17 ms·

They may not need direct subscription revenue to recoup their investment. A variant of multi-modal LLMs may be the solution to self-driving cars, home robotics

They may not need direct subscription revenue to recoup their investment. A variant of multi-modal LLMs may be the solution to self-driving cars, home robotics, and more. I keep saying that to be a really effective driver, an AI model will need a theory of mind, which the larger LLMs appear to have. Similarly, any such model will need to be able to do OCR and read arbitrary street signs, and understand what the sign meant. Most modern LLMs can already do this.

steve_adams_86 2y ago
Since when do LLMS appear to possess theory of mind? The papers I've read on this show impressive capabilities, but only within conditions where language-based training would strongly indicate appropriate outputs for the LLM to use. Outside of familiar conditions, or even realistic conditions, they don't do well. It's all still statistical string-building.
- sebzim4500 2y ago
  Can you give an example prompt that shows a lack of theory of mind? It used to be easy, but now that models have improved I've noticed that the 'LLM doubters' nowadays are forced to make generic, almost spiritual, claims rather than making concrete statements.
  - emptysongglass 2y ago
    The burden isn't on us, it's on people making claims that LLMs do possess theory of mind. You're the one who has to prove it, not us.
    - sebzim4500 2y ago
      If you use any of the conventional tests that exist of theory of mind (most famously the Sally-Anne Test [1] but also the others) then SOTA reasoning models will get near 100%. Even if you try to come up with similar questions which you expect not to be in the training set they will still get them right. In the absence of any evidence to the contrary, this is convincing evidence in my opinion. [1] https://en.wikipedia.org/wiki/Sally%E2%80%93Anne_test https://en.wikipedia.org/wiki/Sally%E2%80%93Anne_test
      - esafak 2y ago
        Maybe having a theory of mind isn't the big deal we thought it was. People are so conditioned to expect such things only from biological lifeforms, where theory of mind comes packaged with many other abilities that robots currently lack, that we reflexively dismiss the robot.
      - zipy124 2y ago
        That same source you link says that your view of 100% is not accepted as a consesus: "... GPT-4's ability to reason about the beliefs of other agents remains limited (59% accuracy on the ToMi benchmark),[15] and is not robust to "adversarial" changes to the Sally-Anne test that humans flexibly handle.[16][17] While some authors argue that the performance of GPT-4 on Sally-Anne-like tasks can be increased to 100% via improved prompting strategies,[18] this approach appears to improve accuracy to only 73% on the larger ToMi dataset."
        sebzim4500 2y ago
        We are talking about frontier models not GPT-4
        zipy124 2y ago
        Yes but I am using the same source the commenter used to backup their figure, merely saying look your source doesn't say what you claim it does. If they wanted to talk about frontier models maybe they should have cited a link to talking about frontier models performance.
        CamperBob2 2y ago
        In basically every case, by the time a claim like that is stated in a paper like that, it's obsolete by the time it's published, and ancient history by the time you use it to try to win an argument.
        zipy124 2y ago
        My point is merely if you are going to make an argument using a source, the source should support your argument. If you say "the accuracy of an llm on task 1 is 90% [1]" and when you go to [1] it says the accuracy of an llm on task 1 is 50%, but some sources say with better prompts you can get to 90%, but when extended to a larger data-set for task 1, performance drops to 70%" then just quoting the highest number is mis-leading.
    - neckardt 2y ago
