Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

ChatGPT Search

openai.com

Post reply on HN
Print view View original

471–480 of 1001 posts

Re: ChatGPT Search

#471

Post by Bromeo »

rurban wrote: post #464

I just detected perplexity for a proper AI search yesterday. My test question today was: Which cbmc solver is multi-threaded? Perplexity gave me the correct and best answers, with links to the relevant arxiv papers. The new ChatGPT search gave me only cadical as answer, plus 2 irrelevant wrong answers (not multi-threaded), but missed all other multi-threaded solvers. => It's crap. Neither Google nor ddg gave me any r…

Kagi answers that there is "CBMC", which is single-threaded, but that there are extensions "Deagle" and "Yogar-CBMC" that provide multi-threading for CBMC. It gives links to the papers for all three, however some of them are closed access (or in other words, fact checked, unlike arxiv).

Re: ChatGPT Search

#472

Post by Etheryte »

rurban wrote: post #464

I just detected perplexity for a proper AI search yesterday. My test question today was: Which cbmc solver is multi-threaded? Perplexity gave me the correct and best answers, with links to the relevant arxiv papers. The new ChatGPT search gave me only cadical as answer, plus 2 irrelevant wrong answers (not multi-threaded), but missed all other multi-threaded solvers. => It's crap. Neither Google nor ddg gave me any r…

It's hard to cross check since you didn't mention what you consider are good answers, but the quick answer snippet on Kagi says:

> The Yogar-CBMC and JCBMC solvers are notable multi-threaded variants of the CBMC (C Bounded Model Checker) framework: ...

Followed by further details and references. The search results themselves look relevant and reasonable to me, but again, outside my area of expertise.

Re: ChatGPT Search

#473

Post by parsimo2010 »

freediver wrote: post #241

Been thinking about this a lot [1]. Will this fundamentally change how people find and access information? How do you create an experience so compelling that it replaces the current paradigm? The future promised in Star Trek and even Apple's Knowledge Navigator [2] from 1987 still feels distant. In those visions, users simply asked questions and received reliable answers - nobody had to fact-check the answers ever. C…

there still has to be some ranking feature for the backend search database to return the top n results to the LLM. So pagerank isn't over, it's just going to move to a supporting role, and probably modified as the SEO arms race continues.

Re: ChatGPT Search

#474

Post by aloer »

There is so much value in blue links turning purple and showing the same content on repeat visits.

This kind of permanence is a huge loss

Re: ChatGPT Search

#475

Post by faizshah »

hatthew wrote: post #238
cjf101 wrote: post #183

If the current iteration of search engines are producing garbage results (due to an influx of garbage + SEO gaming their ranking systems) and LLMs are producing inaccurate results without any clear method proposed to correct them, why would combining the two systems not also produce garbage? The problem I see with search is that the input is deeply hostile to what the consumers of search want. If the LLM's are partic…

Search engines tend to produce neutral garbage, not harmful garbage (i.e. small tidbits of data between an ocean of SEO fluff, rather than completely incorrect facts). LLMs tend to be inaccurate because in an absence of knowledge given by the user, it will sometimes make up knowledge. It's plausible to imagine that they will cover each other's weaknesses: the search engine produces an ocean of mostly-useless data, an…

You just described the value proposition of RAG.

Re: ChatGPT Search

#476

Post by bamboozled »

whizzter wrote: post #300

This is probably Google's Altavista moment, by making their results crappier by the year in search of Ad dollars everyone has felt that there is a potential for search to be better and once that becomes available they'll be in a continious game of catch-up. Yes, Google has their own AI divisions, tons of money and SEO is to blame for part of their crappiness. But they've also _explicitly_ focused on ad-dollars over a…

No it's not.

Re: ChatGPT Search

#477

Post by maqnius »

encom wrote: post #388
steelframe wrote: post #306

Earlier quoted context omitted.

> Search is pretty much dead to me. I've heard reports that requesting verbatim results via the tbs=li:1 parameter has helped some people postpone entirely giving up on Google. Personally I've already been on Kagi for a while and am not planning on ever needing to go back.

Fuzzy search is cancer. I search for $FOO, click a result, Ctrl-F for $FOO ==> Not found. Many such cases. If there's a way to force DuckDuckGo to actually do what I tell it to, I'd love to hear it.

Try searching for "$FOO", that's what I usually do in those cases. See https://duckduckgo.com/duckduckgo-help-pages/results/syntax/

Re: ChatGPT Search

#478

Post by impulser_ »

Why would I use this over Google? This has the same problem as Preplexity.

Google is instant, why would I wait for a bunch of text to generate just to get basic information.

Re: ChatGPT Search

#479

Post by davidclark »

> Ask a question in a more natural, conversational way

I think this might actually be my main pain point with LLMs. Personally, I don’t want this.

I understand it might be helpful for other people. But, I prefer highly specific, advanced search functionality, such as site: or filetype: in google/ddg searches.

scryfall.com for magic the gathering cards is a great example. I’d much prefer typing a few brief flags such as “id=r” instead of “Get me all red identity cards.” And I know I’m getting all red identity cards with scryfall’s current search functionality.

They are also composable, so I can add/drop ones easily instead of perfectly rephrasing a whole sentence because I wanted to change one clause.

I’d need the same level of trust in the LLM’s filtering capabilities as I do in those boolean or regex matching field filters. An escape hatch to hard filters probably would be best for my experience searching things.

Re: ChatGPT Search

#480

Post by drilbo »

short_sells_poo wrote: post #102
breck wrote: post #46

Earlier quoted context omitted.

If they are making billions from subscriptions, why on earth would they want to switch to an ads business?

Because it is never enough. We see this time and time again. Once they are making billions, the people in charge will demand that they start making dozens of billions, and then hundreds. The growth must never cease, because the moment you stop growing, you can't sell the dream that supports ridiculous PE ratios anymore. Google was a very profitable business 10 years ago and the search was still decent. In the last de…

>It's an interesting showcase of essentially an evolutionary algorithm/swarm optimizer falling into a local optimum while a much better global optimum is out of reach because the real world is something like a Rastrigin function with copious amounts of noise with an unknowable but fat tailed distribution.

I've never heard it framed like this before, that's beautiful.

Post reply on HN