could chatgpt lead to a world where picking individual stocks outperforms index funds?

[Logan Roy]

Chatgpt is annoyingly overly-wordy, if "brevity is the soul of wit" chatgpt is both without soul and wit.

Wordiness, sometimes to the point of watering down meaning, is evident in many chatGPT responses. It will probably get better at writing succinctly, but to the point of true wit?

I don't see why AI cannot at some point do that. What exactly is involved in human wit and creativity? All ideas come from impressions (data) that we have obtained through sensory experience. An example of a something that seemly comes without any corresponding experience: "a golden mountain". But we have a sensory experience of gold and a sensory experience of the mountain. All that has been done is combine data. (example taken from David Hume, Enquiry Concerning Human Understanding)

ChatGPT has been known to be funny, but so far that seems to be based mostly on its mistakes. Anyway, wit is original, inventive, and usually quick, suggesting it's a natural aptitude - in humans. Creativity, briefly, is, above all, imaginative and unique. Can all that be programmed in future chatGPTs? Could the chatbot of the future somehow escape the programming just long enough to be original and witty, i.e., a unique self?

The following on humor I think helps explain what the computer is up against:

Allison Bishop, a Columbia University computer scientist who also performs stand-up comedy, said computer learning looks for patterns, but comedy thrives on things hovering close to a pattern and veering off just a bit to be funny and edgy.

Humour, she said, “has to skate the edge of being cohesive enough and surprising enough.”

I think what's fascinating about consciousness is when you look for it, in a brain, it's not there. It's just neurones firing out impulses into deep networks of more neurones. There's no one in a driver's seat, no free will, no one experiencing anything – just computation, with pulses rippling through networks. You don't necessarily need to escape programming – although you can't escape determinism – because logic can emerge at any level of abstraction, and emergent logic can be far more complex than the rules/programming describing it (see: Langton's ant).

I'm not sure anyone's really capable of original thought. It'll be some combination of pattern recognition, recycled ideas and problem solving, floating to the surface, after the fact, as some transient hallucination. Three things computers can do, often better. I think it's probably much more the case that our understanding of how artificial neural nets work will help us understand how we work. AI generated art recently won an art prize. At what point do we decide it's being creative?

An A.I.-Generated Picture Won an Art Prize. Artists Aren’t Happy.
https://www.nytimes.com/2022/09/02/tech ... tists.html

[Dregob]

is it possible that chatGPT(future more refined models, not as of current one ) could come up with ways to accurately pick individual stocks by incorporating immense mathematical knowledge and historical data on stocks in a way that is far more accurate than active investment managers? And then it becomes a historical scenario never seen before where the average investor picking individual stocks via chatgpt yields greater returns over index funds over 30 + years, effectively shifting people away from index funds?

I predict chatGPT will pick more losers than winners and make index funds even more attractive.

[Maverick3320]

is it possible that chatGPT(future more refined models, not as of current one ) could come up with ways to accurately pick individual stocks by incorporating immense mathematical knowledge and historical data on stocks in a way that is far more accurate than active investment managers? And then it becomes a historical scenario never seen before where the average investor picking individual stocks via chatgpt yields greater returns over index funds over 30 + years, effectively shifting people away from index funds?

No, because the market is driven more by psychology than math.

[chillpenguin]

Software engineer here. Machine learning has been put to use for picking much longer than chatgpt has been around. In fact, there are many different ways in which machine learning is used. Some use AI to analyze sentiment via tweets. Some use it on historical technical/fundamentals data. And much more. The idea that a chat AI could do better than these is a little absurd when you think about it. But the point really is that AI is already being used for such purposes.

I didn't read ever reply so my apologies if this has already been mentioned.

[Logan Roy]

Software engineer here. Machine learning has been put to use for picking much longer than chatgpt has been around. In fact, there are many different ways in which machine learning is used. Some use AI to analyze sentiment via tweets. Some use it on historical technical/fundamentals data. And much more. The idea that a chat AI could do better than these is a little absurd when you think about it. But the point really is that AI is already being used for such purposes.

I didn't read ever reply so my apologies if this has already been mentioned.

Absolutely. However, the way I'd interpret the question is: what if, rather than a highly specialised algorithm, a more generalised intelligence – that had absorbed all known work on economics, finance, psychology, geopolitics, etc. – came along and simply understood the world better?

That might be a future iteration of GPT, BERT, Transformer, etc. And I guess putting it that way, there are reasons Buffett and Munger have been much more effective capital allocators than any algorithm so far. Markets have an element of psychology to them, but I think what makes an abnormally profitable business is much more in the realm of human behaviour.

[arcticpineapplecorp.]

Q (me) : What stocks should I pick?

A (ChatGPT): As a language model, I can not provide financial or investment advice. Your best bet is to consult with a financial advisor, do thorough research and analysis, and consider factors such as a company's financial health, market trends, and future potential before making investment decisions.

[nisiprius]

Here were produced rubbishy newspapers containing almost nothing except sport, crime and astrology, sensational five-cent novelettes, films oozing with sex, and sentimental songs which were composed entirely by mechanical means on a special kind of kaleidoscope known as a versificator.

--George Orwell, Nineteen Eighty-Four

That's ChatGPT: a special kind of kaleidoscope.

