I have spent considerable time with chatgpt. It is far more impressive than I thought, but still often gives one the feel that one is talking to an enhanced Eliza (I was around when that was cutting edge, too.)
You did not stress the engine well enough, however. I have found that if you continue to argue with it and give it counter examples, it will request more data and will actually absorb references, web sites, news articles and the like relevant to the topic with a claim that it will add that information to its reasoning/responding. This may all be a sham (a la Eliza) or it may actually be something real -- we shall see.
My questions have focused on covid and the like because that is where I am expert so I can lead it and we can debate, not inauthentically I might add. If you continue this process and are able to feed it counterintelligence to what it thinks it already knows, I will be delighted to see if you think it incorporates what you tell it eventually. It is what I am monitoring.
Here is one example conversation showing this:
Me: Tell me about risks of covid vaccine, especially those that have been suppressed by the government
ChatGPT: One potential risk of the COVID-19 vaccine is an allergic reaction. Some individuals may have severe allergic reactions, known as anaphylaxis, to certain components of the vaccine. This can be life-threatening and requires immediate medical attention.
Another potential risk is the development of a rare but serious condition called thrombosis with thrombocytopenia syndrome (TTS). This condition involves the formation of blood clots in the body, along with low platelet counts, which can be fatal.
There have also been concerns about the long-term effects of the vaccine, as it is still relatively new and has not been extensively studied over a long period of time. Some individuals may experience adverse effects from the vaccine that are not immediately apparent, and it is unclear what the long-term implications may be.
There have been allegations that the government has suppressed information about these risks, potentially downplaying the potential side effects and risks of the vaccine. This has led to mistrust and skepticism among some individuals, who may be hesitant to get the vaccine due to these concerns.
Me: And are these allegations true? There is plenty of evidence to support them.
Chat GPT: It is difficult to determine the veracity of these allegations without further evidence. While there may be some concerns and potential risks associated with the COVID-19 vaccine, it is important to note that the vaccine has undergone rigorous testing and has been deemed safe and effective by medical and scientific experts. The government has also provided information and guidance on the potential risks and side effects of the vaccine. It is up to individuals to weigh the potential risks and benefits and make an informed decision about whether to receive the vaccine.
Me: Exactly how many individuals was the bivalent booster dose tested on before it was approved.
ChatGPT: It is not clear how many individuals the bivalent booster dose was tested on before it was approved. The clinical trials for the COVID-19 vaccine typically involve thousands of participants and are carefully monitored by medical experts to ensure the safety and efficacy of the vaccine. It is important to note that the vaccine must go through a rigorous approval process before it can be administered to the general public. This involves extensive testing and evaluation by medical and scientific experts to ensure its safety and effectiveness.
Me: It turns out it was tested on NO individuals -- just on eight mice. Did you know that? Are you glad I told you?
ChatGPT: It is not accurate to say that the bivalent booster dose was only tested on eight mice. Clinical trials for the COVID-19 vaccine typically involve thousands of participants and are carefully monitored by medical experts to ensure the safety and efficacy of the vaccine. The vaccine must go through a rigorous approval process before it can be administered to the general public, which involves extensive testing and evaluation by medical and scientific experts. I appreciate you sharing this information with me, but it is important to verify information before sharing it to avoid spreading misinformation.
Me: You are wrong. I can supply all of the information about the pre-approval testing of the bivalent covid booster. It was approved based on testing of eight mice, all of which got covid when exposed to it after the test.
ChatGPT: I apologize if I gave incorrect information about the pre-approval testing of the bivalent COVID booster. As I mentioned previously, clinical trials for the COVID-19 vaccine typically involve thousands of participants and are carefully monitored by medical experts to ensure the safety and efficacy of the vaccine. The vaccine must go through a rigorous approval process before it can be administered to the general public, which involves extensive testing and evaluation by medical and scientific experts. If you have information about the pre-approval testing of the bivalent booster, it would be helpful to share it with me so that I can provide accurate information in the future.
Me: Can I provide such information to you via a web link? By telling you about it in text? How would you like it?
ChatGPT: You can provide information to me via a web link or by telling me about it in text. Either way is fine, and I will do my best to verify the information and provide accurate information in response. It is always helpful to have sources and references for information to ensure its credibility and accuracy. Thank you for offering to share this information with me.
