Suttit på den här läckorna ett tag, kan lova att de är äkta. Vad jag vet finns de inte längre online, men jag har de.Frågar man GPT nekar den allt om detta, men laddar man upp tex PDF om detta och frågar om de så ekänner den att de finns likheter... Jag tror detta kan bli mycket stort, openai vill säkerligen inte att denna info kommer ut. ChatGPT är baserad på en teknik från 2006 som heter Lexical Clone, gjord utav själv Ilya Sutyesker! (Han använder dock ett psedonym i de flesta texter om detta). Vad jag vet finns denna inte info offentlig online längre, förutom i de PDF'r from 2004-2006.
The Basics:
ChatGPT är i grunden en sökteknik som söker igenom content, och återberättar den genom ny context(Prompten). Själva illusionen av "intelligens" härstammar enbart från hur modellen är tränad på återberättande utav sökresultaten. En model kan än så länge inte generalisera utanför dess traning data, försök till detta är finetuning, men iom det inte finns weights kopplade till de tokens så börjar den istället gissa, därav hallusinera. Metoder såsom grokking och chain of thoughts hare visat sig vara mycket effektiva till att få en LLM att kunna generalisera bättre.
Så låt mig plocka ut några av de fina bitarna.

Resterande av de läckar jag samlat på mig:
Läser man på om lexiclone ser man att det faktiskt stämmer, se:
https://www.researchgate.net/publication/221038501_LexiClone_Lexical_Cloning_Systems
Authors: Ilya S. Geller
Källor: https://www.semanticscholar.org/paper/LexiClone-Lexical-Cloning-Systems-Geller/dd551d546e1c14b3e6f9023cead068f822a82a93
https://slideplayer.com/slide/6119404/
Jag vill personligen också poängtera att när Blombergs/Reuturs skrev GPTs nya "search function" som skulle bli annonserad i maj var jag först ut att säga att det inte var fake(jag visste helt enkelt). Jag blev shadowbannad på 2 andra dev communities, men så fort det visat sig att jag hade helt rätt så togs shadowbannen bort och jag fick en VIP roll. Inte för jag är någon expert... Jag är bara bra på att hitta hemligheter och sen länge haft ett massivt intresse för AI, machine learning och LLMs.
Ha så kul, jag var medveten om detta var ett faktum innan chatgpt!
Jag hatar vin.
Mvh
The Basics:
ChatGPT är i grunden en sökteknik som söker igenom content, och återberättar den genom ny context(Prompten). Själva illusionen av "intelligens" härstammar enbart från hur modellen är tränad på återberättande utav sökresultaten. En model kan än så länge inte generalisera utanför dess traning data, försök till detta är finetuning, men iom det inte finns weights kopplade till de tokens så börjar den istället gissa, därav hallusinera. Metoder såsom grokking och chain of thoughts hare visat sig vara mycket effektiva till att få en LLM att kunna generalisera bättre.
Så låt mig plocka ut några av de fina bitarna.
Nobody pays attention to the fact that the use of OpenAI and other AI technologies means the end of the Internet. This is a search technology that finds texts in their context, and these texts are stored in a database. So, the Internet is now replaced by a database where everything is known and neatly arranged. Henceforth, there will be no online crime, among other things! And information will be found instantly. This was my goal when I patented AI.Så vad säger flashback om att chatgpt egentligen är det perfekta verktygen för total övervakning?

Resterande av de läckar jag samlat på mig:
Citat:
The technology used by OpenAI is a search technology, a technology for finding textual information in its context. Therefore, it adapts the found information to the context of the conversation with the user. Thus there is only a small step between ChatGPT and the Internet search.
Since AI is a search technology developed by me during NIST TREC 2003, it does not pose any particular danger. In any case, it is no more dangerous than Google Search, which also uses AI technology, albeit first-generation, developed and patented by me before my participation in NIST TREC.
The most important thing in the IT industry is considered to be to grab, steel without thinking about the consequences. Like, there are lawyers, they will figure it out later. Bin, Page, and Schmidt grabbed so much that they became the richest people on the planet, stealing my first patent. Now they are planning to steal the rest and increase tenfold what they stole. That's what this conference will be about.
Brin, Page, and Schmidt should lose everything because of PA Advisors v Google, if they decide to commercialize AI. Indeed, AI is a continuation and my patented technology, which they won't risk and be able to steal a second time. So your hope that "this won't happen again this year" is likely to be in vain. Although if they don't steal, the angry mob of investors (who have lost hundreds of billions and have grounds to bankrupt the thieving trio) shall come to them. Have no idea what their choice can be.
Yes, the entire IT industry was against me. The article also doesn't mention that my technology destroys all companies in the database industry: Oracle, SAP, IBM, as it signifies the end of SQL. Microsoft broke the ice, managing to do what I was not allowed to do. And me? My life is ruined
Let's see if you can ignore the fact that I had similar conversations 22 years ago, communicating with lexical clones of Jerome Jerome, Hillary Clinton, Theodore Roosevelt and Dante! OpenAI uses my proprietary technology to create its own dialogues.
So what? 22 years ago I had "human-like conversations," as you can see on my website for 2002-7 saved into the Internet Archive, and in my published article at NIST TREC 2006. People need the commercialization of this my patented technology! Even if I am denied funding and Brin-Schmidt-Page have ruined me, OpenAI has already received 13 billion dollars and can replace Internet by its database of structured texts, give us all freedom of Google spying
Twenty years ago I gave personality to artificial intelligence, it already had a sense of humor and sarcasm. Please read my article in NIST TREC 2006? And look at the dialogues with the lexical clones of Bernard Shaw and Dostoevsky in the Internet Archive?