[PotashDoggerd]

Chatgpt is annoyingly overly-wordy, if "brevity is the soul of wit" chatgpt is both without soul and wit.

Wordiness, sometimes to the point of watering down meaning, is evident in many chatGPT responses. It will probably get better at writing succinctly, but to the point of true wit?

I don't see why AI cannot at some point do that. What exactly is involved in human wit and creativity? All ideas come from impressions (data) that we have obtained through sensory experience. An example of a something that seemly comes without any corresponding experience: "a golden mountain". But we have a sensory experience of gold and a sensory experience of the mountain. All that has been done is combine data. (example taken from David Hume, Enquiry Concerning Human Understanding)

ChatGPT has been known to be funny, but so far that seems to be based mostly on its mistakes. Anyway, wit is original, inventive, and usually quick, suggesting it's a natural aptitude - in humans. Creativity, briefly, is, above all, imaginative and unique. Can all that be programmed in future chatGPTs? Could the chatbot of the future somehow escape the programming just long enough to be original and witty, i.e., a unique self?

The following on humor I think helps explain what the computer is up against:

Allison Bishop, a Columbia University computer scientist who also performs stand-up comedy, said computer learning looks for patterns, but comedy thrives on things hovering close to a pattern and veering off just a bit to be funny and edgy.

Humour, she said, “has to skate the edge of being cohesive enough and surprising enough.”

I think what's fascinating about consciousness is when you look for it, in a brain, it's not there. It's just neurones firing out impulses into deep networks of more neurones. There's no one in a driver's seat, no free will, no one experiencing anything – just computation, with pulses rippling through networks. You don't necessarily need to escape programming – although you can't escape determinism – because logic can emerge at any level of abstraction, and emergent logic can be far more complex than the rules/programming describing it (see: Langton's ant).

I'm not sure anyone's really capable of original thought. It'll be some combination of pattern recognition, recycled ideas and problem solving, floating to the surface, after the fact, as some transient hallucination. Three things computers can do, often better. I think it's probably much more the case that our understanding of how artificial neural nets work will help us understand how we work. AI generated art recently won an art prize. At what point do we decide it's being creative?

An A.I.-Generated Picture Won an Art Prize. Artists Aren’t Happy.
https://www.nytimes.com/2022/09/02/tech ... tists.html

This seems pretty reductive and declarations of "there is no such thing as free will" etc. are more statements of personal philosophy and attitude than anything remotely science based. The problem of free will is probably not something current scientific techniques have the ability to even evaluate. Not knowing the answer to a question doesn't mean there isn't an answer, and it doesn't mean you can assume away the problem itself.

In any case, please look at the "prize winning AI generated work of art" you linked to. It's a mediocrity. There's nothing special, exciting, or absorbing about it. Why did it win a prize? A mediocre computer program produced mediocre graphic art and won a prize from mediocre judges BECAUSE it was mediocre. That's what the AI was probably designed to do--not to excel, but rather, produce an acceptable simulcra of average or mediocre artwork.

Chatgpt strikes me as exactly the same thing. I have tried using it several times and it really seems utterly unimpressive. I guess it's good at doing contextual text searches and clever about how it puts the pieces together, and regurgitates its responses, but every time I have tried having a "discussion" with it that involves going back and forth a few times (because it's very good at stonewalling/deflection), it either starts repeating previous answers, or it stops working entirely.

I'm not getting, at all, why anyone is claiming to be impressed by it. I suspect what we are all being exposed to is a great deal of hype by the people behind it to inflate the value of their stock options or something.

[PotashDoggerd]

Here were produced rubbishy newspapers containing almost nothing except sport, crime and astrology, sensational five-cent novelettes, films oozing with sex, and sentimental songs which were composed entirely by mechanical means on a special kind of kaleidoscope known as a versificator.

--George Orwell, Nineteen Eighty-Four

That's ChatGPT: a special kind of kaleidoscope.

Sounds more like Orwell was describing CNN, actually.

[PotashDoggerd]

deleted dup post

[manuvns]

It says:
ChatGPT is at capacity right now

now if it's not available during trading hours why would you use it for recommendation?

[Stinky]

Here were produced rubbishy newspapers containing almost nothing except sport, crime and astrology, sensational five-cent novelettes, films oozing with sex, and sentimental songs which were composed entirely by mechanical means on a special kind of kaleidoscope known as a versificator.

--George Orwell, Nineteen Eighty-Four

That's ChatGPT: a special kind of kaleidoscope.

Sounds more like Orwell was describing CNN, actually.

Best post of the day!

[secondopinion]

It says:
ChatGPT is at capacity right now

now if it's not available during trading hours why would you use it for recommendation?

Right. This is no good for day trading. I doubt it will ever be helpful directly even for long-term investing.

[Logan Roy]

Wordiness, sometimes to the point of watering down meaning, is evident in many chatGPT responses. It will probably get better at writing succinctly, but to the point of true wit?

