I agree to a large extent. I've been an evangelist for openness throughout my career, but I now see the trusted semi-commons as a refuge--not for scoundrels but for those still committed to an honest exchange of views. Hopefully, such mechanisms won't be echo chambers, but hope springs eternal. People don't like their orthodoxies challenged, although at least such an approach would presumably avoid the worst of the bad actors and actions.
Does anyone know the state of AI use for medical diagnosis? That would seem to be a good fit for it and could be effective in producing better treatment outcomes and lowering medical costs.
Wouldn't FB be the most vulnerable out of the five? A reveal that [major scandal] was entirely generated by AI could corrode trust in the platform and cause users to flee when the *ahem* Meta project had been just hemorrhaging money when upper management seems wed to pushing it.
But overall, I think the "we are forked" attitude is overblown. It undervalues humanity's capacity to adapt to new technologies, overlooks positive uses of the technologies, presents what I think is a simplistic vision of the human mind as easily replicated, and exemplifies an alarmist tendency in neoconservatism, where everything is Munich, everything is HAL, everything is doomed.
Here is an exchange with CHatGPT about the two sentences. I have a strange sense that she (ChatGPT) is tired of hearing about these sentences.
ME: AI researchers often discuss these two sentences: 1. "The committee denied the group a parade permit because they advocated violence." 2. "The committee denied the group a parade permit because they feared violence." Can you tell who the pronoun "they" represents in sentence 1? Who does it represent in sentence 2? Explain how you know the answer.
ChatGPT: In sentence 1, the pronoun "they" represents the group that the committee denied a parade permit to. This is because the sentence states that the reason for denying the permit was because "they advocated violence."
In sentence 2, the pronoun "they" represents the committee. This is because the sentence states that the reason for denying the permit was because "they feared violence" implying that the committee feared violence.
ME: How do you know this means the committee feared violence? It could mean the group feared violence.
CHAT: You are correct that the sentence can be interpreted in different ways and "they" in sentence 2 could refer to the group, who feared violence. The context of the statement is not clear and it's possible that it could be interpreted that the group feared violence. The given sentence by itself does not provide enough information to determine who "they" refers to with certainty.
ME: That is a good answer. Can you point to a reason why it is more likely the committee fears violence than the group?
CHAT: Without additional context, it is not possible to determine with certainty who "they" refers to in sentence 2 and whether the group or the committee fears violence. However, in a situation where a committee is responsible for granting or denying permits for events, it is more likely that the committee would be the ones fearing violence, as they are responsible for ensuring the safety of the community. The group, on the other hand, may not necessarily have the same level of concern for safety and would be more interested in organizing the parade.
ME: Good answer!
CHAT: Thank you, I'm glad you found it helpful. If you have any more questions, feel free to ask.
Not interesting? Hmmm Historically, they virtually define the existence of human thought, including art, science, religion, and philosophy. Maybe not important to you, but these are the questions that have defined human culture since we looked at the stars and told stories around fires.
These things are not interesting in this context because this AI model cannot yet deal with them. If this were physics and we were discussing a single-body equation that could not be used in a multi-body simulation, the multi-body simulation would not be "interesting." Perhaps it would be more accurate to say "not relevant."
I am not a physicist. I edit and translate papers for physicists. They often talk about models that work with one or two particles, but fail with more complex systems.
The ChatGPT AI program has no actual knowledge of anything. It has no more intelligence than an old fashioned library paper index card system. There are some AI programs that attempt to simulate knowledge, and apply logic to problems, but this one does not. At least, it did not when I read about it a few years ago.
Thank you for the context of your comment. I am curious. Is there is a way to prevent bias when doing scientific inquiry, outside mathematics? My knowledge about AI is very limited.
As an aging Boomer tech writer and tester for a firm that is developing this type of technology I have resisted delving into it as being way above my pay grade. It's starting to dawn on me that, like smart phones and smart cities and smart everything else, I'd better get myself briefed on this despite being a closeted Luddite. Thanks for the newsletter!
I work with engineers who would rather crawl naked over broken glass than talk or write to end users. <smile> [Mind you, they are brilliant, admirable, and loving human beings - they just REALLY, REALLY hate to get interupted. from their algorithms,]
"In an AI world, you type 'how to change a tire' and ChatGPT simply explains how to change a tire to you."
Speaking of semantics vs syntax, there's no simple way to explain how to transform an object from being a tire into being me. That's a very complex process. And how would ChatGPT know that what you wanted to change a tire into was me? The inquiry was open-ended.
It may be clumsy, but your interpretation of an alternate meaning is incorrect. If I wanted a tire to become me, I'd say "How to change a tire into me." I'm sure there are better examples out there; this isn't one of them.