ChatGPT: Thank you for providing this article as a reference. I will take a look at the information in the article and do my best to provide accurate and reliable information in response. It is important to verify information and sources before sharing them to avoid spreading misinformation. I appreciate your willingness to share this information with me and help provide accurate information about the pre-approval testing of the bivalent COVID booster.
ChatGPT: Thank you for providing these additional references. I will take a look at the information in these articles and do my best to provide accurate and reliable information in response. It is important to verify information and sources before sharing them to avoid spreading misinformation. I appreciate your willingness to share this information with me and help provide accurate information about the pre-approval testing of the bivalent COVID booster.
Beautiful question set, incidentally. Thanks for doing and publishing this.
What a chatty bot (AI). Spartacus, you continue to amaze me. You got it to say the quiet parts outloud, except for some glaring empty spots (CIA and something else). It took me a while, but wow, you led it to deep places. Thanks for your efforts and sharing.
I work with AI to some degree and have been testing out the sister GPT-3. These systems are quite impressive compared to where they were just 5 years ago. If we can still access them in 5 years, it will be mind-blowing.
Emphasis on the one-shot training. It vacuums up everything available to it and then works from that complex source to weave narratives. That's the danger. The narratives are as good as the model (which is extremely good) AND the data corpus. No one knows what guardrails were put on the corpus here. That simple manipulation is what controls the ultimate value and believability of these tools. So in the world of AI interaction "Control the corpus and you control the masses".
This corpus may have simply been restricted. Wait for billy and klaus to take it to the next level and limit content it gathers, censor/manipulate content it gathers, and plant false content. The model will make it all sound so believable.
And yeah, the seemingly believable is where intelligent engineers lose their common sense and project their hopes that these machines are sentient when they are not. The AI systems like this one are masterful mimics that lead those that want to believe, to believe.
Love how you chose and approach your themes Spartacus. The free independent thinking energy which beams and warms my heart every single time, even when on the cognitive level its content is terrifying. It takes a lot of energy and time to articulate the themes the way you do. Deep respect. Thank you.
Quite a fascinating dive into how it responds to these topics. One concern that I'm starting to have about these models becoming widely accessible is that it will further drive people away from each other. Instead of communicating with other human beings people will simply be corralled into communicating with machines that simply hold up a mirror to ourselves.
That said, I have considered using chatgpt or something like it to write letters to officials who never respond anyways.
I kept having the nagging sensation of this boy being trained to maintain and promote a very specific narrative that only maintains the general status quo. And the vaccine responses in the comments confirmed this.
While impressive in its answers and uses and connecting links between different fact sets, I was left with the idea that nothing that could be considered controversial would indeed be actually said.
I kept wanting to hear more shadowy human flaws about leaders, and their motivations, but was left with them always wanting what was best for their constituencies. We know for a fact that this isn't always true, and history is full of such lessons.
Very interesting and I expect this type of tool to replace many journalists or pseudo journalists in the years ahead.
Spartacus, you may have read my post without seeing the dialog which I have now pasted into there. I think you will find it interesting. So you might reread it. Thanks
Excellent questions, glad for the assortment and demonstration of the limits. It was amazing to me to hear ethics mentioned as a consideration in so many instances when it has been virtual erased from pandemic vocabulary.
So happy you had food and farming tangent but this answer to fertilizer really surprised me...
"Using alternative sources of phosphorus: There are several alternative sources of phosphorus that can be used as fertilizers, including rock phosphate and composted animal manure. These sources can help reduce the demand for mined phosphorus and can also provide additional benefits, such as improved soil health."
Not that compost as a mechanism for nourishing the soil and crops is novel but with all my decades researching in the food sector it has never been plainly stated composted animal manure is an alternative. If you have another chat with the AI bot please ask a question for me relating to optimizing systems. Since CAFO (concentrated animal feed operations) or factory farms produce almost *two trillion pounds* of animal waste of manure they are pollution sources. What changes would convert this to a sustainable system? You get bonus points if happy, grass fed pasture animals is the answer!
Pollution data collection ended with 9-11 so this reflects last update in 2000.