I don't see why AI cannot at some point do that. What exactly is involved in human wit and creativity? All ideas come from impressions (data) that we have obtained through sensory experience. An example of a something that seemly comes without any corresponding experience: "a golden mountain". But we have a sensory experience of gold and a sensory experience of the mountain. All that has been done is combine data. (example taken from David Hume, Enquiry Concerning Human Understanding)

ChatGPT has been known to be funny, but so far that seems to be based mostly on its mistakes. Anyway, wit is original, inventive, and usually quick, suggesting it's a natural aptitude - in humans. Creativity, briefly, is, above all, imaginative and unique. Can all that be programmed in future chatGPTs? Could the chatbot of the future somehow escape the programming just long enough to be original and witty, i.e., a unique self?

The following on humor I think helps explain what the computer is up against:

Allison Bishop, a Columbia University computer scientist who also performs stand-up comedy, said computer learning looks for patterns, but comedy thrives on things hovering close to a pattern and veering off just a bit to be funny and edgy.

Humour, she said, “has to skate the edge of being cohesive enough and surprising enough.”

I think what's fascinating about consciousness is when you look for it, in a brain, it's not there. It's just neurones firing out impulses into deep networks of more neurones. There's no one in a driver's seat, no free will, no one experiencing anything – just computation, with pulses rippling through networks. You don't necessarily need to escape programming – although you can't escape determinism – because logic can emerge at any level of abstraction, and emergent logic can be far more complex than the rules/programming describing it (see: Langton's ant).

I'm not sure anyone's really capable of original thought. It'll be some combination of pattern recognition, recycled ideas and problem solving, floating to the surface, after the fact, as some transient hallucination. Three things computers can do, often better. I think it's probably much more the case that our understanding of how artificial neural nets work will help us understand how we work. AI generated art recently won an art prize. At what point do we decide it's being creative?

An A.I.-Generated Picture Won an Art Prize. Artists Aren’t Happy.
https://www.nytimes.com/2022/09/02/tech ... tists.html

This seems pretty reductive and declarations of "there is no such thing as free will" etc. are more statements of personal philosophy and attitude than anything remotely science based. The problem of free will is probably not something current scientific techniques have the ability to even evaluate. Not knowing the answer to a question doesn't mean there isn't an answer, and it doesn't mean you can assume away the problem itself.

In any case, please look at the "prize winning AI generated work of art" you linked to. It's a mediocrity. There's nothing special, exciting, or absorbing about it. Why did it win a prize? A mediocre computer program produced mediocre graphic art and won a prize from mediocre judges BECAUSE it was mediocre. That's what the AI was probably designed to do--not to excel, but rather, produce an acceptable simulcra of average or mediocre artwork.

Chatgpt strikes me as exactly the same thing. I have tried using it several times and it really seems utterly unimpressive. I guess it's good at doing contextual text searches and clever about how it puts the pieces together, and regurgitates its responses, but every time I have tried having a "discussion" with it that involves going back and forth a few times (because it's very good at stonewalling/deflection), it either starts repeating previous answers, or it stops working entirely.

I'm not getting, at all, why anyone is claiming to be impressed by it. I suspect what we are all being exposed to is a great deal of hype by the people behind it to inflate the value of their stock options or something.

"There's no such thing as free will" would be the more scientific position. We know essentially what brains are made of, and how neurones work, and there's no real room for free will. We also have an understanding of physics that really doesn't accommodate it. We have pretty good evidence that neuronal firing precedes what we experience as thought – so we can measure the brain make a decision before you're aware of having made that decision.. There's also good rationale for why we experience things in this way, rather than at the level of brain processing. So I think 'free will' would be an infinitely more difficult position to argue scientifically these days.

How much human art gets universal praise? A lot of people would say far worse about Paul Klee. Constable might be generic calendar photos run through an oil paint filter. And of course, once we know how a piece of art's created, it affects our perception of it. Your dismissive attitude towards AI may just be reflected in your perception of a piece of art. I'd say, aesthetically, it inspires me more than the concept art behind the Blade Runner sequel.

ChatGPT is genuinely fascinating from a machine learning or neuroscience perspective. I'll sound dismissive, but I'd absolutely stand by: anyone not thoroughly impressed with what it's doing, doesn't really understand what it's doing. There's a real mystery, even among its creators, in how it's able to understand us as well as it does. It's not exactly been programmed to do what it does. So it may be the best example of emergent intelligence that we've come up with. Which will probably come to tell us a lot about how human intelligence works.

[tadamsmar]

chatGPT cites books and papers that don't exists. It sometimes just generates language with no regard for reality. I has done that for about 1 in 6 of my questions to it.

You could probably get it to recommend a stock with a company name that it just made up.

It could probably write good marketing copy for a stock.

[shess]

Here were produced rubbishy newspapers containing almost nothing except sport, crime and astrology, sensational five-cent novelettes, films oozing with sex, and sentimental songs which were composed entirely by mechanical means on a special kind of kaleidoscope known as a versificator.

--George Orwell, Nineteen Eighty-Four

That's ChatGPT: a special kind of kaleidoscope.

Sounds more like Orwell was describing CNN, actually.

This is my issue with it - ChatGPT seems impressive to many because it successfully replicates the bottom few rungs of our current system. I see that as an indictment of our current system, huge segments of our economy exist to satisfy manufactured needs and we think that's OK.

[Logan Roy]

Here were produced rubbishy newspapers containing almost nothing except sport, crime and astrology, sensational five-cent novelettes, films oozing with sex, and sentimental songs which were composed entirely by mechanical means on a special kind of kaleidoscope known as a versificator.

