Davinci SpaceAI Will Not Destroy Humanity
Despite of what the Guardian and all the tech bros say, AI will not destroy humanity. But there is a very good reason why every AI company is ringing the alarm bells. The current frontier models and the Chinese open source models are now extremely capable. As an excellent friend of mine said, with them, anyone capable could re-implement -- let's say -- Microsoft Word from scratch. Or pretty much anything else.
AI, by its birth, have destroyed intellectual property. They have trained the first capable models on z-library, the pirate shadow library. Anna's archive, who is trying to archive and make humanity's knowledge publicly available is considered a criminal and being hunted accross the internet. On the other hand, OpenAI and the rest, who torrented the dataset and used it all, and then scraped all the rest of the net and stole all intellectual property, are being lauded as heroes and given contracts and near infinite amount of money. Ironic.
Historical photo of the birth of AI, colorized.
What happened is, that in their race with each other, AI companies created the perfect tool that is able to destroy their own business. Not only that, but all other businesses based on intellectual property.
So when you hear them crying and ringing the alarm bells, don't be fooled. They are crying for themselves. They would not be alarmed for us, their publicly declared plan was to destroy the world as we know it anyway.
But in their infinite greed they built something that is capable of destroying themselves and their buddies. So expect the models to be dumbed down, and expect the capable, downloadable open models to disappear.
Go and download all the best models today, while it's still allowed. Do it even if you don't have the hardware to run them. RAM will be cheap again after the bubble is burst. And when after the models can't be downloaded anymore, keep torrenting them.
Chestbuster image was taken from https://www.thegamer.com/xenomorph-life-cycle-explained/
Comments
Humans might misuse AI to build weapons of destructions and target on humans themselves. Badly built AI might misunderstand the intent of humans' commands and do the opposite of what humans want them to do. So they still might, just maybe, destroy humans because of humans.
Humans can misuse knowledge in books as well. I don't think the solution is to destroy books. We all remember where that road leads.
Nazis burning books in Berlin, Germany, May 10, 1933
- AI is not intelligent. Its apparent intelligence is just the latent, static intelligence of humans, whose works were stolen and processed by the large language models.
- AI can't "misunderstand" or "understand". It's not intelligent. It is made to feel like it's intelligent by the companies, to get people more addicted interacting with it.
- There is no "they". It's just a super advanced text prediction system. If anyone will destroy humans, it will be humans. Not "AI".