4 ms·

Humans get things wrong too. A better question is: what error rate is acceptable for this task? Jobs where higher error rates are acceptable, or where errors a

Humans get things wrong too. A better question is: what error rate is acceptable for this task? Jobs where higher error rates are acceptable, or where errors are easier to detect, will succumb to automation first. Art and poetry fit both of these criteria. The claim is that as the model and training data sizes increase, these errors will get more and more rare. We will see... I am very optimistic about the far future. However, there will be a transition period where some jobs have been automated away but not others. There will be massive inequality between the remaining knowledge workers and manual laborers. If I was in a role on the early automation side of the spectrum then I would be retraining ASAP.

worrycue 4y ago
Humans can self correct / think critically. AIs like ChatGPT can’t do that at all. You know sometimes you have a “bright idea” then after thinking about it for a second you realise it’s nonsense. With AI like ChatGPT, the “thinking about it for a second” part never happens.
- zamnos 4y ago
  There are logs where ChatGPT initially gives the wrong answer, but then corrects itself when asked to explain the wrong answer. Is that the second part you're thinking of?
  - danparsonson 4y ago
    The crucial difference there is the presence of an external agent intelligent enough to spot that the answer is wrong; humans can do that for themselves. ChatGPT doesn't self-reflect.
    - drusepth 4y ago
      Interestingly, many (most?) humans don't self-reflect or correct themselves unless challenged by an external agent as well — which doesn't necessarily have to be another human. Also of note, GPT-4 seems to show huge improvements so far over GPT-3 when it comes to "thinking out loud" to come to a (better) answer to more complex problems. Kind of a front-loaded reflection of correctness for an overall goal before diving into the implementation weeds — something that definitely helps me (as a human) avoid unnecessary mistakes in the first place.
      - danparsonson 4y ago
        > Interestingly, many (most?) humans don't self-reflect or correct themselves unless challenged by an external agent as well Disagree with you here - why do you say this? Maybe we don't apply self-reflection consistently (for example when it comes to political beliefs) but even toddlers know when they haven't achieved the goal they were aiming for. ChatGPT has no clue unless you prod it, because it doesn't know anything - it's stringing words together using probability.
  - impalallama 4y ago
    ChatGPT will do the exact same thing with correct answers.

worrycue 4y ago

Humans can self correct / think critically. AIs like ChatGPT can’t do that at all. You know sometimes you have a “bright idea” then after thinking about it for a second you realise it’s nonsense. With AI like ChatGPT, the “thinking about it for a second” part never happens.

zamnos 4y ago
There are logs where ChatGPT initially gives the wrong answer, but then corrects itself when asked to explain the wrong answer. Is that the second part you're thinking of?
- danparsonson 4y ago
  The crucial difference there is the presence of an external agent intelligent enough to spot that the answer is wrong; humans can do that for themselves. ChatGPT doesn't self-reflect.
  - drusepth 4y ago
    Interestingly, many (most?) humans don't self-reflect or correct themselves unless challenged by an external agent as well — which doesn't necessarily have to be another human. Also of note, GPT-4 seems to show huge improvements so far over GPT-3 when it comes to "thinking out loud" to come to a (better) answer to more complex problems. Kind of a front-loaded reflection of correctness for an overall goal before diving into the implementation weeds — something that definitely helps me (as a human) avoid unnecessary mistakes in the first place.
    - danparsonson 4y ago
      > Interestingly, many (most?) humans don't self-reflect or correct themselves unless challenged by an external agent as well Disagree with you here - why do you say this? Maybe we don't apply self-reflection consistently (for example when it comes to political beliefs) but even toddlers know when they haven't achieved the goal they were aiming for. ChatGPT has no clue unless you prod it, because it doesn't know anything - it's stringing words together using probability.
- impalallama 4y ago
  ChatGPT will do the exact same thing with correct answers.