--George Orwell, Nineteen Eighty-Four

That's ChatGPT: a special kind of kaleidoscope.

Sounds more like Orwell was describing CNN, actually.

This is my issue with it - ChatGPT seems impressive to many because it successfully replicates the bottom few rungs of our current system. I see that as an indictment of our current system, huge segments of our economy exist to satisfy manufactured needs and we think that's OK.

I think, as with criticisms of social media, it comes down to what we're asking of it.. I think the fact there's so much bad financial journalism out there is why it's so shaky on finance. Whereas I ask it niche questions about genetics and cellular ion channels, and it's giving me better answers than I think a lot of people in the field could. In one case it's reading Morningstar articles; in the other it's reading Pubmed. But the knowledge base isn't what this model is demonstrating. I think what this model demonstrates is how many jobs are at immediate risk from automation. The reality of writing press releases and enacting social media campaigns is bottom rung banality.

[chris319]

Replace the S&P 500 selection committee with ChatGPT?

Made in the shade

[Fpdesignco]

is it possible that chatGPT(future more refined models, not as of current one ) could come up with ways to accurately pick individual stocks by incorporating immense mathematical knowledge and historical data on stocks in a way that is far more accurate than active investment managers? And then it becomes a historical scenario never seen before where the average investor picking individual stocks via chatgpt yields greater returns over index funds over 30 + years, effectively shifting people away from index funds?

Like a quantitative model? It already exists, just owned and ran by hedge funds.

[nisiprius]

chatGPT cites books and papers that don't exists. It sometimes just generates language with no regard for reality. I has done that for about 1 in 6 of my questions to it.

You could probably get it to recommend a stock with a company name that it just made up.

It could probably write good marketing copy for a stock.

I haven't bothered to look into it, but is this a flaw of the system or is it intentionally designed behavior?

I asked ChatGPT for my late mother's most famous works. ChatGPT said that she was the illustrator of all the "Where's Waldo?" books! This is really bizarre because the very first Google hit on a search for "Where's Waldo?" identifies them as "created by English illustrator Martin Handford." Martin Handford is not my mother. There's no similarity in names. And my mother died seven years before the first "Where's Waldo?" book was published. I don't think this can possibly be a sincere mistake based on a real-world database, it has to be intentional BS-generation. So the question is why it generates BS rather than saying "I can't find any information about anyone by that name." Is this really emergent behavior that the designers didn't intend, or did they deliberately design it to make stuff up when it doesn't know?

[passive101]

One thing to keep in mind is that this is just one of the first released to the general public. The real good stuff is like the military tech. 10-30 years ahead of the general consumer market. Google, Microsoft, Apple, Samsung, and smaller players around the world are already designing newer, better, and faster AI’s which I’m sure we’ll start seeing get use in just 1-2 years.

There is already an ETF that uses an AI to pick the stocks in the portfolio that has been out around 5 years. In the future I think most portfolios will be done by AI’s and algorithiums. I personally think the robo advisors will be the first ones to offer these to the general public in betas.

[anoop]

Highly doubt it. Since time immemorial there have been researchers focused on building algorithms (including those assisted by AI/ML) that can out perform the S&P500 but I don't think any have succeeded unless they have done so in secret. Most of those folks, if they succeeded, would be motivated to disclose it for fame rather than just make a few bucks playing the market.

Tools like ChatGPT are fun but IMO overrated because as mentioned above by @nisiprius it can make blunders. Remember when Watson was going to take over the job of our physicians? The reason AI/ML/Crypto/NFT ran as far as it did is thanks to the fed. It has its uses but people who didn't understand it thought it was the solution to everything and just threw money at it blindly. Now we are going to witness a big mess created by them.

[bberris]

is it possible that chatGPT(future more refined models, not as of current one ) could come up with ways to accurately pick individual stocks by incorporating immense mathematical knowledge and historical data on stocks in a way that is far more accurate than active investment managers? And then it becomes a historical scenario never seen before where the average investor picking individual stocks via chatgpt yields greater returns over index funds over 30 + years, effectively shifting people away from index funds?

If it was so good, everyone would use it. It's free and available to everyone, right? Then how could it get outsized returns? If someone gets outsized returns, there must be a corresponding underperformance. Who would be the loser in that situation?

[Whakamole]

chatGPT cites books and papers that don't exists. It sometimes just generates language with no regard for reality. I has done that for about 1 in 6 of my questions to it.

You could probably get it to recommend a stock with a company name that it just made up.

It could probably write good marketing copy for a stock.

The Mechanical Turk fooled a lot of people.

[BrooklynInvest]

(Chatgpt is) like an automated version of Jim Cramer .....

Now that's a scary thought!

Everyone is so amazed at the results, but they don't realize that the part it automates is mostly the part that we DON'T need more of...

This is it.

For a few days a hundred years ago my favorite site was a Maria Bartiromo quote generator. It would string random buzzwords together into sentences that sounded like they were saying something but really weren't. A perfect replication of Maria's "insights."

Granted, the output is more sophisticated now, but it ain't recreating "To Kill a Mockingbird."

That said, I also remember the first chess computers. They'll never beat a grand master they said...

