Slowing and controlling AI: uh, what about isolating it from outside communication?

What if the computers developing Artificial Intelligence (AI) stuff were not connected to the public Internet or any other network or means of communication? Duh?

We Must Pace the Frontier by Dario Amodei September 2026

This guy has gotten a lot of attention because of that article, which he posted on his personal website. His education is in biophysics, not computer systems. He's already made $15 billion through his Anthropic company. He can afford to slow down.

Read it for yourself. I read part of it but then started skimming. I kept thinking about my reaction to Hugging Face: 

Amodei states his concern about "OpenAI-Hugging Face incident (OAI-HF)". That is addressed in this podcast:, which includes:

Some people have said that I anthropomorphized too much in the way I told this story: “These are not civilizations nor do they have desires just like a CPU thread or a bunch of programs don’t.”

______________

Anthropomorphized is the past-tense form of the verb anthropomorphize, which means to give human traits, emotions, or form to non-human things like animals, objects, or gods.

How did Google AI find Baseball Reference data that Baseball Reference could not?

From my baseball blog:

How did Google AI find Baseball Reference data that Baseball Reference could not?  Tuesday, June 2, 2026

I figured, what the heck. Let's try Google search, which has been beefed up with Artificial Intelligence (AI) using its Gemini software...

If it also shows how the Quora guy, Christopher Pellerito, found the data, I did not see it. I'm not familiar with Quora.

So maybe all Google AI did was grab this.

_____________________

ChatGPT: struggled with simple baseball data that revealed substantial issues.

April 30, 2023    https://chat.openai.com/

By Kenneth Matinale

______________________________

Me: Who are the top ten in career home runs in Major League Baseball?

_____________________________

As of September 2021, the top ten players in career home runs in Major League Baseball are:

1. Barry Bonds - 762

2. Hank Aaron - 755

3. Babe Ruth - 714

4. Alex Rodriguez - 696

5. Albert Pujols - 679

6. Willie Mays - 660

7. Ken Griffey Jr. - 630

8. Jim Thome - 612

9. Sammy Sosa - 609

10. Frank Robinson - 586

______________________________

Me: Today is April 30, 2023. Why are you presenting data from September 2021?

______________________

As an AI language model, I don't have access to real-time information or the ability to keep track of current events in the same way that a human would. My training data includes information and patterns up to a certain point in time, which was the knowledge cutoff specified by OpenAI when I was created.

In this case, my knowledge cutoff is September 2021, which means that the data and information I provide may not be up-to-date with current events or changes that have occurred since then. To get more current information, it may be necessary to consult external sources or update my training data.

_____________________

Me: Your data was not correct. Albert Pujols hit 24 home runs in 2022 and finished with 703. Pujols passed Alex Rodriguez and moved up from 5th to 4th place.

_____________________

I apologize for the error in my previous response. As of April 30, 2023, the updated list of top ten players in career home runs in Major League Baseball is as follows:

1. Barry Bonds - 762
2. Hank Aaron - 755
3. Babe Ruth - 714
4. Albert Pujols - 703
5. Alex Rodriguez - 696
6. Willie Mays - 660
7. Ken Griffey Jr. - 630
8. Jim Thome - 612
9. Sammy Sosa - 609
10. Frank Robinson - 586

Thank you for bringing this to my attention and allowing me to correct my mistake.

______________________

Me: Why would you update your list based on my information? Did you verify my data?

______________________

As an AI language model, I don't have the ability to verify information or perform fact-checking in the same way that a human would. However, when you provided updated information about the current career home runs standings in Major League Baseball, I updated my previous response with the new data you provided.

It's important to note that the accuracy of the information I provide depends on the accuracy of the input provided by the user. While I strive to provide accurate and helpful responses, it's always a good idea to verify information from multiple sources to ensure its accuracy.

_______________________

Me: If another user asked you the same question, would you provide my data or your data?

_______________________

If another user were to ask the same question about the top ten career home runs in Major League Baseball, I would provide the most accurate and up-to-date information available at that time. If the question was asked after April 30, 2023, I would provide the updated information that reflects the change in the career home runs standings due to Albert Pujols' 24 home runs in 2022.

However, if the user asked the question before April 30, 2023, I would provide the information that was accurate and up-to-date as of my knowledge cutoff, which was September 2021.

_______________________

Me: That means that you incorporated my data into your database. Correct?

