Hacker News is, in my experience, and despite its benefits, susceptible to a belief in pseudoscience. I love the optimism behind the refrain of “what’s next?” But, at the end of the day, I wish the community, myself included, did a better job of acknowledging that this optimism too frequently sacrifices comprehension and understanding for the dopamine released when we contemplate possible futures.
Crypto mania really highlighted how gullible the average user is in buying into the hype. Worse yet, people peddling things that are objectively regressive versus existing solutions.
> Never did I expect so many dorks to turn to the dark side like this in my life. :/
I think this is a little overly optimistic about “dorks”, but I will also note the shift in the ratio of dorks to business people.
I’ve recently seen backlash against the term “tech bro”. Bill Burr recently called it out as a bullshit title to separate the bad actors from oneself, but I’ve always held it identifies a class of interest in the tech sector.
University fraternities get large sums of money from alumni and as such are financially motivated to put their pledges in degrees of that result in high paying positions.
In the 70s this was law, the 80s it was finance, 90s it was medicine, 00s it became tech.
6 figure entry level position for someone without any knowledge of an if statement were du jour for this group of people and so they flooded in. Bringing their ideals and sociopathy with them.
Woz is a dork and the brains that built apple, but his narrow interests allowed the business end to make the globally affecting decisions and we are still suffering the consequences.
At the end of the day, I guess I just don't really see what the issue is with AI hallucinations or how they make them less intelligent.
My brother in law who is extremely smart, and very well trained -- MIT PhD -- and exposed to orders of magnitude more data than ChatGPT, given that the human senses ingest gigabytes of data per day, has -- like everyone else -- persistent errors in his thinking. I single him out because of his credentials, but everyone has these.
As a particularly funny example, we were talking about steers and bulls and he was like "Oh, right, I remember, some male cattle are born bulls and some are born steers". My wife, who is similarly educated, believed for a very long time that watermelon's grew on trees. At my undergrad school, a woman who was getting her physics degree and then who eventually went on to get a PhD, fundamentally believed that yeast would not rise if you talked around it -- I guess her parents had told her that and she never questioned it.
That is to say, humans lack full knowledge. We're trained on significantly more data than ChatGPT et al. We still have persistent inaccuracies despite all of that. We're not perfect.
That doesn't make us less useful. Half the time we just rely on other humans trained on similar, but slightly different data, to iron those out, and indeed AI systems that employ those methods get better accuracy.
Now again, my brother in law has fewer errors than most people since he is very smart. However, the vast majority of people are really not that bright (rumor has it that fifty percent have below average intelligence), and yet, simply through verification with other humans, they are able to accomplish useful things.
Sometimes even very capable people have persistent thinking errors. This doesn't make them less smart. No one can claim perfect intelligence. That's the only thing that's bullshit.
I just want to point out that I think ChatGPT has seen more data than a single human being. With 45 TB of text data [1], it would cost a human around 70k years of continuous reading (assuming 1 GB is 200M words, and humans read 225 words per minute) to see all the training data.
Humans have a much richer IO system than any GPT model. Estimates say humans ingest about 75GB of data per day. That means ChatGPT has seen about 600 days worth of human data. Or about two years. So if it's smarter than a two year old, it's safe to say that it surpasses human ability.
The mistake is thinking that 'data' means only text.
> We're trained on significantly more data than ChatGPT et al.
I don't think that's true anymore. GPT-4 is rumored to have been trained on a dataset with >1 trillion tokens. An average human would take a few thousand years to read through that dataset once.
That means for a human baby to be as smart as GPT they would have to go from zero to speaking English within two weeks of birth.
Your error is mistaking training data for text only. Humans are particularly inefficient with text because it takes a lot of visual data to get to the thought of a particular word or character due to how human senses work (we have no 'token' sense).
> At the end of the day, I guess I just don't really see what the issue is with AI hallucinations or how they make them less intelligent.
At the end of the day, you wouldn't be bothered if Average Joe would add arsenic to your favourite pie, because AI told it would enhance it's taste, right?
Thing is, Average Joe wouldn't be the one asking your brother about bulls and steers, or your wife about watermellons. They would ask ChatGPT. And 99% it would be all fine, until some orphanage cook would try to improve some dish.
I can live with bullshit coming from ChatGPT, but the bullshit coming from tech companies about how the earth-shattering capabilities of their latest LLM offerings justify their valuations is tougher to swallow.
Frankfurt’s book “On Bullshit” explains that lies and bullshit are very distinct:
- Both people who are lying and people who are telling the truth are focused on the truth. The liar wants to steer people away from discovering the truth and the person telling the truth wants to present the truth.