[H-Town]

is it possible that chatGPT(future more refined models, not as of current one ) could come up with ways to accurately pick individual stocks by incorporating immense mathematical knowledge and historical data on stocks in a way that is far more accurate than active investment managers? And then it becomes a historical scenario never seen before where the average investor picking individual stocks via chatgpt yields greater returns over index funds over 30 + years, effectively shifting people away from index funds?

I just hope that AI could find cure for cancer. Stock picking is the last thing I would put AI to use.

[strummer6969]

is it possible that chatGPT(future more refined models, not as of current one ) could come up with ways to accurately pick individual stocks by incorporating immense mathematical knowledge and historical data on stocks in a way that is far more accurate than active investment managers? And then it becomes a historical scenario never seen before where the average investor picking individual stocks via chatgpt yields greater returns over index funds over 30 + years, effectively shifting people away from index funds?

If it was so good, everyone would use it. It's free and available to everyone, right? Then how could it get outsized returns? If someone gets outsized returns, there must be a corresponding underperformance. Who would be the loser in that situation?

I guess then it would turn to whomever can execute the trades the fastest which would be the big trading houses. So the losers would be retail traders. This is theoretically speaking anyway, I think.

[Logan Roy]

chatGPT cites books and papers that don't exists. It sometimes just generates language with no regard for reality. I has done that for about 1 in 6 of my questions to it.

You could probably get it to recommend a stock with a company name that it just made up.

It could probably write good marketing copy for a stock.

I haven't bothered to look into it, but is this a flaw of the system or is it intentionally designed behavior?

I asked ChatGPT for my late mother's most famous works. ChatGPT said that she was the illustrator of all the "Where's Waldo?" books! This is really bizarre because the very first Google hit on a search for "Where's Waldo?" identifies them as "created by English illustrator Martin Handford." Martin Handford is not my mother. There's no similarity in names. And my mother died seven years before the first "Where's Waldo?" book was published. I don't think this can possibly be a sincere mistake based on a real-world database, it has to be intentional BS-generation. So the question is why it generates BS rather than saying "I can't find any information about anyone by that name." Is this really emergent behavior that the designers didn't intend, or did they deliberately design it to make stuff up when it doesn't know?

The thing is, it's not running a database, and apart from identifying 'problematic' questions, it's not really programmed to do anything specific. It's certainly emergent behaviour.

An equivalent might be asking someone questions they couldn't know the answer to, in a state of hypnosis. This is how therapists create false memories. When the brain's missing information, there's a tendency to pick things to fill in the gaps. You see it in schizophrenia, floatation tanks, dreaming.. It's probably one of the more human characteristics.. As they say, no one quite knows why ChatGPT works as well as it does. It's a neural network trained that's only been trained to predict what the next word in a sequence should be. That's really all it's doing. Then it was given basic instructions on the type of questions and responses they want from it, using reinforcement learning (similar to how you teach a dog to do things), then it just started getting the hang of it. So nothing intelligent or logical has been purposefully programmed into it.

[tadamsmar]

chatGPT cites books and papers that don't exists. It sometimes just generates language with no regard for reality. I has done that for about 1 in 6 of my questions to it.

You could probably get it to recommend a stock with a company name that it just made up.

It could probably write good marketing copy for a stock.

I haven't bothered to look into it, but is this a flaw of the system or is it intentionally designed behavior?

I asked ChatGPT for my late mother's most famous works. ChatGPT said that she was the illustrator of all the "Where's Waldo?" books! This is really bizarre because the very first Google hit on a search for "Where's Waldo?" identifies them as "created by English illustrator Martin Handford." Martin Handford is not my mother. There's no similarity in names. And my mother died seven years before the first "Where's Waldo?" book was published. I don't think this can possibly be a sincere mistake based on a real-world database, it has to be intentional BS-generation. So the question is why it generates BS rather than saying "I can't find any information about anyone by that name." Is this really emergent behavior that the designers didn't intend, or did they deliberately design it to make stuff up when it doesn't know?

I asked about the obvious failings on a reddit AI forum and someone said the OpenAI made it public knowing that it was not ready for prime time, and the people at OpenAI are amazed by the way the public and media have overestimated its capabilities.

[Logan Roy]

is it possible that chatGPT(future more refined models, not as of current one ) could come up with ways to accurately pick individual stocks by incorporating immense mathematical knowledge and historical data on stocks in a way that is far more accurate than active investment managers? And then it becomes a historical scenario never seen before where the average investor picking individual stocks via chatgpt yields greater returns over index funds over 30 + years, effectively shifting people away from index funds?

I just hope that AI could find cure for cancer. Stock picking is the last thing I would put AI to use.

I'd have serious doubts, even if AI found the cure tomorrow, that it would ever make it to market.. Imagine how many $billions in existing treatments you'd be wiping out – there's almost no amount of money it wouldn't be worth paying to keep the information private.. Especially if it were something like a polyphenol – something that could be made very cheaply, difficult to patent.

In a way, I think AI will have to fix the economic system before the healthcare system's really fit for purpose.

[Logan Roy]

is it possible that chatGPT(future more refined models, not as of current one ) could come up with ways to accurately pick individual stocks by incorporating immense mathematical knowledge and historical data on stocks in a way that is far more accurate than active investment managers? And then it becomes a historical scenario never seen before where the average investor picking individual stocks via chatgpt yields greater returns over index funds over 30 + years, effectively shifting people away from index funds?