Basically, GPT is using Bayesian statistics to fill the role of "model of the Universe". It is possible to imagine some surreal anime art piece where a tire transforms into a human being, but the overwhelming percentage of inquiries with the phrase "how to change a tire to you" will, in the wider context, refer to the act of replacing a (presumably damaged) tire with another one. However if we are a tad more precise with our language - as we would be discussing this with another human-- GPT has no trouble with the concept (see transcript below)
ME: Can you explain how to change a tire to me?
GPT: Sure! Here are the basic steps for changing a tire:
Find a safe place to change your tire, away from traffic.
Turn on your hazard lights to warn other drivers.
Use your car's jack to lift the vehicle off the ground. The jack should be placed on a solid surface and under the designated jack point on your car (consult your vehicle's owner's manual if you're unsure where this is).
Use a lug wrench to remove the lug nuts that hold the flat tire in place. Turn them counter-clockwise to loosen them.
Carefully remove the flat tire and put on the spare tire, making sure it's facing the correct direction.
Replace the lug nuts and tighten them by turning them clockwise.
Lower the vehicle back to the ground using the jack.
Use the lug wrench to give the lug nuts an extra tightening.
Remove the jack and tools from the car.
Check the tire pressure on the spare tire and inflate if needed.
Remember, your spare tire is not meant to be driven on for long distances. Get your flat tire repaired or replaced as soon as possible.
ME: Can you explain to me how to transform into a tire?
GPT: I'm sorry, but it is not possible for a human being to physically transform into a tire. Tires are inanimate objects and humans are living organisms with a different physical structure. The statement is not based on any scientific or biological facts.
Addendum: it strikes me that when we get videoGPTX, and ask it to create a video showing a tire transforming into a human being, it will gladly oblige...
Thank you, Sandy. That is a terrific article and I will be passing in along to some of my colleagues and friends.
Where things really get murky is when one asks questions like, "Does life have meaning?" "What is the proof of the existence of God?"
"Can something come from nothing," i.e., the Universe.
These, and many other questions like these are almost considered unanswerable through the lens of any human being or social structure. Some fall under the category of ontology. Some may fall under the subcategory of metaphysics. They certainly pose problems to the limits of our ability to know. So, "Why is there something rather than nothing?"would be a question that may be more appropriate for astrophysicist or cosmologists rather than philosophers or theologians. Yet these questions and answers are difficult to fathom on any level. So, I had fun asking the AI bot these q's and admired its ability to resign itself to agnosticism. Or its claim the answers to be ascertainable only through the lens of objective reality; something that is unreachable by humans.
CHAT: The question of whether life has meaning is a complex and deeply personal one, and different people may have different perspectives on it. Some people may believe that life has inherent meaning, while others may believe that meaning is something that we assign to our experiences. Some may find meaning through religion, others through relationships, and others through personal accomplishments or goals. Ultimately, whether or not life has meaning is a subjective question and may vary from person to person.
That seems like a good answer to me.
These questions do not seem difficult or profound to me. My answers are:
1. Life is the product of evolution, which has no inherent meaning or purpose, so neither does life.
2. I do not know of any proof of the existence of God.
3. Cosmologists say the universe came from nothing. I wouldn't know about that.
You and I may be twin sons of different mothers. And yes, I did ask the AI that question. There have certainly been billions of people wrestling with these type of questions for a very long time, and still do. It seems intrinsic to the human experience if taken in the aggregate. I don't think they are lacking intelligence. I just wonder why these questions are so damn important to them. So if I say my life has no meaning, value, or purpose. That feels pretty bleak to say... Whether true or not. It would be hard to discount emotion out of the human experience since I've read that 75% of all decisions made by humans are generally illogical and made from a point of emotion. It's also hard for me to talk about humans as if I'm above it and not in it. I'm in it. I'm moving on to the philosophy of aesthetics. Artistic expression, including music, may be an entry to answering some of these questions with a little different flavor. Don't forget poetry and storytelling.
Well said!🙂
Well said!
Good points. Perhaps new forms of trust enhancement will be developed, but the existing approaches do not breed confidence.
I agree to a large extent. I've been an evangelist for openness throughout my career, but I now see the trusted semi-commons as a refuge--not for scoundrels but for those still committed to an honest exchange of views. Hopefully, such mechanisms won't be echo chambers, but hope springs eternal. People don't like their orthodoxies challenged, although at least such an approach would presumably avoid the worst of the bad actors and actions.
Does anyone know the state of AI use for medical diagnosis? That would seem to be a good fit for it and could be effective in producing better treatment outcomes and lowering medical costs.
I believe AI chats will improve customer service. There wouldn’t be a wait and the agent wouldn’t be in Bangalore.
Both “committee” and “crowd” are singular so I’m going to assume “they” are third parties somewhere.