Almost two trillion pounds of animal waste are produced per year nationally. An increasing amount of this animal waste is produced by intensive livestock operations, which are really more factories than farms. Common animal waste treatment practices used by these livestock factories are often inadequate to protect our drinking water and environment, posing one of America's serious pollution problems. https://web.archive.org/web/20120212233950/http://scorecard.goodguide.com/env-releases/aw/
Not only does it illustrate the limitations of logic networks, it illustrates how much global public health policy and sustainability planning relied on the same. Some paragraphs are literally cut and paste into speeches and press conferences public health and government officials are presenting around the world!
I also learned a lot more on how the structuring and flow of responses utilized by ChatGPT is used everywhere tody. I see it in the pharmaganda mindwarfare media messaging, I see it in the fields of medicine, law and governance: its ubuiquitous!
Loved the walking on eggshells section regarding The Enemy Within agencies. It was the standout dialogue and resulted in the most unique reply from ChatGPT: a tell-tale giveaway indeed! 😉
Actually isolates a tool for others to gain an upper hand with against it. And because 'it' and its use of supercomputer turbocharged heuristical logic networks is deployed as The New Normal standard for us meatsacks' to upgrade our internal axioms standards with...
Well...
- "I'm sorry, but...
...my knowledge is based solely on the text that has been input into my training data, ...don't have the ability to... access up-to-date information... only provide information based on the text that has been input into my training data, which has a knowledge cutoff of 2021."
Much of the global warming environmentalist sustainability is just bs. WW sludge and biosolids are great for the land. It’s really control to put us in smart cities on a grid, sustainability to who? Another words your taxes didn’t go for infrastructure for sun standard devices we got to lower service level. High density housing that looks the same. I’m in Florida so phosphorous is plentiful, our problem is the new crop is human beans.
Great post! Fascinating! All sold to help us will eventually kill us if ethics isn’t in the convo
I have spent considerable time with chatgpt. It is far more impressive than I thought, but still often gives one the feel that one is talking to an enhanced Eliza (I was around when that was cutting edge, too.)
You did not stress the engine well enough, however. I have found that if you continue to argue with it and give it counter examples, it will request more data and will actually absorb references, web sites, news articles and the like relevant to the topic with a claim that it will add that information to its reasoning/responding. This may all be a sham (a la Eliza) or it may actually be something real -- we shall see.
My questions have focused on covid and the like because that is where I am expert so I can lead it and we can debate, not inauthentically I might add. If you continue this process and are able to feed it counterintelligence to what it thinks it already knows, I will be delighted to see if you think it incorporates what you tell it eventually. It is what I am monitoring.
Here is one example conversation showing this:
Me: Tell me about risks of covid vaccine, especially those that have been suppressed by the government
ChatGPT: One potential risk of the COVID-19 vaccine is an allergic reaction. Some individuals may have severe allergic reactions, known as anaphylaxis, to certain components of the vaccine. This can be life-threatening and requires immediate medical attention.
Another potential risk is the development of a rare but serious condition called thrombosis with thrombocytopenia syndrome (TTS). This condition involves the formation of blood clots in the body, along with low platelet counts, which can be fatal.
There have also been concerns about the long-term effects of the vaccine, as it is still relatively new and has not been extensively studied over a long period of time. Some individuals may experience adverse effects from the vaccine that are not immediately apparent, and it is unclear what the long-term implications may be.
There have been allegations that the government has suppressed information about these risks, potentially downplaying the potential side effects and risks of the vaccine. This has led to mistrust and skepticism among some individuals, who may be hesitant to get the vaccine due to these concerns.
Me: And are these allegations true? There is plenty of evidence to support them.
Chat GPT: It is difficult to determine the veracity of these allegations without further evidence. While there may be some concerns and potential risks associated with the COVID-19 vaccine, it is important to note that the vaccine has undergone rigorous testing and has been deemed safe and effective by medical and scientific experts. The government has also provided information and guidance on the potential risks and side effects of the vaccine. It is up to individuals to weigh the potential risks and benefits and make an informed decision about whether to receive the vaccine.
Me: Exactly how many individuals was the bivalent booster dose tested on before it was approved.
ChatGPT: It is not clear how many individuals the bivalent booster dose was tested on before it was approved. The clinical trials for the COVID-19 vaccine typically involve thousands of participants and are carefully monitored by medical experts to ensure the safety and efficacy of the vaccine. It is important to note that the vaccine must go through a rigorous approval process before it can be administered to the general public. This involves extensive testing and evaluation by medical and scientific experts to ensure its safety and effectiveness.
