Debate Devil

AI

debate practice chatbot

SHOW HN2023

84points
83comments
#761of 3,109 in 2023
Top 25%of 2023 by points

How HN took it

RoastedHalf or more of the comments are critical

7 positive, 11 neutral or asking, 1 mixed, 29 critical, out of 48 comments about the project. Thread score -0.43.

Execution: 1 for, 22 againstIdea: 5 for, 4 againstDesign: 1 for, 3 against

The thread, comment by comment

Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.

positive neutral or mixed critical about another commenter or something else the maker

What HN said

What people liked

  • Good idea for rubberducking and practicing arguments

    From 3 comments
    • This seems like a very useful way to rubberduck ideas that are not yet well formed. I like it! Are there limits to a conversation? The input prompt disappeared in the middle of…
      collinvandyck76 on HN
    • I think this is genius. Thank you. This is the best use I've ever heard of for an LLM. Online "debates" are always abysmal. People talking past each other, repeating arguments…
      jfengel on HN
    • this seems like it might be useful for arguing on HN
      turmeric_root on HN
  • Clean, no-login user experience

    From 1 comment
    • Setting the actual product aside for a moment, I must commend this as one of the most outstanding demonstrations I have encountered on Hacker News in recent memory. The seamless…
      user00012-ab on HN

Objections

  • AI judge declares winner inconsistently or favors the AI regardless of argument quality

    From 4 comments
    • 1. I have a character limit, it doesn't 2. It throws multiple arguments at you 3. Then it stops the debate after a certain point 4. It repeats its arguments 5. then the AI judge…
      Brendinooo on HN
    • The AI ended up conceding my initial position word for word, and yet the AI judge said the AI won.
      aerhardt on HN
    • Fairly terrible. I asked the AI to defend Nazism (I said it was bad), it refused to do so, "agreed" with me that Nazism is bad, then, when I pointed out that this means I won the…
      boringuser2 on HN
    • Very neat idea. However the debate I tried was farcically bad. The bot just provided "facts" that go counter to any statistics on the matter. Then the "AI judge" ruled the bot to…
      numbers_guy on HN
  • Arguments are shallow, repetitive or factually wrong

    From 4 comments
    • It just spouts fake news and propaganda, and then seems to get stuck in a loop when its challenged in a way it can't respond to.
      danpalmer on HN
    • It isn't great. I argued "the sun is hot" - lots of ways to argue against this, but it got stuck on basically repeating "that might not be the case" followed by statements that…
      Chris2048 on HN
    • Very neat idea. However the debate I tried was farcically bad. The bot just provided "facts" that go counter to any statistics on the matter. Then the "AI judge" ruled the bot to…
      numbers_guy on HN
    • The AI's core tenet to my assertion "Python is an awful language for beginners" (due to whitespace formatting and having a non transparent "this"-self- having to manually be…
      SillyUsername on HN
  • Character limit for user is much lower than the AI's

    From 2 comments
    • The AI writes many paragraphs but I am limited to just a tiny response. That is frustrating.
      calderknight on HN
    • The character limit for the user being dramatically lower than the AI’s limit is really annoying, otherwise I had fun with this!
      virgildotcodes on HN
  • Stays within guardrails and doesn't argue all sides

    From 2 comments
    • It won’t actually argue all sides of an issue. Take any controversial one, and it stays inside its guardrails no matter which opinion you want it to counter.
      mrjbq7 on HN
    • Fairly terrible. I asked the AI to defend Nazism (I said it was bad), it refused to do so, "agreed" with me that Nazism is bad, then, when I pointed out that this means I won the…
      boringuser2 on HN
  • Name 'Devil' is an odd choice

    From 1 comment
    • I wouldn't use a product called "devil". Like why take a concept that's evil incarnate and associate it with your product when there's a million other names to choose from.
      user3939382 on HN

Requests

  • Ability to share conversation transcripts without an account

    From 2 comments
    • Is there a way to share my transcript?
      rain1 on HN
    • Very cool! I asked "Is there a God?" on Devil's advocate mode and won! Prompt: Is there a God? To begin, the AI asked me what my thoughts on the existence of God were: My…
      sunnysidedown on HN
  • Persona that uses logical fallacies for practice

    From 1 comment
    • I gave this a try and it’s really pretty cool. Something I am looking for is the ability to have a debate with someone who purely uses logical fallacies. It would be interesting…
      minimally on HN
  • Allow longer user responses

    From 2 comments
    • The AI writes many paragraphs but I am limited to just a tiny response. That is frustrating.
      calderknight on HN
    • The character limit for the user being dramatically lower than the AI’s limit is really annoying, otherwise I had fun with this!
      virgildotcodes on HN

Questions

  • Can the AI judge ever rule the user wins?

    From 1 comment
    • 1. I have a character limit, it doesn't 2. It throws multiple arguments at you 3. Then it stops the debate after a certain point 4. It repeats its arguments 5. then the AI judge…
      Brendinooo on HN
  • Is there a limit to conversation length?

    From 1 comment
    • This seems like a very useful way to rubberduck ideas that are not yet well formed. I like it! Are there limits to a conversation? The input prompt disappeared in the middle of…
      collinvandyck76 on HN

Alternatives named

  • Plain ChatGPT

    From 2 comments
    • zero value add to a regular chatgpt besides some kind of web wrapper around it
      wellthisisgreat on HN
    • The AI's core tenet to my assertion "Python is an awful language for beginners" (due to whitespace formatting and having a non transparent "this"-self- having to manually be…
      SillyUsername on HN

Category, keywords, summaries and labels on this page were made by AI (Thread Reader v2) and not checked by hand. Each point links to the comments it came from.

Add a badge to your READMEShown on HackerShow
[![Shown on HackerShow](https://hackershow.org/badge.svg)](https://hackershow.org/show/35568799)