Wouldn't FB be the most vulnerable out of the five? A reveal that [major scandal] was entirely generated by AI could corrode trust in the platform and cause users to flee when the *ahem* Meta project had been just hemorrhaging money when upper management seems wed to pushing it.
This post inspired me to reread Krauthammer's 1997 warning about AI. https://www.washingtonexaminer.com/weekly-standard/be-afraid-9802
I'm all for looking at the downsides of this technology, which I think are considerable. I've done some of that myself recently at https://www.splicetoday.com/digital/reality-and-hype-in-the-2020s
But overall, I think the "we are forked" attitude is overblown. It undervalues humanity's capacity to adapt to new technologies, overlooks positive uses of the technologies, presents what I think is a simplistic vision of the human mind as easily replicated, and exemplifies an alarmist tendency in neoconservatism, where everything is Munich, everything is HAL, everything is doomed.
Here is an exchange with CHatGPT about the two sentences. I have a strange sense that she (ChatGPT) is tired of hearing about these sentences.
ME: AI researchers often discuss these two sentences: 1. "The committee denied the group a parade permit because they advocated violence." 2. "The committee denied the group a parade permit because they feared violence." Can you tell who the pronoun "they" represents in sentence 1? Who does it represent in sentence 2? Explain how you know the answer.
ChatGPT: In sentence 1, the pronoun "they" represents the group that the committee denied a parade permit to. This is because the sentence states that the reason for denying the permit was because "they advocated violence."
In sentence 2, the pronoun "they" represents the committee. This is because the sentence states that the reason for denying the permit was because "they feared violence" implying that the committee feared violence.
ME: How do you know this means the committee feared violence? It could mean the group feared violence.
CHAT: You are correct that the sentence can be interpreted in different ways and "they" in sentence 2 could refer to the group, who feared violence. The context of the statement is not clear and it's possible that it could be interpreted that the group feared violence. The given sentence by itself does not provide enough information to determine who "they" refers to with certainty.
ME: That is a good answer. Can you point to a reason why it is more likely the committee fears violence than the group?
CHAT: Without additional context, it is not possible to determine with certainty who "they" refers to in sentence 2 and whether the group or the committee fears violence. However, in a situation where a committee is responsible for granting or denying permits for events, it is more likely that the committee would be the ones fearing violence, as they are responsible for ensuring the safety of the community. The group, on the other hand, may not necessarily have the same level of concern for safety and would be more interested in organizing the parade.
ME: Good answer!
CHAT: Thank you, I'm glad you found it helpful. If you have any more questions, feel free to ask.
On the Media dedicated the full hour this week to a discussion of artificial intelligence.
I'd highly recommend listening to this if you're interested in the topic. Its a very detailed and sophisticated discussion:
https://www.wnycstudios.org/podcasts/otm/episodes/on-the-media-its-a-machines-world
Not interesting? Hmmm Historically, they virtually define the existence of human thought, including art, science, religion, and philosophy. Maybe not important to you, but these are the questions that have defined human culture since we looked at the stars and told stories around fires.
These things are not interesting in this context because this AI model cannot yet deal with them. If this were physics and we were discussing a single-body equation that could not be used in a multi-body simulation, the multi-body simulation would not be "interesting." Perhaps it would be more accurate to say "not relevant."
I looked your name up. If you're the physicist, I will take your word for this. I have no counter comment.
My purpose, by asking an AI program about objective reality has epistemological implications for gaining knowledge, going forward.
I am not a physicist. I edit and translate papers for physicists. They often talk about models that work with one or two particles, but fail with more complex systems.
The ChatGPT AI program has no actual knowledge of anything. It has no more intelligence than an old fashioned library paper index card system. There are some AI programs that attempt to simulate knowledge, and apply logic to problems, but this one does not. At least, it did not when I read about it a few years ago.
Thank you for the context of your comment. I am curious. Is there is a way to prevent bias when doing scientific inquiry, outside mathematics? My knowledge about AI is very limited.
As an aging Boomer tech writer and tester for a firm that is developing this type of technology I have resisted delving into it as being way above my pay grade. It's starting to dawn on me that, like smart phones and smart cities and smart everything else, I'd better get myself briefed on this despite being a closeted Luddite. Thanks for the newsletter!
Gosh, how can you be a tech writer and a Luddite at the same time??
I work with engineers who would rather crawl naked over broken glass than talk or write to end users. <smile> [Mind you, they are brilliant, admirable, and loving human beings - they just REALLY, REALLY hate to get interupted. from their algorithms,]
"In an AI world, you type 'how to change a tire' and ChatGPT simply explains how to change a tire to you."
