4 ms·

> “As I see it, ‘intelligence’ needs something like the ability to know or understand some area,” he said. “If something can’t actually understand things we sho

> “As I see it, ‘intelligence’ needs something like the ability to know or understand some area,” he said. “If something can’t actually understand things we shouldn’t say it’s intelligent, not even a little intelligent, but people are using the term artificial intelligence for bullshit generators.” I wouldn't go as far in blaming AI. It is certainly error prone and has very short autonomy. It's not better than subject matter experts in any field. But it can solve multi-step problems even when they don't look very close to anything in the training set. I thought that as a pioneer in the free software movement, he would fully support LLMs that democratize skills and make them universally applicable. LLMs benefit from the knowledge of some and extend that help to others, much like the principles of free software. In fact AI is much more radical in making skills free than software ever was.

varelse 3y ago
[dead]
vidarh 3y ago
His thinking here leans on assuming that there's a qualitative difference between LLM's "understanding" and our "understanding" that justifies using the term for us, but not for LLMs. Maybe there is, but their reasoning ability is sufficient that I don't think is at all a given, and at a minimum an assertion that should make you ask someone how they define "understanding", because defining that in a way that is focused on reasoning-ability that will exclude LLMs but not exclude some people seems tricky.
- ad404b8a372f2b9 3y ago
  I don't know that his argument is based on essentialism since he goes on to describe algorithms which he considers artificially intelligent: "There are programs that can look at a photo of some magnified cells and tell you, with greater likelihood of being right than any human doctor, whether it’s cancerous or not,”. And he puts it in contrast to his previous statement about LLMs not being artifically intelligent. It seems to me he wants people to look past the impressive facade of correct grammar and see that a lot of the LLMs often spew out bullshit. Personally I've seen them perform intelligence tasks, generate correct code and so forth. But I think many people ascribe to them a general intelligence that they don't have, which is why you see them quoted increasingly often online as if they were knowledgeable or correct-by-default.
  - vidarh 3y ago
    I think he's seriously contradicting himself there in that the algorithms he argue are AI certainly do not meet his stated standard, and it seems like he's drawing a line where he refuses to use the term about anything that might be seen as intelligent that doesn't meet his standard for intelligence, but is fine with using it for techniques where there's no chance of confusion. I think if he wasn't making an essentialist argument about LLMs he worded himself exceedingly poorly. Arguing against the danger of blindly believing LLMs is good, but at the same time he's destroying his own credibility on the subject by exaggerating how bad they are vs. focusing on the dangers of blind trust. This is especially notable when he then goes on to point to measurable success as a metric for other AI tech - he'd have made a far better point if he argued that LLMs also need to be evaluated carefully on a case by case basis like these other systems, rather than be taken on trust.
    - simiones 3y ago
      > This is especially notable when he then goes on to point to measurable success as a metric for other AI tech - he'd have made a far better point if he argued that LLMs also need to be evaluated carefully on a case by case basis like these other systems, rather than be taken on trust. The examples he gives are technologies where you can point the AI to a problem and trust the results will work as intended. You don't need to check every individual output of those AIs, you can just trust that it will be better than the output you'd get from a human expert (even though it's of course not 100%, and a combination of AI + human expert may be better). What is the equivalent for LLMs?
      - vidarh 3y ago
