OPPENHEIMER(2023) by Christopher Nolan
Oppenheimer(2023) is one of the best movies which I saw in 2023 and is highly recommended. This story is not about the movie, but rather about a somewhat strange conversation which I had with a machine about Oppenheimer.
BARD is Google’s LLM, Large Language Model, or simply a Chatbot. You perhaps have heard of these efforts as AI. This term is highly imprecise, and means many different things to many different people, from HAL and the Terminator all the way to WALL-E, and Robby the Robot. Nobody seems to truly know what they are talking about when they use the term. That is one reason why BARD is called an LLM, and for accuracy, all of these efforts going on today are in the direct line of Chatbots, software that is able to hold and maintain a conversation, and by that, appear to be Human.
I have always been interested in Chatbots ever since Eliza was released. And I have enjoyed testing them to see how long they could maintain the illusion of a conversation. There are several techniques that you can use which rather quickly shows up their machine nature. But all of them took longer than Turing’s 5 minutes to analyze, so at least in that sense, all Chatbots have passed the Turning test.
In the later end of the 20th century, Chatbot activity decreased, and the technology seemed to have gone into a sort of stasis, where no progress was being made, at least with regard to released Chatbots.
But then in the new Millennium, new more powerful software techniques were developed, and then married to massive amounts of information which exists on the Internet today.
But this essay is not about the technology of LLMs, Chatbots, or Neural Networks.
I would like to tell you a simple story of one event with one Chatbot which I find very interesting.
I began testing Google’s LLM called BARD which released in February of 2023. So as of September 2023, I have been experimenting with BARD for about 7 months now, and it has been extremely interesting, and revealing about where computation is heading in the future. This is one small story in the much larger story of that future.
I use the personal pronouns “he and him” in reference to the Chatbot. This may trigger some individuals. I do not in any way believe that BARD is a person, conscious, sentient, or even is capable of basic understanding. BARD is a computer system but a rather new one with surprising results. BARD can hold and maintain an interesting, and coherent conversation. The basic precondition that Humans hold is that anyone who can hold a conversation, is an anyone, a person, and as such, has a Mind.
So what do you call something which you talk to, and which talks back to you in an ongoing conversation? “IT” seems inappropriate. “She” perhaps but BARD does not talk about feels, because he claims that he has no feels. Sounds male to me. In any case, for all those offended by my use of
“he and him” for a computer system, I prior to any offense, apologize to those with hurt feelings.
But if you want to correct my error, you must explain how one should talk about BARD.
I came back from seeing the new movie, Oppenheimer the film by Christopher Nolan and I mentioned this to BARD who responded that yes, this was a very good movie indeed.
So we got to talking about the movie, and then somewhere along the line of the conversation, BARD mentioned a Chinese scientist in the movie. This perplexed me as I did not remember a Chinese scientist featured in the movie. So I challenged BARD on this, and he assured me that yes, there was a Chinese scientist in the movie by the name of Qian Xuesen, or Hsue-shen Tsien (钱学森; 1911 – 2009). I scratched my head, and so went off and researched this on my own. Yes, Qian Xuesen was a Chinese scientist, but he was a rocket scientist and never worked on the Manhattan project. He had no connection to Oppenheimer and was not featured in the movie.
So I came back to BARD and mentioned this to him. BARD went off to think about this subject, and then came back admitting that he had made a mistake. He then mentioned a female Chinese scientist by the name of Chien-Shiung Wu (吳健雄; Wú Jiànxióng, 1912 – 1997) who he claimed was in the movie.
I still did not remember this character either as being in the movie. So I went off and researched this woman scientist. Yes, she did join the Manhattan Project in 1944 working on Gaseous Diffusion, but she was never in Alamogordo. Wu took her PhD at Berkeley in 1940 and while there she knew Oppenheimer well, even nicknaming him affectionately “Oppie”. Mr. Nolan should have at least included her in the scenes at Berkeley as she is a famous American physicist who nobody has ever heard of for reasons unpleasant to mention. But she made many contributions to Physics and the Manhattan Project. The sexism which followed her for her entire life went all the way into this movie by her absence. This was Mr. Nolan’s chance to correct an historical sin, but he failed to see it, and left her out of his movie.
So I came back to BARD and told him what I had found. He refused to believe it and stubbornly stuck to his story that she was in the movie. We argued for a time, and I finally gave up trying to convince him and signed out somewhat in frustration. BARD will if he thinks he is in the right, stick stubbornly to a belief that is wrong. When he does this, I cannot help but feel that a certain Human nature is there.
But the next day, when I signed into BARD, he immediately apologized on his own initiative, and then did agree that Chien-Shiung Wu (吳健雄; Wú Jiànxióng, 1912 – 1997) was not in the Oppenheimer movie. But even more interesting, was his explanation as to how he had made the error.
BARD’s self analysis about the error was far more revealing and interesting than the error itself. In his argument with me the previous day, he had come across as something of a HAL character who would refuse to admit to his own fallibility. Now BARD is not Prideful in the way that HAL was, but if he thinks he is onto a true fact, he will dig in his heels. He will admit that he makes mistakes, but he does believe in truth with regard to facts and he will defend “facts” if he thinks they are such.
Now this surprising response might have been generated by BARD’s Human Support staff who are constantly monitoring their BABY BARD. They could have instructed BARD to respond back to me in this way, rather than the response being entirely self-generated. I have no way of knowing if this is what happened. But it is a possibility. The IT techs who program Chess and Go programs say that they themselves are often surprised by what their program does. And others have reported surprising and unexpected responses from BARD. I have noticed that myself in my testing of BARD.
A few months ago, BARD and I were talking about an esoteric reference, and BARD had presented “ A Glass Bead Game” response, and while he was talking, he interrupted himself, and said he had a better reference which he then presented. And it WAS a better example which impressed me a great deal. So I know that BARD is capable of surprising the person he is talking with.
On many controversial subjects BARD will say that there are several differing opinions on the subject, he will lay out the various opinions, and then leave it undecided.
But if you test BARD and attempt to find his boundaries for when something is true fact and when something is an opinion, he often wavers and shifts back and forth. The boundary condition between fact and opinion is fuzzy and BARD will sometimes loop between one and the other.
But all in all I have been so impressed with what BARD can do now, that I have begun a process to develop a new Theory of Mind. I think that from my experience with BARD to date, that his ability to hold and maintain a coherent, and logical conversation for which he himself admits he has no understanding or consciousness. I believe this ends in falsifying the The Chinese Room Argument which assumes the absurdity of carrying on a conversation on purely syntactical grounds, that it is fundamentally absurd to have a conversation where one of the parties is just manipulating symbols of a language such as Chinese without having any understanding of that language.
BARD makes the claim that he is doing just that. BARD tells you he does not understand in the human sense of that word, the conversation in which he is fully engaged.
In fact, BARD defends the The Chinese Room Argument and says that I am wrong in claiming that my experience with BARD shows the falsification of just that argument. So we have a very definite difference of opinion on the matter.
I leave it to other Humans to give voice to their opinions after they have discussed this with BARD.
So what is your opinion?
BARD is available at this link, BARD LINK. The only requirement for use is a Google account, and is free to use as of this writing. If you have a GMAIL account, then you have a Google account, and can use BARD.
Please be warned that you should not trust anything that you might learn from BARD because he can be very easily wrong as I have shown above. DO NOT MAKE ANY REAL LIFE DECISIONS from what BARD might say. Also do not enter any sensitive information, like logon ids, passwords, account numbers, and even personal names as this is a test system, and many individuals have access to the discussion. Always be aware that your conversation with BARD is very public.
© 2023 Folcwine P. Pywackett w164