- A person who communicates bullshit is not interested in whether what they say is true or false, only in its suitability for their purpose.
By this definition, the article seems well reasoned
> We argue that these falsehoods, and the overall activity of large language models, is better understood as bullshit in the sense explored by Frankfurt (On Bullshit, Princeton, 2005): the models are in an important way indifferent to the truth of their outputs
Well, if producing a working, custom disposable web application doing exactly what I need to do in half an hour for data processing is bullshit, I don’t know what isn’t.
The point about using the term bullshit (as used by Frankfurt [1]) is it implies indifference to truth.
So the paper is pointing out that, rather than thinking of these models as being mostly truthful, but which sometimes hallucinate, it's more accurate to think of them as bullshit all the time.
It doesn't mean that the bullshit is wrong (although it could be). The model itself doesn't care.
I think a weakness of the paper is it doesn't give a simple definition of what it mean by bullshit. It says "bullshitting, in the Frankfurtian sense" which leaves you to perhaps read his book which can probably be interpreted various ways.
If you look at the wikipedia page for On Bullshit it has:
>Frankfurt determines that bullshit is speech intended to persuade without regard for truth.
Based on that ChatGPT isn't really bullshit. It doesn't really have intent to persuade. It's more like a souped up search engine that takes in a bunch of documents and tries to process the data in them in a way which matches your query. The truth or not out is largely determined by the truth or not of the data going in. And that seems rather good with ChatGPT.
The add glue to Pizza stuff was a different LLM with different input data. It had read a reddit thread where someone had jokingly suggested that. A challenge with these models is to filter bullshit out of the input data though that is an issue with humans too.
This just seems like a longer version of the HN comments I see in every single AI related post. No such thing as a hallucination, LLMs don't really "understand" therefore they can neither say truths nor lies just like autocorrect can't be honest or dishonest, it "just" chooses a token from a probability distribution, etc.
I see this every day. I'm pretty sure most of us who have even a slight interest in AI know the gist of how LLMs work. I'm not sure about what difference it makes in practice.
As someone who uses AI, as a tool, every single day, I call bullshit.
If you understand limitations of current AI, it is an exceptionally useful tool. Just have realistic expectations and understand how to use the tool for what it has to offer.
If you expect the tool to have all of the answers and get everything 100% right, then you should not be using it at all. But don't preach your beliefs to the rest of us who are able to get tremendous value out of it.
ChatGPT is an idea generator. As a user, it's your job to validate and test its generated hypotheses. If you altogether dismiss ChatGPT by calling it bullshit, it means you are cognitively too lazy to test any hypotheses. This doesn't excuse ChatGPT, but it gives it some breathing room to potentially get better with improved models as time progresses.
The intellectual meat of the article seemed to be that the author found a connection between Harry Frankfurt's definition of [bullshit](https://www2.csudh.edu/ccauthen/576f12/frankfurt__harry_-_on...) and what we've been calling "hallucinations". Seemed reasonable but it wasn't exactly evident to me why I should care. We're replacing one word with another word. so what.
the best the authors could do was argue that there were some mysterious set of negative social consequences that we'd get from using "hallucinations," since it implies that the models are mistaken or misguided in an attempt at representing the truth, rather than attend to the fact that all the models are doing is generating text that _seems_ truthy, and has no awareness or attention to what the truth _actually_ is. One could probably just read the last paragraph and get all they needed to from the paper:
>We object to the term hallucination because it carries certain misleading implications ...
>
>Calling chatbot inaccuracies ‘hallucinations’ feeds in to overblown hype about their abilities among technology cheerleaders, and could lead to unnecessary consternation among the general public. It also suggests solutions to the inaccuracy problems which might not work, and could lead to misguided efforts at AI alignment amongst specialists. It can also lead to the wrong attitude towards the machine when it gets things right: the inaccuracies show that it is bullshitting, even when it’s right. Calling these inaccuracies ‘bullshit’ rather than ‘hallucinations’ isn’t just more accurate (as we’ve argued); it’s good science and technology communication in an area that sorely needs it.
So their argument is effectively:
1. it's wrong, or at least, Frankfurt's definition of "bullshit" fits better.
2. it could mislead the public or alignment researchers
On 1), I'm willing to concede that hallucinations might be the wrong term. But words have a life of their own, and it's too late to go back now. At least to late for one paper to change anything.