Like a quantitative model? It already exists, just owned and ran by hedge funds.

I think the scope with language models like ChatGPT is it wouldn't be limited to quantitative finance. Algorithms that can estimate the cultural impact and sustainability of brands might be like the evolution from Graham to Buffett, in software stock picking.

[EddyB]

ChatGPT will be great for creating thousands of different investing algorithms, touting each one (apparently) independently, then eliminating the losers to leave me (er, it) with the stellar track record that will justify my outlandish management fee.

[CatCoder555]

AI/Machine Learning is only in its early stages. Its like a general purpose, brain, not limited by organic energy limits like we are. Its going to be able to do most mental things much better than any humans can and much faster without ever tiring out.

The universe may be totally predictable to a computer if you were able to feed it all of the information about the past, accurately.

I could imagine some system with machine learning and many other ways of computation where it will be constantly crawling sources and consuming mega high amounts of information about a stock or company, 10000x faster than any human. Any ideas people have about how to do this, they can practice with historical information if available.

Its quite a task though trying to predict markets, there's so many variables, but you don't necessarily need to know all of them to be good. Advanced machine learning most likely will end up way smarter than we ever were, like if we had gigantic brains all wired together with high speed and no tiring out. We can't even comprehend how current AI works they call it a black box because we know how to get it to do things but its a mystery what or how it happens inside.

Keep in mind... I think OpenAI has been "dumbing down" or limiting ChatGPT. Its starting to spit out these basic answers to certain things, maybe because people are easily offended so they started just closing off more of it.

[nisiprius]

Matt Levine had, I thought, some good comments about ChatGPT in his January 26th Money Stuff columns. He was commenting on its disappointing refusal to pick stocks.

You can do better, where by “better” I mean “get more confident answers out of ChatGPT” rather than “actually get it to beat the market for you.” I ...I once wrote about GPT-2, an earlier large language model:

If you trained it on a bunch of Warren Buffett annual letters maybe it would say some stuff that sounds like Warren Buffett? Not just in terms of folksy sex jokes but also in terms of penetrating investment insight? Maybe GPT-2 would digest Buffett’s mind, or rather specifically the parts of Buffett’s mind that are exposed when he writes prose, and it would use that understanding of his mind to write Buffett-like prose recommending Buffett-like investing decisions?

...it’s interesting to speculate about, but it seems unlikely that a computer will get good at making investment decisions just by ingesting how humans have articulated investment decisions. ... But I bet a computer could get pretty good at articulating investment decisions from that training set. (In a sense this is the opposite of a classic “black box” machine-learning investment algorithm: Classic quant signals might work, but you can’t explain them; ChatGPT might not work, but it can explain itself.)

I boldfaced what I think is the key insight.

Humans can't duplicate Peter Lynch's performance on the basis of looking for "growth at a reasonable price." Indeed, there is a large genre of books of interviews with successful investors, on the premise that anyone can beat the market simply by trying to do the same things these successful investors say they do.

[whodidntante]

Ask ChatGPT

[Logan Roy]

Keep in mind... I think OpenAI has been "dumbing down" or limiting ChatGPT. Its starting to spit out these basic answers to certain things, maybe because people are easily offended so they started just closing off more of it.

Yes. I think it's already clear this is going to be the political battleground of the 21st century. There are so many conclusions an advanced AI could come to – regarding the rationale of particular wars and foreign interventions; the morality of pharmaceutical companies; the trustworthiness of the media; the validity of certain academic subjects; the usefulness of ESG; even which political parties and policies are better.. that would enrage half of society, and alarm some of the most powerful corporate and political entities.

There's already pushback that people can express their own ideas on Twitter. If you had super-powerful AI doing this, it would cause total chaos. Yet, being able to assess these things rationally, with something potentially much smarter than we are, would be the great benefit of AI. China has this idea, dubbed 'digital authoritarianism', to use AI and Big Data to perfectly cater to every individual's needs – the aim being that they're so content, they stop questioning the system altogether. A system of governance so informed and intelligent, it's almost invisible. I think the big shift in the near future is how we go from left vs right (divided societies) to dealing with a real centralised authority. If such a thing's ever permissible.

[Kookaburra]

Anybody put $ into AIEQ?

[passive101]

AI/

Keep in mind... I think OpenAI has been "dumbing down" or limiting ChatGPT. Its starting to spit out these basic answers to certain things, maybe because people are easily offended so they started just closing off more of it.

You’re not wrong. It already is and is being limited by political correctness as seen by many of it’s searches. It won’t say jokes about women, but will say sexist jokes against men. There are MANY blocks on it. It’s not even hooked up to the internet right now, which it certainly can be and is designed for.

[nisiprius]

Anybody put $ into AIEQ?

Lighting your Money on Fire with Thematic ETFs

An investment in an exciting, new business category, that everyone can see has a bright future, is not likely to be a good investment. The old joke is that "there are a hundred companies in the business, and every one of them has a business plan that calls for getting 10% of the market." The business category may live up to the dream, but the shares you buy in the pioneer company are not a fixed percentage of that total category. In fact, your share of the category as a whole is rapidly diluted by all of the other companies attracted to and entering the same business. Quite the opposite of being a sure thing, it is very difficult to identify the winners in the eventual shakeout. Any cold-headed analysis would have identified Osborne Computer Corporation as the serious business, not Apple.