Speaking of semantics vs syntax, there's no simple way to explain how to transform an object from being a tire into being me. That's a very complex process. And how would ChatGPT know that what you wanted to change a tire into was me? The inquiry was open-ended.
It may be clumsy, but your interpretation of an alternate meaning is incorrect. If I wanted a tire to become me, I'd say "How to change a tire into me." I'm sure there are better examples out there; this isn't one of them.
You had me there for a second or two...
Basically, GPT is using Bayesian statistics to fill the role of "model of the Universe". It is possible to imagine some surreal anime art piece where a tire transforms into a human being, but the overwhelming percentage of inquiries with the phrase "how to change a tire to you" will, in the wider context, refer to the act of replacing a (presumably damaged) tire with another one. However if we are a tad more precise with our language - as we would be discussing this with another human-- GPT has no trouble with the concept (see transcript below)
ME: Can you explain how to change a tire to me?
GPT: Sure! Here are the basic steps for changing a tire:
Find a safe place to change your tire, away from traffic.
Turn on your hazard lights to warn other drivers.
Use your car's jack to lift the vehicle off the ground. The jack should be placed on a solid surface and under the designated jack point on your car (consult your vehicle's owner's manual if you're unsure where this is).
Use a lug wrench to remove the lug nuts that hold the flat tire in place. Turn them counter-clockwise to loosen them.
Carefully remove the flat tire and put on the spare tire, making sure it's facing the correct direction.
Replace the lug nuts and tighten them by turning them clockwise.
Lower the vehicle back to the ground using the jack.
Use the lug wrench to give the lug nuts an extra tightening.
Remove the jack and tools from the car.
Check the tire pressure on the spare tire and inflate if needed.
Remember, your spare tire is not meant to be driven on for long distances. Get your flat tire repaired or replaced as soon as possible.
ME: Can you explain to me how to transform into a tire?
GPT: I'm sorry, but it is not possible for a human being to physically transform into a tire. Tires are inanimate objects and humans are living organisms with a different physical structure. The statement is not based on any scientific or biological facts.
Addendum: it strikes me that when we get videoGPTX, and ask it to create a video showing a tire transforming into a human being, it will gladly oblige...
https://openai.com/dall-e-2/ can generate a picture based on "a tire transforming into a human being." The image is not convincing.
It does not do a good job with "a wooden doll such as Pinocchio transforming into a human being" either. Maybe it is off day for dall-e-2.
Not a good job on this one either: "From the Greek myth, show the statue of Galatea made by Pygmalion coming to life."
Forking good article ...
Thank you, Sandy. That is a terrific article and I will be passing in along to some of my colleagues and friends.
Where things really get murky is when one asks questions like, "Does life have meaning?" "What is the proof of the existence of God?"
"Can something come from nothing," i.e., the Universe.
These, and many other questions like these are almost considered unanswerable through the lens of any human being or social structure. Some fall under the category of ontology. Some may fall under the subcategory of metaphysics. They certainly pose problems to the limits of our ability to know. So, "Why is there something rather than nothing?"would be a question that may be more appropriate for astrophysicist or cosmologists rather than philosophers or theologians. Yet these questions and answers are difficult to fathom on any level. So, I had fun asking the AI bot these q's and admired its ability to resign itself to agnosticism. Or its claim the answers to be ascertainable only through the lens of objective reality; something that is unreachable by humans.
Have you tried asking ChatGPT these questions?
ME: Does life have meaning?
CHAT: The question of whether life has meaning is a complex and deeply personal one, and different people may have different perspectives on it. Some people may believe that life has inherent meaning, while others may believe that meaning is something that we assign to our experiences. Some may find meaning through religion, others through relationships, and others through personal accomplishments or goals. Ultimately, whether or not life has meaning is a subjective question and may vary from person to person.
That seems like a good answer to me.
These questions do not seem difficult or profound to me. My answers are:
1. Life is the product of evolution, which has no inherent meaning or purpose, so neither does life.
2. I do not know of any proof of the existence of God.
3. Cosmologists say the universe came from nothing. I wouldn't know about that.
You and I may be twin sons of different mothers. And yes, I did ask the AI that question. There have certainly been billions of people wrestling with these type of questions for a very long time, and still do. It seems intrinsic to the human experience if taken in the aggregate. I don't think they are lacking intelligence. I just wonder why these questions are so damn important to them. So if I say my life has no meaning, value, or purpose. That feels pretty bleak to say... Whether true or not. It would be hard to discount emotion out of the human experience since I've read that 75% of all decisions made by humans are generally illogical and made from a point of emotion. It's also hard for me to talk about humans as if I'm above it and not in it. I'm in it. I'm moving on to the philosophy of aesthetics. Artistic expression, including music, may be an entry to answering some of these questions with a little different flavor. Don't forget poetry and storytelling.