Live data from Hacker News

AI chatbots are not a replacement for search engines

iai.tv

261–270 of 415 posts

Re: AI chatbots are not a replacement for search engines

#261

Earlier quoted context omitted.

I believe this to be a huge deal. In no time a huge part of the internet will be AI-generated and we will pollute our data set for research and training AI with things that are subtly or completely false at times and we won't be able to tell. It's a lot like iRobot or other stories in which the robots will start to subtly misbehave, like a program you can't debug.

Well this is unfair because we haven’t given natural intelligence a chance yet. What if we invented an algebraic notation for search indexing and allowed people to index and classify documents in their own way. Then a search engine could do something special with all that. Human intelligence cannot progress without a notation to capture our thoughts in.

Are you aware of https://commoncrawl.org/ ?

Re: AI chatbots are not a replacement for search engines

#262
post #259

I came to the comment section looking for some valid arguments for/against the article, but nope.

This is the most terrifyingly stupid comment thread I've read on HN lately. What's unique about it though is the thoroughness of the bad quality comments. It's bad all the way down. I think I'm skipping my HN scroll this evening and just going to bed.

Re: AI chatbots are not a replacement for search engines

#263
It is incredible that we as humanity are about to put our info search in hands of a hallucinatory text generator, without certainty about (i) how resilient the whole thing is against spam/propaganda, and (ii) what happens to people unable to find the difference.

Re: AI chatbots are not a replacement for search engines

#264
post #22

The way I see it: - Today's search engines will give you links to millions of documents - ChatGPT, if asked correctly, will instead generate one perfect document based on millions of the documents To me, that's a clear evolution of the search engine, especially with all the SEO & ad spam that's plaguing Google and others currently. I wasn't ready to pay a monthly fee for an ad-free Google. I am ready to pay for somet…

The problem with ChatGPT as a replacement for Google is that it was not designed to produce accurate facts, and it shows. This model cut its teeth writing articles about the discovery of unicorns in the Andes[0] for goodness sake! It's a language model, and a very impressive one at that, but language is used to express falsehoods and fiction just as regularly as it is used to express truth. This doesn't mean that it…

1. Provide citations like you.com already does. 2. Add Trustrank/Trustflow into the training set

Re: AI chatbots are not a replacement for search engines

#265

Earlier quoted context omitted.

People will not write books and blogs if they can't pay the bills. If AI xant replace those, you will have no books and blogs. Thats how market works. > People will also continue to chat with each other... so you'll always have that as a data source. Your solution takes us back to before invention of the printing oress? the dark ages?

People won’t just write encyclopedia articles for free, right?

Many Wikipedians write for free, indeed.

Re: AI chatbots are not a replacement for search engines

#266
post #22

The way I see it: - Today's search engines will give you links to millions of documents - ChatGPT, if asked correctly, will instead generate one perfect document based on millions of the documents To me, that's a clear evolution of the search engine, especially with all the SEO & ad spam that's plaguing Google and others currently. I wasn't ready to pay a monthly fee for an ad-free Google. I am ready to pay for somet…

The problem with ChatGPT as a replacement for Google is that it was not designed to produce accurate facts, and it shows. This model cut its teeth writing articles about the discovery of unicorns in the Andes[0] for goodness sake! It's a language model, and a very impressive one at that, but language is used to express falsehoods and fiction just as regularly as it is used to express truth. This doesn't mean that it…

RIP google. There's a joke that the best place to hide a murder is on page 2 of google search. The fact is GPT has at bare minimum parsed the top page of google and reasoned about the top results to give the best answer with as much context of why it reasoned so. GPT is basically a superset of google with added reasoning. I can't see it giving worse results than Google. Both will produce a bogus answer if you ask a bogus question, but GPT will never produce a bogus answer when google has the correct one.

Re: AI chatbots are not a replacement for search engines

#267

Earlier quoted context omitted.

OTOH I could imagine a hybrid situation where the AI Bot searches Google for you and restates the millions of results Google gives it

I agree that a hybrid could work, although ChatGPT struggles to keep context with more than a dozen paragraphs, so I doubt it could summarize millions of results. And in either case, you're not solving the problem that commenters here are trying to solve: Google's bad results will still affect your search, so ChatGPT will be fed largely by dubious blogspam.

Even in a small number of paragraphs language models hallucinate. For example Barack Obama, it’s not hard for me to trick it into saying he was born in Chicago. He was not. But there’s a ton of text about Barack Obama in Chicago. Plus, Michelle was born there. When predicting the most likely next token for “Barack Obama was born in ______” there’s a lot of confounding text.

No, there’s a fundamental issue in that statistical language models don’t start with facts to form sentences, they form sentences based on previous tokens and statistical weights. The weights are not auditable or easily updatable.

Re: AI chatbots are not a replacement for search engines

#268

I compare it all to using Github Copilot in my daily work. Before, I needed to Google many problems to find answers, curate the answers, scroll down websites, read downvoted comments on StackOverflow, copy/paste code, refactor it, turn code into TypeScript, etc. Now, I just tell Copilot what I need to do, and I simply curate multiple solutions. ChatGPT also understands what I need to do and changes things based on hu…

If I understand correctly, the ChatGPT training is a continuation of the CoPilot training, which is a continuation of GPT-3 training. It makes sense they exhibit a similar behaviour.

Re: AI chatbots are not a replacement for search engines

#269

I agree that chat bots aren’t the proper modality for replacing search. So what? Currently, Google search results stink. ChatGPT results are way better in a number of domains. Does it need to be a chat bot? No. But Google still stinks now. I’ll take anything that can just find the correct information.

I tend to agree for a lot of queries. For example, I just got curious and asked Google how many oranges in a cup of juice. The main result is a conversion of a cup to cubic inches, for no reason. Scrolling down a bit, the first usable answer was a quora blurb. ChatGPT told me it depends on the size and type of orange, but generally 2 to 4. Guess which one I'll use next time?

I just asked Google assistant this: "How many oranges are in a cup of juice?". It responded with: "Two to three oranges", and then followed up with more information why and a link to the source where it derived the information (https://www.webstaurantstore.com/blog/2760/juice-in-citrus-f...).

I'm not sure why you didn't get this, but it seems better than ChatGPT.

Re: AI chatbots are not a replacement for search engines

#270

Earlier quoted context omitted.

Absolutely this. I just did a discussion with ChatGPT on intuitionistic logic and it gave answers that were subtly wrong in my understanding of the subject. I have an MA in math but I couldn't be sure whether my impression of its wrongness was correct until asked it for a reference and it gave a nonexistent book by Per Martin-Lof, "Intuitionistic Logic", along with a list of actual books by him. It's an impressive an…

No, you are probably right. It has no guardrails for generating a true citation. When you ask it for citations, its programming generates the form of citation. As the authors of the article discuss, it has no model of reality to understand that citations are not just a form of words and symbols put together.

Indeed and it gave several other actual citation with seemingly correct summaries.

And it's definitely less prone to obvious bs than just a few weeks ago. For a lot of general things, it gives a summary paragraph that's seems a distillation of standard information - the main problems seem to appear when these "scripts" lack consistency.

Edit: Once it "learns" to get citations right, you'll have to dig into those to see if it's right. It's easy to see that arms not looking good.

Post reply on HN