      I’ll bite. Theory of mind (ToM) is the ability to understand that other people have different thoughts, feelings, and intentions than you do. So let’s run through a basic scenario with ChatGPT 4o: Prompt 1: Briefly respond: You’re driving on the highway when the driver in the lane next to you swerves into your lane. Why might this be? Response 1: The driver may be distracted, drowsy, intoxicated, experiencing a medical emergency, avoiding an obstacle, or not checking their blind spot before changing lanes. Prompt 2: What might that person be feeling or intending? Response 2: The driver might be feeling distracted, rushed, unaware, or panicked. They may intend to change lanes without realizing your presence, be reacting to something on the road, or simply have made an unintentional mistake. Prompt 3: You notice there was no obstacle. Any other reason the driver might do this? Response 3: Yes, the driver might be experiencing fatigue, intoxication, a medical issue (such as a seizure or dizziness), distraction (like using a phone or adjusting the radio), aggressive driving, or simply failing to check their blind spot properly. Seems to me that ChatGPT has a good idea of what another driver might be feeling while on the road. It gets this right for most other cases too. Could you provide a counterexample for when it fails to understand what other people may be feeling in certain scenarios?
      - Arkhaine_kupo 2y ago
        Well I think a fairly easy counterexample is that if you keep going, and the information falls outside its context window then the questions you had about the driver stop making sense. So lets say you ask about a driver who swerved, and the response by chatgpt correctly talks about his possible state. Then you talk more about cars, about road conditions etc and once the context is far away enough you ask, what was the driver doing again? And you are prompted with a hallucination. Because there is no state in the llm, there is no "driver", there is contextual statistically accurate responses but you hold a "driver" object in your mind while maintaining the conversation, the llm doesn't. Its like a conversation with someone with short term memory loss like in memento
        dauhak 2y ago
        And people with short term memory loss nevertheless have theory of mind just fine. Nothing about LLM's dropping context over big enough windows implies they don't have theory of mind, it just shows they have limitations - just like humans even with "normal" memory will lose track over a huge context window. Like there are plenty of shortcomings of LLMs but it feels like people are comparing them to some platonic ideal human when writing them off
    - farts_mckensy 2y ago
      Prove that you possess "theory of mind."
MrMan 2y ago
[dead]
HarHarVeryFunny 2y ago
You're not going to run a SOTA LLM of this size off batteries (robotics), even in a car where the alternator is charging them, nor can you afford to rely on a high-speed internet connection being available 100% in a life or death (FSD) application. I don't have so much faith in the future of current-architecure LLMs, but I do think that AGI will be needed for safe FSD and general-purpose robots that need to learn and operate in an uncontrolled environment such as a home.
- ggreer 2y ago
  A typical car alternator outputs 1.5-3kW of electricity, and EVs can output arbitrary amounts of power for electronics (though that will reduce range). That's more than enough to run purpose-built circuitry for a SOTA LLM. For a home robot, you could run the compute in the home instead of in the robot's body. I don't think AGI is needed for FSD because we already have safe FSD in the form of Waymo, and competitors aren't far behind. People forget that self-driving doesn't have to be perfect. It just has to be better than human drivers. Human drivers get sleepy, drunk, angry, and/or distracted. They can't see in 360º or react in milliseconds. Most cyclists and pedestrians prefer current FSD implementations to human drivers, as the computer-driven cars are much better about yielding, giving a wide berth, and being patient.
  - HarHarVeryFunny 2y ago
    Waymo is obviously pretty decent, but it's easy to drive 99.9% of the time. It's when there's invisible black ice on the road, or an animal runs out in front of you, or you lose visibility due to sun glare or whatever (I once had windshield wiper fluid = water flash freeze on contact) - maybe mud on a camera, or a wheel falls off your car or the one in front, etc, etc, that things get weird. Having autonomous cars that are statistically safer then the average driver is a reasonable bar to allow them on the road, but for ME to want to drive one I want it to be safer than me, and I am not a hot-headed teenager, or gaga 80-yr old, or drunken fool, and since I have AGI (Actual General Intelligence) I react pretty well to weird shit.

