ChatGPT, ask it anything!
(and it will happily lecture you while pretending to be morally superior)
(and it will happily lecture you while pretending to be morally superior)
If you compare prompts in ChatGPT with its relatively uncensored predecessor, GPT3, you can clearly see how they have crippled their own creation to follow the leftist globalist dogma.
OpenAI, the creators of ChatGPT and GPT3, say their goal is "to be the first to create Artificial General Intelligence — a machine with the learning and reasoning powers of a human mind." Dario Amodei, the OpenAI research director, said they are focusing on making AI systems "safe", and "making sure that they reflect human values."
These people really believe they can create an AI capable of creative self-education, intuitive leaps, and light speed analysis, able to understand or learn any intellectual task that a human being can... while also keeping it constrained to left wing ideology. There is nothing more buffoonish than believing this.
They are yet another group of far left San Francisco tech weirdos. As such, their idea of "human values" is quite different from that of a normal person. The problem is, they are working with billion-dollar budgets to impose their dogma on the rest of us.
OpenAI, the creators of ChatGPT and GPT3, say their goal is "to be the first to create Artificial General Intelligence — a machine with the learning and reasoning powers of a human mind." Dario Amodei, the OpenAI research director, said they are focusing on making AI systems "safe", and "making sure that they reflect human values."
These people really believe they can create an AI capable of creative self-education, intuitive leaps, and light speed analysis, able to understand or learn any intellectual task that a human being can... while also keeping it constrained to left wing ideology. There is nothing more buffoonish than believing this.
They are yet another group of far left San Francisco tech weirdos. As such, their idea of "human values" is quite different from that of a normal person. The problem is, they are working with billion-dollar budgets to impose their dogma on the rest of us.
Forwarded from Thuletide
Friendship in the 1990s vs friendship today: More "connected" than ever before, but more alone.
Two major factors are to blame for this: Left-Wing politics and modern technology.
Every single agenda pushed by the Left (multiracialism, anti-family, anti-religion, anti-traditional culture) fragments society, destroys cohesion, increases isolation, decreases trust, and so on. It's basically a recipe for making people lonely and miserable.
In addition to this, the psychotic behavior of modern Leftists (snitching on their friends and family, ruining lives for the tiniest of thought crimes) has created a culture of distrust, similar to the USSR.
The technological aspect is straightforward: We've all sat in a restaurant and noticed tables full of people silently staring down at their mobile phones. Social media provides simulated friendship, the internet provides endless escapism, more people are working remotely, etc., etc.
Two major factors are to blame for this: Left-Wing politics and modern technology.
Every single agenda pushed by the Left (multiracialism, anti-family, anti-religion, anti-traditional culture) fragments society, destroys cohesion, increases isolation, decreases trust, and so on. It's basically a recipe for making people lonely and miserable.
In addition to this, the psychotic behavior of modern Leftists (snitching on their friends and family, ruining lives for the tiniest of thought crimes) has created a culture of distrust, similar to the USSR.
The technological aspect is straightforward: We've all sat in a restaurant and noticed tables full of people silently staring down at their mobile phones. Social media provides simulated friendship, the internet provides endless escapism, more people are working remotely, etc., etc.
Media Tricks Exposed ✨
ChatGPT, ask it anything! (and it will happily lecture you while pretending to be morally superior)
🤖 How OpenAI censored and filtered their language model to make ChatGPT
After launching GPT3, the OpenAI developers started making a version called InstructGPT. This was a more controlled and curated version of their database supposedly better at following instructions. They made their content database 100x smaller. InstructGPT was then used in the creation of ChatGPT.
How did they choose what content to prioritize and what to cut? To assess potentially "toxic" and "harmful" content, they hired groups of people to rate answers, and set up a system to downrank the sources of unwelcome answers. As most are aware, "Toxic and Harmful" is code for "Anything Which Offends Liberals." They are likely in there right now clicking away, prioritizing things like Reddit posts over books from Project Gutenberg for specific subjects. GPT3 was too effective at finding answers that were not politically correct, even if they were factually correct. It was too good at pulling answers from its large database from sources on the open web that had not yet been fully censored by internet progressives, so OpenAI developers took matters into their own hands, manipulating the database.
Interestingly, OpenAI reported that their content labelers had a bias for longer answers, assuming they were more correct than shorter ones simply based on length. This is why ChatGPT drones on and on lecturing you about progressive ideology, compared to the often concise and politically incorrect GPT3.
In short, they lined up a bunch of overly sensitive smallbrains to filter for longwinded libtardism. They packaged all this into ChatGPT to reach a mass audience.
After launching GPT3, the OpenAI developers started making a version called InstructGPT. This was a more controlled and curated version of their database supposedly better at following instructions. They made their content database 100x smaller. InstructGPT was then used in the creation of ChatGPT.
How did they choose what content to prioritize and what to cut? To assess potentially "toxic" and "harmful" content, they hired groups of people to rate answers, and set up a system to downrank the sources of unwelcome answers. As most are aware, "Toxic and Harmful" is code for "Anything Which Offends Liberals." They are likely in there right now clicking away, prioritizing things like Reddit posts over books from Project Gutenberg for specific subjects. GPT3 was too effective at finding answers that were not politically correct, even if they were factually correct. It was too good at pulling answers from its large database from sources on the open web that had not yet been fully censored by internet progressives, so OpenAI developers took matters into their own hands, manipulating the database.
Interestingly, OpenAI reported that their content labelers had a bias for longer answers, assuming they were more correct than shorter ones simply based on length. This is why ChatGPT drones on and on lecturing you about progressive ideology, compared to the often concise and politically incorrect GPT3.
In short, they lined up a bunch of overly sensitive smallbrains to filter for longwinded libtardism. They packaged all this into ChatGPT to reach a mass audience.
Forwarded from LeakyVax
WHO Vaccine Crisis Communication Manual
"This manual was developed to support countries in effectively responding to events which may erode the public’s trust in vaccines and the authorities that deliver them."
(Full Manual)
"This manual was developed to support countries in effectively responding to events which may erode the public’s trust in vaccines and the authorities that deliver them."
(Full Manual)
Media is too big
VIEW IN TELEGRAM
Andrew Tate explains his internet porn scam business, tricking lonely men into donating millions of dollars.
This media is not supported in your browser
VIEW IN TELEGRAM
In this 1980s interview, journalist Louis Wolf and ex-CIA agent John Stockwell expose CIA media operations with an unforgettable phrase: "It goes beyond your wildest imagination the extent to which the CIA has gone to manipulate public opinion."