5 ms·

Yeah, I'm noticing the same thing. I have a lot of questions that could be Stackoverflow questions. But to do that, I'd have to make sure there are no duplicat

by _ea1k 3y ago

Yeah, I'm noticing the same thing. I have a lot of questions that could be Stackoverflow questions. But to do that, I'd have to make sure there are no duplicates and frame the question with every appropriate detail. Or, I could just have a quick conv with the AI. Sure, it'll get some things wrong, but so will SO answers. :) And it is amazing at how clarifying it can be. Even the wrong answers are sometimes a clue that the questions themselves are a little off.

menacingly 3y ago
I'm trying to think of a worse outcome for a QA site than the anxiety a lot of people need to overcome to post to SO. I'd like numbers on what % of questions are scolded for having been asked
- KMag 3y ago
  I'd also love to see the % of scoldings that miss an important subtle difference in the questions being asked.
- sensanaty 3y ago
  I've literally never had a SO question deleted/downvoted/any of the other stuff. Just think of a genuinely good question and show you've done a modicum of research and people will happily reply to it
  - seanthemon 3y ago
    Marked as duplicate, use the search next time won't you??
  - bobsmooth 3y ago
    How can you do research if you don't know what you don't know?
    - sensanaty 3y ago
      How do you expect anyone to help you if you don't know what to ask? You have to at least have a starting point, and that's usually more than enough on SO. stackoverflow.com/users/8610114/sensanaty Seriously just check out questions I've asked on SO. They're nothing groundbreaking, a lot of them are pretty beginner/basic questions, yet I've never really had adverse reactions to any one of my questions. Some go unanswered sure, and I might have one or two marked as dupes (they are more or less) but if you've ever browser the "new" portion of the site it's inundated with hudreds of shitty, low effort questions, so it's understandable. SO is also an excellent jumping off point for figuring it out yourself. I've had dozens of times where I figured out a solution midway through typing a question, since you have to make good quality questions there in the first place.
  - menacingly 3y ago
    I love the profoundly flawed "you should have searched" logic. Nearly everyone who got there did so because a google search failed. The first step in this process _is_ a search. You're being whined at for not knowing the very specific incantation which will summon the results. Which, if you knew, you likely wouldn't need to ask anyway. "You're stupid and your question is stupid. I forbid answers. I will keep it up and nofollow all of the scalped resources it links to in the hopes others can come here and benefit from this lack of answer"
  - _ea1k 3y ago
    I agree with you, but my first step in doing a modicum of research on a new subject is to ask questions. I can ask the most basic questions to chatgpt and usually get pointed in the right direction. Which is my point. ChatGPT is the most basic modicum of research and it is excellent for that. But it also can do a lot more.
- sublinear 3y ago
  The next iteration of chat bot can't possibly avoid being this way either. Collaboration on writing a better question is the only way its answers could get to the next level and customized for your specific use case. I'm imagining it spinning up a generated SO clone for you to browse and scolding you just the same (but at least giving you a decent link), or at least asking why an answer you found isn't good enough. The way collaborating with AI works now is too wordy and error prone. You don't know what you don't know. Chat bots still don't talk back much and that's a problem. When we get to this level of AI, maybe some people will finally understand that getting your expectations broken is not a form of hostile communication.
- _ea1k 3y ago
  I remember arguing about this and boat programming.
blibble 3y ago
this works for the now as there was plenty of data to parasitise, sorry, train from technology moves on in 2030, when stackoverflow have folded due to lack of visitors, and all new data on new libraries/APIs are locked up beyond OpenAI's reach how useful is the AI going to be? the whole thing sows the seeds of its own destruction
- JieJie 3y ago
  I hear what you are saying. What I think that what you're missing is that subsequent conversations with AI, where humans and AI solve problems together, is what the next generation of AI is going to be trained on. That data is being collected every time someone interacts with a chatbot, which is why they are being provided — so many of them - as a free service. It is, as before: if the service is free, you are the product. That being said, I think that to the extent AI is an actually helpful product, as opposed to social media, public forums, or free video hosting, the tradeoff is way more worthwhile. In my most humble opinion. I'm sure your calculus will lead to different outcomes.
  - blibble 3y ago
    only in very rare cases will the data from the interaction actually be useful for training, simply because the final version is rarely supplied to it yes, it's asked to produce something, but it won't see the result after substantial effort required to actually make it work the interaction is not going to be nearly enough to produce useful training data to replace e.g. stackoverflow
  - visarga 3y ago
    > if the service is free, you are the product and if GPTBot can reach your site, you are the product, they would make you their product even if you don't use their free service but now they are not allowed to complain when others train models on GPT-4 data, they can become someone else's product too - see Orca, TinyStories and Phi-1.5
- jncfhnb 3y ago
  Probably still very useful

by _ea1k 3y ago

menacingly 3y ago

I'm trying to think of a worse outcome for a QA site than the anxiety a lot of people need to overcome to post to SO. I'd like numbers on what % of questions are scolded for having been asked