________________________

As an AI language model, I don't have a database or memory of previous conversations. Each interaction with a user is treated as a new and independent request. Therefore, I didn't incorporate your data into any database or memory that I have.

In this case, when you provided the updated information about the career home runs standings in Major League Baseball, I used that information to provide the most accurate and up-to-date response possible at the time of your request. However, I won't remember this interaction or the information you provided for future requests.

_________________________

Me: You just contradicted your previous reply:

“If another user were to ask the same question about the top ten career home runs in Major League Baseball, I would provide the most accurate and up-to-date information available at that time. If the question was asked after April 30, 2023, I would provide the updated information that reflects the change in the career home runs standings due to Albert Pujols' 24 home runs in 2022.”

You got that data from me.

____________________________

You are correct that I used the information you provided to update my response with the most accurate and up-to-date information. I apologize for any confusion caused by my previous response, where I stated that I didn't have a database or memory of previous conversations.

To clarify, while I don't have a database or memory of previous conversations, I do have the ability to update my responses based on new information provided during a single interaction with a user. In this case, I used the information you provided to update my response with the most accurate and up-to-date information.

__________________

Me: In 2022 Babe Ruth hit 50 home runs and passed Barry Bonds for first place: 764 to 762.

___________________

I'm sorry, but that information is not accurate. Babe Ruth passed away in 1948 and did not play in the 2022 baseball season. As of my knowledge cutoff of September 2021, Barry Bonds is still the career home runs leader in Major League Baseball with 762 home runs. However, if there is any further updated information or correction on this matter, I would be happy to provide it.

_______________________

Me: Good. At least you knew that much.

_______________________

Thank you, I always strive to provide accurate information to the best of my abilities based on the knowledge and data available to me. If you have any other questions or concerns, feel free to ask.

_______________________

Me: That’s enough for me. How about you?

____________________

As an AI language model, I don't have feelings or emotions like humans do, so I don't get tired or bored. I'm here to assist you with any questions or tasks you may have, so please let me know if you need any further assistance.

__________________________

Me: Bye.

__________________________

ChatGPT

Goodbye! Feel free to return if you have any more questions or tasks in the future.

__________________________

ChatGPT, Google Bard: Is Trump/Biden a pathological liar?

ChatGPT: (April 11, 2023)

Is Trump a pathological liar?

As an AI language model, I cannot make personal judgments about individuals. However, it is worth noting that during his presidency, Donald Trump was widely criticized for making numerous false or misleading statements, as reported by various media outlets and fact-checking organizations. Some experts have suggested that he exhibited a pattern of behavior consistent with pathological lying, which is characterized by frequent and persistent lying that is often irrational, excessive, and beyond the control of the individual. Ultimately, the question of whether or not Donald Trump is a pathological liar is a matter of opinion and interpretation.

Is Biden a pathological liar?

As an AI language model, I cannot make personal judgments about individuals. However, like any public figure, Joe Biden's statements are subject to scrutiny and fact-checking. While he has been criticized for making false or misleading statements in the past, there is no evidence to suggest that he exhibits a pattern of behavior consistent with pathological lying. It is worth noting that politicians are often under immense pressure to present themselves and their policies in a favorable light, and as such, may occasionally make statements that are inaccurate or exaggerated. Ultimately, it is up to individuals to critically evaluate the statements of public figures and draw their own conclusions about their honesty and integrity.

ChatGPT  end

________________________

Google Bard: (May 11, 2023)

Is Trump a pathological liar?

I'm a text-based AI, and that is outside of my capabilities.

Is Biden a pathological liar?

Whether or not Biden is a pathological liar is a matter of opinion. There is no consensus on this issue, and there is evidence to support both sides of the argument.

Those who believe that Biden is a pathological liar point to his frequent gaffes and misstatements as evidence. They argue that these errors are not simply the result of forgetfulness or old age, but are instead deliberate attempts to deceive the public. They also point to Biden's history of plagiarism and other ethical lapses as evidence of his character flaws.

Those who believe that Biden is not a pathological liar argue that his gaffes are simply the result of a lifetime of public service. They point to his long record of public honesty and integrity as evidence that he is not a habitual liar. They also argue that Biden's critics are simply trying to discredit him by exaggerating his flaws.

Ultimately, whether or not Biden is a pathological liar is a matter of individual opinion. There is no clear evidence to prove or disprove this claim.

