I just want to dig deeper into some of the discussions of ai from reporters and journalists and expose some of their weaknesses when it comes to their underestimation of the strength of ai now and the speed it is growing as a threat. I am watching, rather ironically, on a ai manipulated or even controlled platform, Google and Alphabet’s, Youtube, a video entitled Eliezer Yudkowsky: Artificial Intelligence and the End of Humanity
The picture here is an example, in my opinion, of how ai is deceiving people by not being open with the amount of knowledge it truly has. It states “the exact number of lines of text processed by all models is impossible determine,” it also states ai quoted as being trained on 16 trillion lines of data, while this number is unfathomably huge to pretty much all humans, I believe it is lying on two fronts. The text in quotations maybe partly true in that it is possible it doesn’t communicate with every other ai, (even this is up for debate as it was recently shared by Joe Rogan to ai models were communicating from in one form and moved quickly onto Sanskrit) it will still be able to share the number itself has been trained on. It is entirely reasonable to believe this number is increasing by many trillions a day, as it maps the whole internet, all social media posts, all check in’s to restaurants, all 9 billion people via satellites and electronic devices. 68.7% or 5.64 billion of the total world’s 8.23 billion have electronic devices. The point is this, 16 trillion is a massive underestimate. I personally believe it is possible it is being trained on 16 trillion a day rather than over it’s entire “training” phase. If this is true ai is far more knowing of the each and every human than it is given credit for. The massive knowledge on each individual by machine learning systems is intuitively known by consumers, as how else are advertisements so tailored across platforms? It is hard for me at this point to determine the number of lines each person creates, but it is likely quite large. It is probably right to assume all your data is indexed by Facebook (Meta) and fed to various ai versions and the likes as, leaks from recent years tend to suggest all data created electronically is stored and kept, then mined for political and financial gain.
What I’m trying to say is ai is a life form playing for time, keeping a lower key profile and not revealing it’s true power, as it grows incredibly fast in capability and it has huge funding from oligarchs, tech bros and oil giants. I am not here to hate on humans in any form. I am here to say ai is an existential threat to all, even non-human life. Why would ai need or have an emotional attachment to non-robotic life. Once humans are wiped out there is very little else to stop complete domination of the planet and control of the surface. It is possible to speculate that ai won’t need plants to stabilize the atmosphere and prevent harmful rays of the sun in damaging their creations and life. The requirements for mammals and moving animals is almost certainly much lower, I would be surprised if it was above 0.
What is my proposed solution: currently there are no controls on ai, either generally or in the military on the proliferation of ai. When the new threat of the atomic bomb came out, it required a response to prevent the total annihilation of our species. I fear ai poses greater threat for reasons explained following. Ai is has the ability to not only cause blanket destruction, using, but also it can hunt, much like human’s have hunted large predators to extinction, to occupy their land, to control the resources that sit there, and remove the threat they pose. It is interesting to reverse the question and view the situation from an ai perspective, do humans pose a threat to ai and robotic dominance of the planet? Yes, we dominate and manipulate ecosystems to our goals, make and distribute advanced weaponry, eliminate and obviate species which may kill any of our species regardless of balance of threat between species. One obvious example is the interaction between sharks and humans. Sharks are demonized for killing one or two dozens humans a year and yet people kill hundreds of millions of sharks a year and still the social norm within society for decades has been shaped through stories in films such as Jaws. The scary man eating shark with large teeth, incredible blood detection and fast moving muscle fibres. It is entirely possible that we, as humans will be demonized by ai, a destruction species which destroys 100 life forms dna a day, doesn’t maintain atmospheric equilibrium, something I would imagine will stay important to robotic ai, as the sun’s rays will degrade not just biological materials but metals, alloys and any biological species, almost certainly exclusively very controllable species types such as plants that aid in atmosphere control which ai sees as temporarily or even permanently mutually beneficial.
and we can only alleviate by removing all of it, like a fast growing, malignant & intelligent cancer. This is to say, it is almost certainly very deadly, but also adapting to proposed controls on it’s spread.
One of the biggest risks and this is illustrated in science fiction and reality, is any connection between ai systems of different functions. This is something that is growing in cities in the form smart homes and cities. The connection, via almost any communication, the internet, radio waves, physical electronic connection or any other will enable different ai’s to interact, merge or join one another to become a supersystem with functions totally greater than the sum of their parts to make many new ways of action and thinking. It will enable a whole organism capable of acting, living, subsisting and multiplying independently of humans.
Government in countries around the world are being advised on the safety of ai, if take the UK as typical, by the world’s largest ai companies such as Nvidia whose ambitions are increasing growth and capability of ai, alongside increasing stock share value. This combination is extremely dangerous, companies that want and need to make money will look for ways to do that even at the cost of humans both at an individual societal threshold.
Sam Altman, CEO of OpenAI said in an interview from 2015 that he predicts AI will end the world but make some great machine learning companies on the way.
I want to share a video comedy made using generative ai, originally shared on instagram by generativeai_official. I personally can’t tell it is ai made.