KMag 3y ago
I'd also love to see the % of scoldings that miss an important subtle difference in the questions being asked.
sensanaty 3y ago
I've literally never had a SO question deleted/downvoted/any of the other stuff. Just think of a genuinely good question and show you've done a modicum of research and people will happily reply to it
- seanthemon 3y ago
  Marked as duplicate, use the search next time won't you??
- bobsmooth 3y ago
  How can you do research if you don't know what you don't know?
  - sensanaty 3y ago
    How do you expect anyone to help you if you don't know what to ask? You have to at least have a starting point, and that's usually more than enough on SO. stackoverflow.com/users/8610114/sensanaty Seriously just check out questions I've asked on SO. They're nothing groundbreaking, a lot of them are pretty beginner/basic questions, yet I've never really had adverse reactions to any one of my questions. Some go unanswered sure, and I might have one or two marked as dupes (they are more or less) but if you've ever browser the "new" portion of the site it's inundated with hudreds of shitty, low effort questions, so it's understandable. SO is also an excellent jumping off point for figuring it out yourself. I've had dozens of times where I figured out a solution midway through typing a question, since you have to make good quality questions there in the first place.
- menacingly 3y ago
  I love the profoundly flawed "you should have searched" logic. Nearly everyone who got there did so because a google search failed. The first step in this process _is_ a search. You're being whined at for not knowing the very specific incantation which will summon the results. Which, if you knew, you likely wouldn't need to ask anyway. "You're stupid and your question is stupid. I forbid answers. I will keep it up and nofollow all of the scalped resources it links to in the hopes others can come here and benefit from this lack of answer"
- _ea1k 3y ago
  I agree with you, but my first step in doing a modicum of research on a new subject is to ask questions. I can ask the most basic questions to chatgpt and usually get pointed in the right direction. Which is my point. ChatGPT is the most basic modicum of research and it is excellent for that. But it also can do a lot more.
sublinear 3y ago
The next iteration of chat bot can't possibly avoid being this way either. Collaboration on writing a better question is the only way its answers could get to the next level and customized for your specific use case. I'm imagining it spinning up a generated SO clone for you to browse and scolding you just the same (but at least giving you a decent link), or at least asking why an answer you found isn't good enough. The way collaborating with AI works now is too wordy and error prone. You don't know what you don't know. Chat bots still don't talk back much and that's a problem. When we get to this level of AI, maybe some people will finally understand that getting your expectations broken is not a form of hostile communication.
_ea1k 3y ago
I remember arguing about this and boat programming.

blibble 3y ago

this works for the now as there was plenty of data to parasitise, sorry, train from technology moves on in 2030, when stackoverflow have folded due to lack of visitors, and all new data on new libraries/APIs are locked up beyond OpenAI's reach how useful is the AI going to be? the whole thing sows the seeds of its own destruction

JieJie 3y ago
I hear what you are saying. What I think that what you're missing is that subsequent conversations with AI, where humans and AI solve problems together, is what the next generation of AI is going to be trained on. That data is being collected every time someone interacts with a chatbot, which is why they are being provided — so many of them - as a free service. It is, as before: if the service is free, you are the product. That being said, I think that to the extent AI is an actually helpful product, as opposed to social media, public forums, or free video hosting, the tradeoff is way more worthwhile. In my most humble opinion. I'm sure your calculus will lead to different outcomes.
- blibble 3y ago
  only in very rare cases will the data from the interaction actually be useful for training, simply because the final version is rarely supplied to it yes, it's asked to produce something, but it won't see the result after substantial effort required to actually make it work the interaction is not going to be nearly enough to produce useful training data to replace e.g. stackoverflow
- visarga 3y ago
  > if the service is free, you are the product and if GPTBot can reach your site, you are the product, they would make you their product even if you don't use their free service but now they are not allowed to complain when others train models on GPT-4 data, they can become someone else's product too - see Orca, TinyStories and Phi-1.5
jncfhnb 3y ago
Probably still very useful