        > The examples he gives are technologies where you can point the AI to a problem and trust the results will work as intended. No, they're technologies where AI's have been tuned and tested and validated for there domain so that we after that process know that they will work as intended sufficiently often to measurably be a net benefit, just as you yourself go on to describe. They'll still make mistakes, just like you yourself point out. But they've been tested to ensure they make few enough mistakes to be worthwhile. > What is the equivalent for LLMs? The equivalent for LLMs is exactly the same process. To quote myself: > he'd have made a far better point if he argued that LLMs also need to be evaluated carefully on a case by case basis like these other systems, rather than be taken on trust. In other words: Test them on your use case, and validate their performance on that use case. Don't assume. We wouldn't take the performance of any domain specific model on trust, and there's no more reason to take the performance of an LLM on trust for a domain we've not tested it thoroughly for, and quite possibly fine tuned it for. The difference is that LLMs do well enough to convince some people of the idea they can skip the testing and validation step, and that point - that people are prone to give them a level of trust that they should not be given - is valid. But extending that to dismissing them as "bullshit generators" as he did is equally ridiculous.
- n4r9 3y ago
  > assuming that there's a qualitative difference between LLM's "understanding" and our "understanding" that justifies using the term for us, but not for LLMs If there is no qualitative difference, then human brains are no more than machines that predict the next word at each step. This could be the case, but instinctively I feel that it's not. You're right that it's very tricky - there's some interesting approaches, such as defining classes of problems or riddles that humans find doable but LLMs do not. The below is potentially a good candidate. Ultimately, I think a decisive test will be to see whether LLMs can take on great unsolved mathematical conjectures. ------------------------------ User There are exactly three people in a room: Andrew, Bob, and Chris. Each of them is engaged in an activity. Andrew is reading a book by himself. Bob is playing table-tennis with someone else. What is Chris doing ChatGPT Given the information provided: Andrew is reading a book by himself. Bob is playing table-tennis with someone else. We can infer that Chris must be doing something as well, but the information given does not specify what Chris is doing. Therefore, we cannot determine Chris's activity based solely on the information provided.
  - Turing_Machine 3y ago
    I just tried this with ChatGPT 4. Response: Chris is playing table-tennis with Bob. Edit: ChatGPT 3.5 gives me the response you reported.
    - n4r9 3y ago
      Interesting, thanks! What is ChatGPT4's response to this, if you don't mind? ---------------- Imagine that I take a road trip starting from Amsterdam, visit the cities Warsaw, Budapest, Brussels, Paris, and Berlin, and then return to Amsterdam. In what sequence should I visit those cities in order to minimise the total driving time? ----------------
      - Turing_Machine 3y ago
        To minimize total driving time, a logical sequence would be: Amsterdam → Brussels → Paris → Berlin → Warsaw → Budapest → Amsterdam. This route forms a relatively compact loop, minimizing backtracking and excessive driving distances. However, for the most accurate and updated information, it's advisable to use a GPS or online mapping service like Google Maps to plan your route.
thesuperbigfrog 3y ago
>> I thought that as a pioneer in the free software movement, he would fully support LLMs that democratize skills and make them universally applicable. RMS worked on early AI research at the MIT Artificial Intelligence lab in the 1970s. LLMs are fundamentally different than many of those early AI systems in both goals and how they work. It is likely that he views them through a different lens and sees their shortcomings in sharp relief compared to the early AI systems he worked on.
boomboomsubban 3y ago
His objection seems predominantly pedantic, and bemoans that "antisocial media" is profiting off them. That is incredibly in character for Stallman, he doesn't seem opposed to them as a whole.
xctr94 3y ago
I normally would argue for how much we dumb down human intelligence to pretend AI is intelligent (rather than saying it’s capable). But… leaving that aside, I think it’s on-brand for Stallman to push a bit against them. They’re a) largely owned by private companies; b) their code, data, and models isn’t open or openly licensed. I wouldn’t expect one of GPL’s biggest crusader to defend ML in its current state.
dncornholio 3y ago
He is not blaming the tech.. He is blaming people that believe the output it intelligent. LLM's are marketed as know-all machines while in reality, they don't 'know' anything.