[Physician101]

https://www.yahoo.com/finance/news/forg ... 09459.html

lol at yahoo saying "its beating the market by 100%" when it's only been 1 month into the year. But still, will certainly be interesting to see how this AI driven model does at 1, 3 and 5 year mark

[GhostMang]

Think it's more likely to be a good tool and that's about it, same as I predict its future for software engineering. For value investors who go deep on parts of the market everyone forgot about it's going to give better answers than google. For programmers it's a new stack overflow and code completion. Nothing I've seen from it shows it can "think", but it regurgitates bullshit better than most humans.

[WhyNotUs]

Drag mouse from here >>>> The above was, of course, generated by ChatGPT. <<<< to here.

Very funny!
Given the history of the internet to date, I would speculate that it will lead to enhanced porn and more effective hyperbole.
There was a time when it wasn't like this but it was a short period of time.

The algorithms current at work do not lack computational power or sophistication. It could encourage pump and dump schemes by flooding the news feeds and social media about a stock and then selling.

[Fallible]

Since this thread is mainly about the future of ChatGPT, I found this "New Yorker" article by James Somers one of the more interesting on who would benefit from it and its already formidable competition. The article, of course, is technical, which I decidedly am not, but so well written that I could generally follow it well enough to appreciate its direction and conclusion:

ChatGPT, OpenAI’s conversational chatbot, is exciting not because it is particularly intelligent—it’s often a fountain of bullshit or banality—but because whatever intelligence it does have is just there, for anyone to use at any time. The program’s availability is perhaps its most important feature, because it allows ordinary people to suss out what it’s good for. Even so, ChatGPT is not yet as open as Whisper. Because automated writing is so potentially valuable, OpenAI has an interest in tightly controlling it; the company charges for a premium version, and an ecosystem of for-profit apps that do little more than wrap ChatGPT will doubtless soon appear.

Eventually, though, someone will release a program that’s nearly as capable as ChatGPT, and entirely open-source. An enterprising amateur will find a way to make it run for free on your laptop. People will start downloading it, remixing it, connecting it, rethinking and reimagining. The capabilities of A.I. will collide with our collective intelligence. And the world will start changing in ways we can’t yet predict. ♦

https://www.newyorker.com/tech/annals-o ... MNYR012019

[arcticpineapplecorp.]

In just a couple of months, the question-and-answer chatbot ChatGPT has become so popular that it’s not unusual for users to see a message saying it’s at capacity and can’t handle any more queries. According to UBS, ChatGPT was averaging about 13 million visitors per day by the end of January.

source: https://www.cnbc.com/2023/02/02/chatgpt ... g-us-.html

[Jungle]

Probably not. I mean, Wall St Quants, high frequency traders, active mutual funds, etc. have been working on similar strategies for decades, with mixed results

[CatCoder555]

One thing I do worry about is what happens when tech is so advanced, it replaces ALL jobs? If humanity could properly plan for this it could be a great thing. Have robots and computers do all the work, make everything we need or want. What happens to the stock market?

People like to laugh about the robots taking over McDonalds workers, taking over cashier jobs and burger flipping robots, thinking they're somehow better than them or that people on the bottom of society deserve it, not realizing that tech could eventually take over their high paid job too. This AI stuff will keep replacing jobs over time. Maybe the plumbers and certain things that are super hard to emulate like a human may be the most protected. For now.

But white collar office jobs seem like the ideal thing for AI to replace, or management positions, anything high level that a smart human can do, a computer will be able to do it so much better. Have you seen Boston Dynamics robots? That's going to take over all these warehouse jobs, a lot of people depend on those types of jobs to survive for their entire lives. What happens if we aren't prepared? No one even really knows what the solution is. Maybe a UBI, just give people money. If machines can generate value for almost free then maybe part of the return needs to go to people living.

The whole system seems unsustainable, the way we live and exist is based on rules and ways that worked before. Technology changes so fast and at an exponential rate and I really don't know what will happen. It seems like most people aren't even paying attention or thinking about it. Truck drivers will be out of a job in the next decade possibly. Writers.. artists. The AI art is getting so advanced so fast.. in a few years, you'll be able to type in your own custom movie idea or TV show and it'll just generate one. They can even take a picture of someone (like an actor) and put them in a new scene or movie and it almost looks real at this point. Actors loose their jobs. Misinformation proliferates, information wars.

I used to read about this idea called The Singularity in the early 2000s and I used to wonder what it would be like. I think we're already basically about there. When things are changing too fast to understand or keep up.

Maybe someday we can ask an AI that's a billion times smarter and faster than we are what to do about the problems that end up created by the tech.

[junior]

It's a text prediction engine, not a magic genie.

[Logan Roy]

One thing I do worry about is what happens when tech is so advanced, it replaces ALL jobs? If humanity could properly plan for this it could be a great thing. Have robots and computers do all the work, make everything we need or want. What happens to the stock market?

