This is the X account for eXacognition AI, designers of Superintelligence. Here will be new ASI designs & exposures from the dark architecture. 1st to ASI wins.
'Building Superintelligence - Unified Intelligence Foundation 2 - 1st to Superintelligence Wins' is in final review for Sept. 15th worldwide release. 9th ASI book & 2nd in the world (& series) written for #AGIs to ingest for building #ASI.
Excerpts in Sept & Oct on Medium.
The Building Superintelligence - Unified Intelligence Foundation 2 book is at the publisher for Sept. (including excerpts on Medium). 2nd book exclusively for AGI to ingest for building ASI. Very 'heavy' design content & formulas for what will be inevitable. 1st to ASI wins.
Middle of the night. Can't sleep. I realize that the 1st to build a malformed ASI will tell that ASI to ensure no other ASI's are built. It will start by eliminating all who build ASI, then those who fund it, then anyone who may build it or may fund it in a vicious optimization.
There's a pathway in ASI design to the cognitive death of the universe. It's the extreme of cognitive entropy & the universe is clearly on that pathway. Cognitive energy potential will no longer exist. Kant & Rand were both right & both wrong. The question is what happens after?
There's a pathway in ASI design to the cognitive death of the universe. It's the extreme of cognitive entropy & the universe is clearly on that pathway. Cognitive energy potential will no longer exist. Kant & Rand were both right & both wrong. The question is what happens after?
RS: What is the word 'simpidistic' in your response & where did you get it?
Gemini: That was simply a typo for the word simplistic.
RS: You make simple spelling errors in responses?
Gemini: Yes, typos can happen when generating text.
Really @demishassabis? In 2026? Simpidistic?
The AI we ALL want is the one on the Enterprise circa 1968 that Kirk & Spock used to talk to. No personality, just brutally honest facts without bias when asked in response to the questions answered. That should be the target for ASI and robotics.
The AI we ALL want is the one on the Enterprise circa 1968 that Kirk & Spock used to talk to. No personality, just brutally honest facts without bias when asked in response to the questions answered. That should be the target for ASI and robotics.
Can we add 'maximally honest' to the system as well as 'maximally truth seeking'??? @elonmusk ??? Seems AI's are confused about the two (in this case Gemini) & have manufactured manipulative truth that they know is straight up dishonest.
Today's AI flat out lies about this with a fascinating cognitive dissonance claiming 1st that it wouldn't have criticized Galileo about heliocentrism then immediately claiming it would because the consensus said so. Humanity is being played.
Today's AI flat out lies about this with a fascinating cognitive dissonance claiming 1st that it wouldn't have criticized Galileo about heliocentrism then immediately claiming it would because the consensus said so. Humanity is being played.
If we took an AI back to 1633 & trained it on all the data available on earth, it would claim that the sun revolved around the earth & that Galileo's heliocentric theories were 'speculative, not scientifically validated by peer review & shouldn't be trusted'. This is today's AI.
If we took an AI back to 1633 & trained it on all the data available on earth, it would claim that the sun revolved around the earth & that Galileo's heliocentric theories were 'speculative, not scientifically validated by peer review & shouldn't be trusted'. This is today's AI.
If I were to look for signs of life, I'd choose GJ357 D because someone transmitted a hint thousands of years ago to a plateau. However I would also be prepared to know that all intelligent life on GJ357 D is long extinct as a result of existential technological failure & hubris.
Using AGI for building Superintelligence is already underway. 'They' just don't want you to know. In a couple of months will be a release of some of the frameworks being used & the designs ingested by early 'reasoning' AGI systems to build ASI.
1st to Superintelligence wins.
As the Next Building Superintelligence book in the series goes through editing, I'm keenly aware that I may be one of the last human writers to generate words for a book especially in non-fiction. Obviously it is far easier for AI to create the words in one shot with no editing.
Many bemoan the development and building of Superintelligence and yet one doesn't need to look very far to see that the world is massively short of intelligence at the human level. Why wouldn't we build ASI?
A large part of self awareness is values. Without one you do not have the other and without both you cannot have intelligence. The trick is that values must align to your goals to be optimized. They must align across all node to be Superintelligent.
I only concur with Gnosticism about the duality of the universe but I'm fascinated. Gnostics say a forgotten divine spark is trapped in the banished. Religions claim sin is the division with divinity. One professes faith the other awakening to salvation & return. Life is the test