varelse 3y ago

[dead]

vidarh 3y ago

His thinking here leans on assuming that there's a qualitative difference between LLM's "understanding" and our "understanding" that justifies using the term for us, but not for LLMs. Maybe there is, but their reasoning ability is sufficient that I don't think is at all a given, and at a minimum an assertion that should make you ask someone how they define "understanding", because defining that in a way that is focused on reasoning-ability that will exclude LLMs but not exclude some people seems tricky.

ad404b8a372f2b9 3y ago
I don't know that his argument is based on essentialism since he goes on to describe algorithms which he considers artificially intelligent: "There are programs that can look at a photo of some magnified cells and tell you, with greater likelihood of being right than any human doctor, whether it’s cancerous or not,”. And he puts it in contrast to his previous statement about LLMs not being artifically intelligent. It seems to me he wants people to look past the impressive facade of correct grammar and see that a lot of the LLMs often spew out bullshit. Personally I've seen them perform intelligence tasks, generate correct code and so forth. But I think many people ascribe to them a general intelligence that they don't have, which is why you see them quoted increasingly often online as if they were knowledgeable or correct-by-default.
- vidarh 3y ago
  I think he's seriously contradicting himself there in that the algorithms he argue are AI certainly do not meet his stated standard, and it seems like he's drawing a line where he refuses to use the term about anything that might be seen as intelligent that doesn't meet his standard for intelligence, but is fine with using it for techniques where there's no chance of confusion. I think if he wasn't making an essentialist argument about LLMs he worded himself exceedingly poorly. Arguing against the danger of blindly believing LLMs is good, but at the same time he's destroying his own credibility on the subject by exaggerating how bad they are vs. focusing on the dangers of blind trust. This is especially notable when he then goes on to point to measurable success as a metric for other AI tech - he'd have made a far better point if he argued that LLMs also need to be evaluated carefully on a case by case basis like these other systems, rather than be taken on trust.
  - simiones 3y ago
    > This is especially notable when he then goes on to point to measurable success as a metric for other AI tech - he'd have made a far better point if he argued that LLMs also need to be evaluated carefully on a case by case basis like these other systems, rather than be taken on trust. The examples he gives are technologies where you can point the AI to a problem and trust the results will work as intended. You don't need to check every individual output of those AIs, you can just trust that it will be better than the output you'd get from a human expert (even though it's of course not 100%, and a combination of AI + human expert may be better). What is the equivalent for LLMs?
    - vidarh 3y ago
      > The examples he gives are technologies where you can point the AI to a problem and trust the results will work as intended. No, they're technologies where AI's have been tuned and tested and validated for there domain so that we after that process know that they will work as intended sufficiently often to measurably be a net benefit, just as you yourself go on to describe. They'll still make mistakes, just like you yourself point out. But they've been tested to ensure they make few enough mistakes to be worthwhile. > What is the equivalent for LLMs? The equivalent for LLMs is exactly the same process. To quote myself: > he'd have made a far better point if he argued that LLMs also need to be evaluated carefully on a case by case basis like these other systems, rather than be taken on trust. In other words: Test them on your use case, and validate their performance on that use case. Don't assume. We wouldn't take the performance of any domain specific model on trust, and there's no more reason to take the performance of an LLM on trust for a domain we've not tested it thoroughly for, and quite possibly fine tuned it for. The difference is that LLMs do well enough to convince some people of the idea they can skip the testing and validation step, and that point - that people are prone to give them a level of trust that they should not be given - is valid. But extending that to dismissing them as "bullshit generators" as he did is equally ridiculous.
n4r9 3y ago
> assuming that there's a qualitative difference between LLM's "understanding" and our "understanding" that justifies using the term for us, but not for LLMs If there is no qualitative difference, then human brains are no more than machines that predict the next word at each step. This could be the case, but instinctively I feel that it's not. You're right that it's very tricky - there's some interesting approaches, such as defining classes of problems or riddles that humans find doable but LLMs do not. The below is potentially a good candidate. Ultimately, I think a decisive test will be to see whether LLMs can take on great unsolved mathematical conjectures. ------------------------------ User There are exactly three people in a room: Andrew, Bob, and Chris. Each of them is engaged in an activity. Andrew is reading a book by himself. Bob is playing table-tennis with someone else. What is Chris doing ChatGPT Given the information provided: Andrew is reading a book by himself. Bob is playing table-tennis with someone else. We can infer that Chris must be doing something as well, but the information given does not specify what Chris is doing. Therefore, we cannot determine Chris's activity based solely on the information provided.
- Turing_Machine 3y ago
  I just tried this with ChatGPT 4. Response: Chris is playing table-tennis with Bob. Edit: ChatGPT 3.5 gives me the response you reported.
  - n4r9 3y ago
    Interesting, thanks! What is ChatGPT4's response to this, if you don't mind? ---------------- Imagine that I take a road trip starting from Amsterdam, visit the cities Warsaw, Budapest, Brussels, Paris, and Berlin, and then return to Amsterdam. In what sequence should I visit those cities in order to minimise the total driving time? ----------------
    - Turing_Machine 3y ago
      To minimize total driving time, a logical sequence would be: Amsterdam → Brussels → Paris → Berlin → Warsaw → Budapest → Amsterdam. This route forms a relatively compact loop, minimizing backtracking and excessive driving distances. However, for the most accurate and updated information, it's advisable to use a GPS or online mapping service like Google Maps to plan your route.

thesuperbigfrog 3y ago

>> I thought that as a pioneer in the free software movement, he would fully support LLMs that democratize skills and make them universally applicable. RMS worked on early AI research at the MIT Artificial Intelligence lab in the 1970s. LLMs are fundamentally different than many of those early AI systems in both goals and how they work. It is likely that he views them through a different lens and sees their shortcomings in sharp relief compared to the early AI systems he worked on.

boomboomsubban 3y ago

His objection seems predominantly pedantic, and bemoans that "antisocial media" is profiting off them. That is incredibly in character for Stallman, he doesn't seem opposed to them as a whole.

xctr94 3y ago

I normally would argue for how much we dumb down human intelligence to pretend AI is intelligent (rather than saying it’s capable). But… leaving that aside, I think it’s on-brand for Stallman to push a bit against them. They’re a) largely owned by private companies; b) their code, data, and models isn’t open or openly licensed. I wouldn’t expect one of GPL’s biggest crusader to defend ML in its current state.

dncornholio 3y ago

He is not blaming the tech.. He is blaming people that believe the output it intelligent. LLM's are marketed as know-all machines while in reality, they don't 'know' anything.