-1

What do you think about the possibility of an AI doing web scraping of the contents of Stack Overflow, of Unix stackexchange for example and after understanding the questions applying the best answers?

Could it revolutionize system administration/other work?

3
  • "Could it revolutionize system administration/other work?" So far all we've seen is that it can make up bullshit. Aside, there's actually been a fight about this between moderators of Stackoverflow and other stackexchange sites. We power-users won: We're allowed to close bullshit on grounds of it being AI-confabulated alone, because checking the texts that AI generates takes more work than the AI has saved anyone. Sadly, that work has to be done by humans, not machines. So, here's your answer. You might really want to read up on StackExchange's official stance on all this;
    – Marcus Müller
    Nov 9 at 16:13
  • it's really already been discussed to exhaustion. And it's been demonstrated at scale (namely, through hundreds if not thousands of accounts trying to farm reputation on Stackoverflow.com and other Stackexchange sites) that at this point in time, and for the foreseeable future, as long as we don't leave the by-design-stupid-but-eloquent LLMs behind, they will not contribute a net positive to knowledge bases.
    – Marcus Müller
    Nov 9 at 16:14
  • ... and after understanding the questions -- but this is a blocker. Current AI bots don't understand anything, they match patterns. When they can understand the questions (and answers), then your question will be a viable one. Until then, it isn't viable IMO.
    – Sotto Voce
    Nov 10 at 6:02

1 Answer 1

Reset to default
4

The following is explicitly against policy, but I thought it would be interesting given the question. I pasted your question into ChatGPT and got the following response:

Using AI to scrape and understand content from websites like Stack Overflow and Unix Stack Exchange to provide solutions to questions or problems in the field of system administration or other work could have both positive and negative implications.

Potential Benefits:

  1. Efficiency: AI-driven web scraping and answer generation could significantly speed up the process of finding solutions to common problems. System administrators and other professionals could save time by quickly accessing relevant information.

  2. Availability: AI can work around the clock, making it possible to provide assistance at any time, even when human experts are unavailable.

  3. Scalability: AI can potentially scale to cover a wide range of topics and respond to a large number of queries simultaneously.

  4. Consistency: AI can provide consistent answers and solutions, reducing variability in problem-solving.

Potential Challenges and Concerns:

  1. Quality and Reliability: The quality of AI-generated answers depends on the quality of data it has learned from. Stack Overflow and similar sites may contain incorrect or outdated information. Relying solely on AI-generated answers without human validation could lead to errors.

  2. Understanding Context: AI might struggle with understanding the full context and nuances of a question or problem. Human experts often provide solutions that take into account specific details and constraints.

  3. Lack of Creativity: AI solutions might be limited to patterns and knowledge present in its training data, which could hinder creative problem-solving.

  4. Ethical and Legal Issues: Scraping websites without permission, particularly if it involves user-contributed content, can raise ethical and legal concerns.

  5. Community and Human Interaction: Part of the value of sites like Stack Overflow is the community of experts who engage in discussions and provide explanations. An AI-only approach could potentially disrupt this dynamic.

In summary, while using AI for web scraping and providing answers has the potential to improve efficiency and availability of solutions, it should be used with caution. A hybrid approach, where AI is used to assist human experts, could be a more balanced approach, combining the speed and scalability of AI with the critical thinking and nuanced understanding of human experts. This might be a more effective way to revolutionize system administration and other fields while maintaining the quality and reliability of solutions.

This seems like a rather lucid answer, but not to your specific question, which is in itself a kind of answer to your question. Would you trust a large language model to apply the “best” answer to a specific problem? We get quite a few questions here from humans not applying answers properly, I doubt it would be much better with an LLM.

Not the answer you're looking for? Browse other questions tagged .