On 2), it seems plausible, but, regrettably, the paper spends less than a paragraph talking about it! None of the claims they're making are that complicated, and yet for some reason they fail to provide even a few falsifiable hypotheses about the main implication of their argument. Ok, sure. I can see that the term "hallucination" might be misleading. But are you really going to publish a paper just so you can "Uh, actually ..." everybody and argue that we're using the wrong word? How much is it going to mislead the public? The public is always mislead - why does this particular instance matter? How can we tell? Are alignment researches really going to be misled by choice of terminology? If they are, could you suggest a mechanism? How much would they be misled? If they are misled, why does it matter? What does "misled" even mean? How do we measure it? I could go on.
I'd like to imagine that a paper warning about the implications of using incorrect terminology would go into some detail about their claim and explore how those implications might play out - this paper's publication might have more to do with its title and topic than it does any important claims or results.
This is a philosophical paper, and "bullshit" is a term of art in philosophy, coined by philosopher Harry Frankfurt (1929-2023) in his 2005 booklet On Bullshit, in which he distinguishes bullshit from lies:
> The liar cares about the truth and attempts to hide it; the bullshitter doesn't care if what they say is true or false.
This academic paper argues that ChatGPT produces bullshit in said technical sense. It is nothing inflammatory, superficial, derogatory, dismissive, or obscene.
38 comments
[ 3.4 ms ] story [ 94.1 ms ] threadOr write papers and their code from scratch.
https://github.com/Technion-Kishony-lab/data-to-paper
Related HN discussion of the annoucement 3 days ago: https://news.ycombinator.com/item?id=40674835
I think this is a little overly optimistic about “dorks”, but I will also note the shift in the ratio of dorks to business people.
I’ve recently seen backlash against the term “tech bro”. Bill Burr recently called it out as a bullshit title to separate the bad actors from oneself, but I’ve always held it identifies a class of interest in the tech sector.
University fraternities get large sums of money from alumni and as such are financially motivated to put their pledges in degrees of that result in high paying positions.
In the 70s this was law, the 80s it was finance, 90s it was medicine, 00s it became tech.
6 figure entry level position for someone without any knowledge of an if statement were du jour for this group of people and so they flooded in. Bringing their ideals and sociopathy with them.
Woz is a dork and the brains that built apple, but his narrow interests allowed the business end to make the globally affecting decisions and we are still suffering the consequences.
My brother in law who is extremely smart, and very well trained -- MIT PhD -- and exposed to orders of magnitude more data than ChatGPT, given that the human senses ingest gigabytes of data per day, has -- like everyone else -- persistent errors in his thinking. I single him out because of his credentials, but everyone has these.
As a particularly funny example, we were talking about steers and bulls and he was like "Oh, right, I remember, some male cattle are born bulls and some are born steers". My wife, who is similarly educated, believed for a very long time that watermelon's grew on trees. At my undergrad school, a woman who was getting her physics degree and then who eventually went on to get a PhD, fundamentally believed that yeast would not rise if you talked around it -- I guess her parents had told her that and she never questioned it.
That is to say, humans lack full knowledge. We're trained on significantly more data than ChatGPT et al. We still have persistent inaccuracies despite all of that. We're not perfect.
That doesn't make us less useful. Half the time we just rely on other humans trained on similar, but slightly different data, to iron those out, and indeed AI systems that employ those methods get better accuracy.
Now again, my brother in law has fewer errors than most people since he is very smart. However, the vast majority of people are really not that bright (rumor has it that fifty percent have below average intelligence), and yet, simply through verification with other humans, they are able to accomplish useful things.
Sometimes even very capable people have persistent thinking errors. This doesn't make them less smart. No one can claim perfect intelligence. That's the only thing that's bullshit.
[1] https://www.businessinsider.com/google-researchers-openai-ch...
The mistake is thinking that 'data' means only text.
I don't think that's true anymore. GPT-4 is rumored to have been trained on a dataset with >1 trillion tokens. An average human would take a few thousand years to read through that dataset once.
Humans ingest about 74GB of data per day: https://kids.frontiersin.org/articles/10.3389/frym.2017.0002...
That means for a human baby to be as smart as GPT they would have to go from zero to speaking English within two weeks of birth.
Your error is mistaking training data for text only. Humans are particularly inefficient with text because it takes a lot of visual data to get to the thought of a particular word or character due to how human senses work (we have no 'token' sense).
At the end of the day, you wouldn't be bothered if Average Joe would add arsenic to your favourite pie, because AI told it would enhance it's taste, right?
Thing is, Average Joe wouldn't be the one asking your brother about bulls and steers, or your wife about watermellons. They would ask ChatGPT. And 99% it would be all fine, until some orphanage cook would try to improve some dish.
Because human intelligence is not so any system that models it would also not be.