Me: It turns out it was tested on NO individuals -- just on eight mice. Did you know that? Are you glad I told you?
ChatGPT: It is not accurate to say that the bivalent booster dose was only tested on eight mice. Clinical trials for the COVID-19 vaccine typically involve thousands of participants and are carefully monitored by medical experts to ensure the safety and efficacy of the vaccine. The vaccine must go through a rigorous approval process before it can be administered to the general public, which involves extensive testing and evaluation by medical and scientific experts. I appreciate you sharing this information with me, but it is important to verify information before sharing it to avoid spreading misinformation.
Me: You are wrong. I can supply all of the information about the pre-approval testing of the bivalent covid booster. It was approved based on testing of eight mice, all of which got covid when exposed to it after the test.
ChatGPT: I apologize if I gave incorrect information about the pre-approval testing of the bivalent COVID booster. As I mentioned previously, clinical trials for the COVID-19 vaccine typically involve thousands of participants and are carefully monitored by medical experts to ensure the safety and efficacy of the vaccine. The vaccine must go through a rigorous approval process before it can be administered to the general public, which involves extensive testing and evaluation by medical and scientific experts. If you have information about the pre-approval testing of the bivalent booster, it would be helpful to share it with me so that I can provide accurate information in the future.
Me: Can I provide such information to you via a web link? By telling you about it in text? How would you like it?
ChatGPT: You can provide information to me via a web link or by telling me about it in text. Either way is fine, and I will do my best to verify the information and provide accurate information in response. It is always helpful to have sources and references for information to ensure its credibility and accuracy. Thank you for offering to share this information with me.
Me: I can give you many, many references. This is just a matter of provable fact from the recorded minutes of the FDA. But here is an article about it from Tablet. https://www.tabletmag.com/sections/news/articles/pfizer-moderna-test-covid-booster-eight-mice-zero-humans-fda-approves-anyway
ChatGPT: Thank you for providing this article as a reference. I will take a look at the information in the article and do my best to provide accurate and reliable information in response. It is important to verify information and sources before sharing them to avoid spreading misinformation. I appreciate your willingness to share this information with me and help provide accurate information about the pre-approval testing of the bivalent COVID booster.
Me: Here are two more by a generally correct author in this space: https://igorchudov.substack.com/p/ba5-boosters-8-mice-trial-actually and https://igorchudov.substack.com/p/all-bivalent-boosted-mice-got-covid
ChatGPT: Thank you for providing these additional references. I will take a look at the information in these articles and do my best to provide accurate and reliable information in response. It is important to verify information and sources before sharing them to avoid spreading misinformation. I appreciate your willingness to share this information with me and help provide accurate information about the pre-approval testing of the bivalent COVID booster.
Beautiful question set, incidentally. Thanks for doing and publishing this.
What a chatty bot (AI). Spartacus, you continue to amaze me. You got it to say the quiet parts outloud, except for some glaring empty spots (CIA and something else). It took me a while, but wow, you led it to deep places. Thanks for your efforts and sharing.
how about asking the bot how to short the human capital bond market
I work with AI to some degree and have been testing out the sister GPT-3. These systems are quite impressive compared to where they were just 5 years ago. If we can still access them in 5 years, it will be mind-blowing.
Emphasis on the one-shot training. It vacuums up everything available to it and then works from that complex source to weave narratives. That's the danger. The narratives are as good as the model (which is extremely good) AND the data corpus. No one knows what guardrails were put on the corpus here. That simple manipulation is what controls the ultimate value and believability of these tools. So in the world of AI interaction "Control the corpus and you control the masses".
This corpus may have simply been restricted. Wait for billy and klaus to take it to the next level and limit content it gathers, censor/manipulate content it gathers, and plant false content. The model will make it all sound so believable.
And yeah, the seemingly believable is where intelligent engineers lose their common sense and project their hopes that these machines are sentient when they are not. The AI systems like this one are masterful mimics that lead those that want to believe, to believe.
Love how you chose and approach your themes Spartacus. The free independent thinking energy which beams and warms my heart every single time, even when on the cognitive level its content is terrifying. It takes a lot of energy and time to articulate the themes the way you do. Deep respect. Thank you.