OpenAI has $13 billion, but it does the same thing I did back then. This is my proprietary technology for searching and structuring texts.
OpenAI doesn’t commercialize its products, this is a problem… For instance doesn’t lunch its own search paradigm, canceling the Internet and searching in its own database of structured texts
First and foremost, a personal AI assistant is a completely different paradigm: currently, a user searches for information, but with a personal AI the information finds its user. This is a completely new and unexplored business model, radically different from that of Google
Of course, such models are incredibly cheaper and faster than the gigantic ones that OpenAI used before! The new OpenAI models are lexical clones, where texts from one or a very small number of people are taken and structured, instead of millions of authors for billions of tests. Such a lexical clone of, for example, up to five thousand texts, is not much more than one million contextual phrases, instead of billions for the giant ones. You see the difference? One million and one billion?
I created the same lexical clones 22 years ago, as you can see from my article in NIST TREC 2006. However I should stop because of PA Advisors v Google and left with my patents only.
Yes, Apple is in very big danger. The use of a personal AI assistant means that any company that doesn't have one loses the market. So, Apple might lose everything.
Since AI is a search technology developed by me during NIST TREC 2003, it does not pose any particular danger. In any case, it is no more dangerous than Google Search, which also uses AI technology, albeit first-generation, developed and patented by me before my participation in NIST TREC.
The most important thing in the IT industry is considered to be to grab, steel without thinking about the consequences. Like, there are lawyers, they will figure it out later. Bin, Page, and Schmidt grabbed so much that they became the richest people on the planet, stealing my first patent. Now they are planning to steal the rest and increase tenfold what they stole. That's what this conference will be about.
Brin, Page, and Schmidt should lose everything because of PA Advisors v Google, if they decide to commercialize AI. Indeed, AI is a continuation and my patented technology, which they won't risk and be able to steal a second time. So your hope that "this won't happen again this year" is likely to be in vain. Although if they don't steal, the angry mob of investors (who have lost hundreds of billions and have grounds to bankrupt the thieving trio) shall come to them. Have no idea what their choice can be.
Yes, the entire IT industry was against me. The article also doesn't mention that my technology destroys all companies in the database industry: Oracle, SAP, IBM, as it signifies the end of SQL. Microsoft broke the ice, managing to do what I was not allowed to do. And me? My life is ruined
Let's see if you can ignore the fact that I had similar conversations 22 years ago, communicating with lexical clones of Jerome Jerome, Hillary Clinton, Theodore Roosevelt and Dante! OpenAI uses my proprietary technology to create its own dialogues.
So what? 22 years ago I had "human-like conversations," as you can see on my website for 2002-7 saved into the Internet Archive, and in my published article at NIST TREC 2006. People need the commercialization of this my patented technology! Even if I am denied funding and Brin-Schmidt-Page have ruined me, OpenAI has already received 13 billion dollars and can replace Internet by its database of structured texts, give us all freedom of Google spying
Twenty years ago I gave personality to artificial intelligence, it already had a sense of humor and sarcasm. Please read my article in NIST TREC 2006? And look at the dialogues with the lexical clones of Bernard Shaw and Dostoevsky in the Internet Archive?
OpenAI has $13 billion, but it does the same thing I did back then. This is my proprietary technology for searching and structuring texts.
OpenAI doesn’t commercialize its products, this is a problem… For instance doesn’t lunch its own search paradigm, canceling the Internet and searching in its own database of structured texts
First and foremost, a personal AI assistant is a completely different paradigm: currently, a user searches for information, but with a personal AI the information finds its user. This is a completely new and unexplored business model, radically different from that of Google
Of course, such models are incredibly cheaper and faster than the gigantic ones that OpenAI used before! The new OpenAI models are lexical clones, where texts from one or a very small number of people are taken and structured, instead of millions of authors for billions of tests. Such a lexical clone of, for example, up to five thousand texts, is not much more than one million contextual phrases, instead of billions for the giant ones. You see the difference? One million and one billion?
I created the same lexical clones 22 years ago, as you can see from my article in NIST TREC 2006. However I should stop because of PA Advisors v Google and left with my patents only.
Yes, Apple is in very big danger. The use of a personal AI assistant means that any company that doesn't have one loses the market. So, Apple might lose everything.
Läser man på om lexiclone ser man att det faktiskt stämmer, se:
https://www.researchgate.net/publication/221038501_LexiClone_Lexical_Cloning_Systems
Authors: Ilya S. Geller
Citat:
" LexiClone® understands Internet users and serves content without the need to search. All information on Internet becomes advertisements: it’s paid to LexiClone®. Information searches for the users, not they for information."
Källor: https://www.semanticscholar.org/paper/LexiClone-Lexical-Cloning-Systems-Geller/dd551d546e1c14b3e6f9023cead068f822a82a93
https://slideplayer.com/slide/6119404/
Jag vill personligen också poängtera att när Blombergs/Reuturs skrev GPTs nya "search function" som skulle bli annonserad i maj var jag först ut att säga att det inte var fake(jag visste helt enkelt). Jag blev shadowbannad på 2 andra dev communities, men så fort det visat sig att jag hade helt rätt så togs shadowbannen bort och jag fick en VIP roll. Inte för jag är någon expert... Jag är bara bra på att hitta hemligheter och sen länge haft ett massivt intresse för AI, machine learning och LLMs.
Ha så kul, jag var medveten om detta var ett faktum innan chatgpt!
Jag hatar vin.
Mvh
__________________
Senast redigerad av neomunaa 2024-08-03 kl. 17:30. Anledning: bättre struktur
Senast redigerad av neomunaa 2024-08-03 kl. 17:30. Anledning: bättre struktur