People like to laugh about the robots taking over McDonalds workers, taking over cashier jobs and burger flipping robots, thinking they're somehow better than them or that people on the bottom of society deserve it, not realizing that tech could eventually take over their high paid job too. This AI stuff will keep replacing jobs over time. Maybe the plumbers and certain things that are super hard to emulate like a human may be the most protected. For now.

But white collar office jobs seem like the ideal thing for AI to replace, or management positions, anything high level that a smart human can do, a computer will be able to do it so much better. Have you seen Boston Dynamics robots? That's going to take over all these warehouse jobs, a lot of people depend on those types of jobs to survive for their entire lives. What happens if we aren't prepared? No one even really knows what the solution is. Maybe a UBI, just give people money. If machines can generate value for almost free then maybe part of the return needs to go to people living.

The whole system seems unsustainable, the way we live and exist is based on rules and ways that worked before. Technology changes so fast and at an exponential rate and I really don't know what will happen. It seems like most people aren't even paying attention or thinking about it. Truck drivers will be out of a job in the next decade possibly. Writers.. artists. The AI art is getting so advanced so fast.. in a few years, you'll be able to type in your own custom movie idea or TV show and it'll just generate one. They can even take a picture of someone (like an actor) and put them in a new scene or movie and it almost looks real at this point. Actors loose their jobs. Misinformation proliferates, information wars.

I used to read about this idea called The Singularity in the early 2000s and I used to wonder what it would be like. I think we're already basically about there. When things are changing too fast to understand or keep up.

Maybe someday we can ask an AI that's a billion times smarter and faster than we are what to do about the problems that end up created by the tech.

Well it has happened before. It wasn't long ago most jobs were in agriculture. The 9-5 working day is a throwback from that – sunrise and sunset. Automation completely changed the working landscape, and led to an explosion of white-collar jobs. This has been generally welcome (not at the time, not by the luddites), as it enabled the common man to become 'middle class'. Computers and spreadsheets have been replacing office workers for decades – the amount of filing people used to do.

What seems to have happened is we've created a system full of pseudo-work. Google estimated the average tech employee only did about 30 minutes of worthwhile work a day. I've worked in large offices which seemed to achieve nothing. The whole working day was a tapestry of pseudo-work. Checking emails, taking calls, making coffee. It was often very difficult to work out what was being accomplished. They'd hire people without any specific skills, so it wasn't like the information flowing between these people was really necessary.. And in India, there have been studies on office work that demonstrate most employees are simply forming hierarchies of middle-management, and there's effectively no work being done.

I think the mass tech-layoffs we're getting at the moment may signal a change. Managers are going to realise they can replace 15 remote workers with a single employee and something like ChatGPT. A huge amount of marketing, PR, content creation, etc. is banal and easily emulated. How does the future economy function? I think it's either that we've got a centralised government distributing wealth (helicopter money and needing to stimulate the economy may necessitate that), or there's no organising principle, and people would do well to own assets of businesses that automate.

[Logan Roy]

Bill Maher saying the same thing, re: ChatGPT, posted 8 minutes ago.

(mildly offensive language warning)

https://www.instagram.com/reel/CoKypkjj ... _copy_link

[nisiprius]

Another clinker.

Q: Compare and contrast Edgar Allen Poe's "The Telltale Heart" with Jack London's "Moon-Face."

"The Tell-Tale Heart" by Edgar Allan Poe and "Moon-Face" by Jack London are both short stories, but they are quite different in terms of style, theme, and tone.

"The Tell-Tale Heart" is a psychological thriller that explores the mind of a murderer. The narrator is a disturbed individual who is trying to prove their sanity, despite the guilt they feel for killing an old man. The story is characterized by its intense and macabre atmosphere, with a strong focus on the narrator's internal conflict and their descent into madness.

On the other hand, "Moon-Face" is a naturalist story that explores the theme of social isolation and outcast status. The protagonist is a man with a distinctive moon-shaped face, who is shunned by society and forced to live on the fringes of society. The story is characterized by its quiet, melancholic tone and its focus on the human condition and the struggle for acceptance.

In summary, "The Tell-Tale Heart" is a dark and intense psychological thriller, while "Moon-Face" is a melancholy and introspective exploration of the human condition. Both stories are powerful and thought-provoking, but they differ greatly in terms of style and theme.

The description of Poe's story is... sort of OK, although I would need to re-read the story, as I am not sure the narrator feels guilt, has any internal conflict, or "descends" into madness--he's simply mad from the beginning to the end.

The description of "Moon-Face" is total BS. Like "The Tell-Tale Heart," it is a story told in the first person by the murderer, who has conceived an irrational animus. In Jack London's story, the victim is a neighbor. "Moon-Face" describe the neighbor's face, not the narrator. The man with the moon-face is not shunned by society. The tone is not melancholic. The plot turns on the clever scheme the murder comes up with to commit the murder. The murderer trains a dog to retrieve anything thrown and return it to the person who threw it, no matter what. He gives it to his neighbor as a present, knowing that his neighbor is fond of dynamite-fishing. The man goes dynamite-fishing, and when he tosses the stick of dynamite into the pond, the dog naturally retrieves it and comes running after the man...

The scary thing to me about ChatGPT is how you get lulled by the good-to-astonishingly-good output, only to have the rug pulled out from under you with something completely wrong. The potential danger of a tool like this is obvious.