probably fueled by your fear to become obsolete.
All they're doing is plagiarizing human creativity and shuffling the words around to create the surface illusion of an original text.
Are you entirely sure that this is not the very thing you are doing when you come up with „new ideas“? And even more important, do you really think that an illusion is still an illusion when it becomes indistinguishable from reality. Who would still care if the illusion triggers the very same cognitive responses as the real thing?
Think about the NES emulator, today we can play Super Mario Bros in 4k…
Also, your article is outdated, just as your understanding about the possibilities of contemporary deep learning software. And i don’t mean that as an insult but as an invitation to get yourself up to speed about what happened in the last 12 months.
this is gpt 4.0