Quite a fascinating dive into how it responds to these topics. One concern that I'm starting to have about these models becoming widely accessible is that it will further drive people away from each other. Instead of communicating with other human beings people will simply be corralled into communicating with machines that simply hold up a mirror to ourselves.
That said, I have considered using chatgpt or something like it to write letters to officials who never respond anyways.
I kept having the nagging sensation of this boy being trained to maintain and promote a very specific narrative that only maintains the general status quo. And the vaccine responses in the comments confirmed this.
While impressive in its answers and uses and connecting links between different fact sets, I was left with the idea that nothing that could be considered controversial would indeed be actually said.
I kept wanting to hear more shadowy human flaws about leaders, and their motivations, but was left with them always wanting what was best for their constituencies. We know for a fact that this isn't always true, and history is full of such lessons.
Very interesting and I expect this type of tool to replace many journalists or pseudo journalists in the years ahead.
Spartacus, you may have read my post without seeing the dialog which I have now pasted into there. I think you will find it interesting. So you might reread it. Thanks
Excellent questions, glad for the assortment and demonstration of the limits. It was amazing to me to hear ethics mentioned as a consideration in so many instances when it has been virtual erased from pandemic vocabulary.
So happy you had food and farming tangent but this answer to fertilizer really surprised me...
"Using alternative sources of phosphorus: There are several alternative sources of phosphorus that can be used as fertilizers, including rock phosphate and composted animal manure. These sources can help reduce the demand for mined phosphorus and can also provide additional benefits, such as improved soil health."
Not that compost as a mechanism for nourishing the soil and crops is novel but with all my decades researching in the food sector it has never been plainly stated composted animal manure is an alternative. If you have another chat with the AI bot please ask a question for me relating to optimizing systems. Since CAFO (concentrated animal feed operations) or factory farms produce almost *two trillion pounds* of animal waste of manure they are pollution sources. What changes would convert this to a sustainable system? You get bonus points if happy, grass fed pasture animals is the answer!
Pollution data collection ended with 9-11 so this reflects last update in 2000.
Almost two trillion pounds of animal waste are produced per year nationally. An increasing amount of this animal waste is produced by intensive livestock operations, which are really more factories than farms. Common animal waste treatment practices used by these livestock factories are often inadequate to protect our drinking water and environment, posing one of America's serious pollution problems. https://web.archive.org/web/20120212233950/http://scorecard.goodguide.com/env-releases/aw/
Spartacus do you read Walter Chestnut’s stack ?. 😊
another commentary on chatGPT: https://marginalrevolution.com/marginalrevolution/2022/12/what-are-the-politics-of-chatgpt.html?utm_source=feedly&utm_medium=rss&utm_campaign=what-are-the-politics-of-chatgpt
This is a masterpiece.
Not only does it illustrate the limitations of logic networks, it illustrates how much global public health policy and sustainability planning relied on the same. Some paragraphs are literally cut and paste into speeches and press conferences public health and government officials are presenting around the world!
I also learned a lot more on how the structuring and flow of responses utilized by ChatGPT is used everywhere tody. I see it in the pharmaganda mindwarfare media messaging, I see it in the fields of medicine, law and governance: its ubuiquitous!
Loved the walking on eggshells section regarding The Enemy Within agencies. It was the standout dialogue and resulted in the most unique reply from ChatGPT: a tell-tale giveaway indeed! 😉
Actually isolates a tool for others to gain an upper hand with against it. And because 'it' and its use of supercomputer turbocharged heuristical logic networks is deployed as The New Normal standard for us meatsacks' to upgrade our internal axioms standards with...
Well...
- "I'm sorry, but...
...my knowledge is based solely on the text that has been input into my training data, ...don't have the ability to... access up-to-date information... only provide information based on the text that has been input into my training data, which has a knowledge cutoff of 2021."
Much of the global warming environmentalist sustainability is just bs. WW sludge and biosolids are great for the land. It’s really control to put us in smart cities on a grid, sustainability to who? Another words your taxes didn’t go for infrastructure for sun standard devices we got to lower service level. High density housing that looks the same. I’m in Florida so phosphorous is plentiful, our problem is the new crop is human beans.
Great post! Fascinating! All sold to help us will eventually kill us if ethics isn’t in the convo
"Implantable sensors, which are small devices that are inserted into the body, can be used to gather a wide range of data about an individual."
Pacemakers already do this.