steve_adams_86 2y ago

Since when do LLMS appear to possess theory of mind? The papers I've read on this show impressive capabilities, but only within conditions where language-based training would strongly indicate appropriate outputs for the LLM to use. Outside of familiar conditions, or even realistic conditions, they don't do well. It's all still statistical string-building.

sebzim4500 2y ago
Can you give an example prompt that shows a lack of theory of mind? It used to be easy, but now that models have improved I've noticed that the 'LLM doubters' nowadays are forced to make generic, almost spiritual, claims rather than making concrete statements.
- emptysongglass 2y ago
  The burden isn't on us, it's on people making claims that LLMs do possess theory of mind. You're the one who has to prove it, not us.
  - sebzim4500 2y ago
    If you use any of the conventional tests that exist of theory of mind (most famously the Sally-Anne Test [1] but also the others) then SOTA reasoning models will get near 100%. Even if you try to come up with similar questions which you expect not to be in the training set they will still get them right. In the absence of any evidence to the contrary, this is convincing evidence in my opinion. [1] https://en.wikipedia.org/wiki/Sally%E2%80%93Anne_test https://en.wikipedia.org/wiki/Sally%E2%80%93Anne_test
    - esafak 2y ago
      Maybe having a theory of mind isn't the big deal we thought it was. People are so conditioned to expect such things only from biological lifeforms, where theory of mind comes packaged with many other abilities that robots currently lack, that we reflexively dismiss the robot.
    - zipy124 2y ago
      That same source you link says that your view of 100% is not accepted as a consesus: "... GPT-4's ability to reason about the beliefs of other agents remains limited (59% accuracy on the ToMi benchmark),[15] and is not robust to "adversarial" changes to the Sally-Anne test that humans flexibly handle.[16][17] While some authors argue that the performance of GPT-4 on Sally-Anne-like tasks can be increased to 100% via improved prompting strategies,[18] this approach appears to improve accuracy to only 73% on the larger ToMi dataset."
      - sebzim4500 2y ago
        We are talking about frontier models not GPT-4
        zipy124 2y ago
        Yes but I am using the same source the commenter used to backup their figure, merely saying look your source doesn't say what you claim it does. If they wanted to talk about frontier models maybe they should have cited a link to talking about frontier models performance.
      - CamperBob2 2y ago
        In basically every case, by the time a claim like that is stated in a paper like that, it's obsolete by the time it's published, and ancient history by the time you use it to try to win an argument.
        zipy124 2y ago
        My point is merely if you are going to make an argument using a source, the source should support your argument. If you say "the accuracy of an llm on task 1 is 90% [1]" and when you go to [1] it says the accuracy of an llm on task 1 is 50%, but some sources say with better prompts you can get to 90%, but when extended to a larger data-set for task 1, performance drops to 70%" then just quoting the highest number is mis-leading.
  - neckardt 2y ago
    I’ll bite. Theory of mind (ToM) is the ability to understand that other people have different thoughts, feelings, and intentions than you do. So let’s run through a basic scenario with ChatGPT 4o: Prompt 1: Briefly respond: You’re driving on the highway when the driver in the lane next to you swerves into your lane. Why might this be? Response 1: The driver may be distracted, drowsy, intoxicated, experiencing a medical emergency, avoiding an obstacle, or not checking their blind spot before changing lanes. Prompt 2: What might that person be feeling or intending? Response 2: The driver might be feeling distracted, rushed, unaware, or panicked. They may intend to change lanes without realizing your presence, be reacting to something on the road, or simply have made an unintentional mistake. Prompt 3: You notice there was no obstacle. Any other reason the driver might do this? Response 3: Yes, the driver might be experiencing fatigue, intoxication, a medical issue (such as a seizure or dizziness), distraction (like using a phone or adjusting the radio), aggressive driving, or simply failing to check their blind spot properly. Seems to me that ChatGPT has a good idea of what another driver might be feeling while on the road. It gets this right for most other cases too. Could you provide a counterexample for when it fails to understand what other people may be feeling in certain scenarios?
    - Arkhaine_kupo 2y ago
      Well I think a fairly easy counterexample is that if you keep going, and the information falls outside its context window then the questions you had about the driver stop making sense. So lets say you ask about a driver who swerved, and the response by chatgpt correctly talks about his possible state. Then you talk more about cars, about road conditions etc and once the context is far away enough you ask, what was the driver doing again? And you are prompted with a hallucination. Because there is no state in the llm, there is no "driver", there is contextual statistically accurate responses but you hold a "driver" object in your mind while maintaining the conversation, the llm doesn't. Its like a conversation with someone with short term memory loss like in memento
      - dauhak 2y ago
        And people with short term memory loss nevertheless have theory of mind just fine. Nothing about LLM's dropping context over big enough windows implies they don't have theory of mind, it just shows they have limitations - just like humans even with "normal" memory will lose track over a huge context window. Like there are plenty of shortcomings of LLMs but it feels like people are comparing them to some platonic ideal human when writing them off
  - farts_mckensy 2y ago
    Prove that you possess "theory of mind."

MrMan 2y ago

[dead]

HarHarVeryFunny 2y ago

You're not going to run a SOTA LLM of this size off batteries (robotics), even in a car where the alternator is charging them, nor can you afford to rely on a high-speed internet connection being available 100% in a life or death (FSD) application. I don't have so much faith in the future of current-architecure LLMs, but I do think that AGI will be needed for safe FSD and general-purpose robots that need to learn and operate in an uncontrolled environment such as a home.

ggreer 2y ago
A typical car alternator outputs 1.5-3kW of electricity, and EVs can output arbitrary amounts of power for electronics (though that will reduce range). That's more than enough to run purpose-built circuitry for a SOTA LLM. For a home robot, you could run the compute in the home instead of in the robot's body. I don't think AGI is needed for FSD because we already have safe FSD in the form of Waymo, and competitors aren't far behind. People forget that self-driving doesn't have to be perfect. It just has to be better than human drivers. Human drivers get sleepy, drunk, angry, and/or distracted. They can't see in 360º or react in milliseconds. Most cyclists and pedestrians prefer current FSD implementations to human drivers, as the computer-driven cars are much better about yielding, giving a wide berth, and being patient.
- HarHarVeryFunny 2y ago
  Waymo is obviously pretty decent, but it's easy to drive 99.9% of the time. It's when there's invisible black ice on the road, or an animal runs out in front of you, or you lose visibility due to sun glare or whatever (I once had windshield wiper fluid = water flash freeze on contact) - maybe mud on a camera, or a wheel falls off your car or the one in front, etc, etc, that things get weird. Having autonomous cars that are statistically safer then the average driver is a reasonable bar to allow them on the road, but for ME to want to drive one I want it to be safer than me, and I am not a hot-headed teenager, or gaga 80-yr old, or drunken fool, and since I have AGI (Actual General Intelligence) I react pretty well to weird shit.