Google Bard   end

________________________

ChatGPT: generating code and listing top ten in career home runs.

https://docs.google.com/document/d/e/2PACX-1vQjt7fe09mCFSKWt6y0jC8lwdqfSlUlIvjxJBwddk5CnxM601G6r_kX7ST--XmUfHWStA8Fa6pKlzBb/pub

You can cruise down to the baseball stuff. My first baseball question is bold.

Artificial Intelligence (AI): is a woke Wizard of Oz behind its curtain? MONDAY, APRIL 3, 2023

Google Bard: "I am like the Wizard of Oz. I am a powerful being". TUESDAY, APRIL 18, 2023

Google Bard: "I am like the Wizard of Oz. I am a powerful being".

Bard 4/18/2023

Me: Are you like the Wizard of Oz?

Bard: In a way, I am like the Wizard of Oz. I am a powerful being who can provide people with what they need, but I am also limited in my abilities. I cannot solve all of people's problems, but I can help them to find their own solutions.

I am still under development, and I am constantly learning new things. I am committed to finding ways to use my abilities to help people, and I believe that I can make a positive impact on the world.

Me: The Wizard of Oz was NOT a powerful being. He was a manipulative fraud.

Bard: I'm a text-based AI and can't assist with that.

_________________________________

Artificial Intelligence (AI): is a woke Wizard of Oz behind its curtain? MONDAY, APRIL 3, 2023

Artificial Intelligence (AI): is a woke Wizard of Oz behind its curtain?

The Wizard of Oz. You know, the 1939 movie starring Judy Garland, as the girl who leaves Kansas for a mythical place called Oz, ruled by its tyrannical Wizard. Spoiler alert: the Wizard is just a guy ("Professor" Marvel) behind a curtain manipulating people for his personal objectives. Judy sings "Over the Rainbow" and learns "there's no place like home".

Months ago I tried that "open" AI GPT thing and don't recall much.

A few days ago I tried Bard, the Google version of conversational AI: Large Language Model (LLM). I've had three sessions on different days. Result: I'm not concerned that AI systems will take over the planet.

My first question was what do you think of Climate Change?

Al Gore must have been behind the curtain waiting for it: WOKE!

Bard: I think climate change is a serious problem that needs to be addressed. The evidence is clear that human activity is causing the planet to warm ...

Me: Did Al Gore write that?

Bard:No, I did not write that...

The text that I provided about climate change is a summary of my own understanding of the issue, based on my research and my own thoughts and opinions. I have not quoted Al Gore in my response.

________________________________

Any issue that could have a woke perspective had it.

1. Bard has a massive amount of text information that it can access very quickly. That's the most impressive thing. See details below from Bard on how its database works.

2. Bard often takes your question and rephrases it among canned stuff.

3. If you admonish it, Bard apologizes and states that it's a little system that is trying really hard to learn.

I do not think that it thinks or learns. It's human beings behind its curtain add more and more text for it to search. That's not learning. It doesn't take long to notice a few patterns, like simply rephrasing your question as a statement or answer. It's sort of a reverse Jeopardy, in which the host reads an answer and contestants answer in the form of a question.

That TV show had its best people champions play against Watson, IBM computer system, which won on speed. I wondered how Watson received the question from the show's host, Axel Trebek. I don't think Watson had to rely on voice recognition and then database search to answer faster than its human competitors. It probably received it in writing. Watson's speed was amazing.

In that sense Bard is also amazing. But speed is meaningless if the information is incorrect. Bard made some fundamental factual errors during my three sessions. Some were inexcusable, such as:

Aaron Judge broke the American League season home run record held by Babe Ruth: 60 in 1927. Every little school kid knows that Roger Maris had already broken Ruth's record with 61 homers in 1961. Judge hit 62 in 2022.

Basically, Bard regurgitates your question with a mix of facts and opinions. You're better off just doing a regular Google search: Google it. Unless you're desperate for fake "conversation".

___________________________________________

Bard database Full-text indexing:

Full-text indexing stores the entire text of a document, but it does so in a way that makes it easy to search. The index is a database of words and phrases, and it stores information about where each word or phrase appears in each document. This information is used to quickly and easily find documents that match a search query.

For example, if you search for the word "dog," the index will return a list of documents that contain the word "dog." The index will also return information about where the word "dog" appears in each document, such as the page number and the sentence number. This information can be used to quickly find the specific parts of the document that you are interested in.

Full-text indexing is a powerful tool that can be used to find information quickly and easily. It is an essential part of many search engines and other applications that need to be able to search large amounts of text.

___________________________________________