- Both people who are lying and people who are telling the truth are focused on the truth. The liar wants to steer people away from discovering the truth and the person telling the truth wants to present the truth.
- A person who communicates bullshit is not interested in whether what they say is true or false, only in its suitability for their purpose.
By this definition, the article seems well reasoned
> We argue that these falsehoods, and the overall activity of large language models, is better understood as bullshit in the sense explored by Frankfurt (On Bullshit, Princeton, 2005): the models are in an important way indifferent to the truth of their outputs
It’s a very long paper that reads to me like an opinion piece designed to get a reaction (and citations).
So the paper is pointing out that, rather than thinking of these models as being mostly truthful, but which sometimes hallucinate, it's more accurate to think of them as bullshit all the time.
It doesn't mean that the bullshit is wrong (although it could be). The model itself doesn't care.
[1] https://reasonandmeaning.com/2017/01/23/harry-frankfurt-on-b...
If you look at the wikipedia page for On Bullshit it has:
>Frankfurt determines that bullshit is speech intended to persuade without regard for truth.
Based on that ChatGPT isn't really bullshit. It doesn't really have intent to persuade. It's more like a souped up search engine that takes in a bunch of documents and tries to process the data in them in a way which matches your query. The truth or not out is largely determined by the truth or not of the data going in. And that seems rather good with ChatGPT.
The add glue to Pizza stuff was a different LLM with different input data. It had read a reddit thread where someone had jokingly suggested that. A challenge with these models is to filter bullshit out of the input data though that is an issue with humans too.
Discussion: https://news.ycombinator.com/item?id=40626692
I see this every day. I'm pretty sure most of us who have even a slight interest in AI know the gist of how LLMs work. I'm not sure about what difference it makes in practice.
If you understand limitations of current AI, it is an exceptionally useful tool. Just have realistic expectations and understand how to use the tool for what it has to offer.
If you expect the tool to have all of the answers and get everything 100% right, then you should not be using it at all. But don't preach your beliefs to the rest of us who are able to get tremendous value out of it.
the best the authors could do was argue that there were some mysterious set of negative social consequences that we'd get from using "hallucinations," since it implies that the models are mistaken or misguided in an attempt at representing the truth, rather than attend to the fact that all the models are doing is generating text that _seems_ truthy, and has no awareness or attention to what the truth _actually_ is. One could probably just read the last paragraph and get all they needed to from the paper:
>We object to the term hallucination because it carries certain misleading implications ... > >Calling chatbot inaccuracies ‘hallucinations’ feeds in to overblown hype about their abilities among technology cheerleaders, and could lead to unnecessary consternation among the general public. It also suggests solutions to the inaccuracy problems which might not work, and could lead to misguided efforts at AI alignment amongst specialists. It can also lead to the wrong attitude towards the machine when it gets things right: the inaccuracies show that it is bullshitting, even when it’s right. Calling these inaccuracies ‘bullshit’ rather than ‘hallucinations’ isn’t just more accurate (as we’ve argued); it’s good science and technology communication in an area that sorely needs it.
So their argument is effectively: 1. it's wrong, or at least, Frankfurt's definition of "bullshit" fits better. 2. it could mislead the public or alignment researchers
On 1), I'm willing to concede that hallucinations might be the wrong term. But words have a life of their own, and it's too late to go back now. At least to late for one paper to change anything.
On 2), it seems plausible, but, regrettably, the paper spends less than a paragraph talking about it! None of the claims they're making are that complicated, and yet for some reason they fail to provide even a few falsifiable hypotheses about the main implication of their argument. Ok, sure. I can see that the term "hallucination" might be misleading. But are you really going to publish a paper just so you can "Uh, actually ..." everybody and argue that we're using the wrong word? How much is it going to mislead the public? The public is always mislead - why does this particular instance matter? How can we tell? Are alignment researches really going to be misled by choice of terminology? If they are, could you suggest a mechanism? How much would they be misled? If they are misled, why does it matter? What does "misled" even mean? How do we measure it? I could go on.
I'd like to imagine that a paper warning about the implications of using incorrect terminology would go into some detail about their claim and explore how those implications might play out - this paper's publication might have more to do with its title and topic than it does any important claims or results.
This is a philosophical paper, and "bullshit" is a term of art in philosophy, coined by philosopher Harry Frankfurt (1929-2023) in his 2005 booklet On Bullshit, in which he distinguishes bullshit from lies:
> The liar cares about the truth and attempts to hide it; the bullshitter doesn't care if what they say is true or false.
This academic paper argues that ChatGPT produces bullshit in said technical sense. It is nothing inflammatory, superficial, derogatory, dismissive, or obscene.
Again, why is it flagged?