5 ms·

ChatGTP is amazing. Search engines are dead

2 points by tus666 3y ago

I am getting continually more impressed by ChatGPT. But this blew me away. I don't think I will bother with search engines ever again, apart from looking for a company webpage. Look at this. I wanted to know about the introduction of Windows XP and how it replaced Win 98, over what time-frames. ----------------- > When did WinXP gain widespread adoption among home users? How long did it take? Windows XP gained widespread adoption among home users in the early to mid-2000s, shortly after its release in 2001. It took a few years for Windows XP to become the dominant operating system in the consumer market, as many home users were still running Windows 98 or Windows 2000 at the time. > "By 2003, Windows XP had surpassed Windows 98 in market share," - what was the overall market share breakdown in percentage figures? According to StatCounter Global Stats, which provides data on operating system usage share based on website visits, the worldwide market share breakdown for desktop operating systems in January 2003 was as follows: Windows 98: 35.02% Windows XP: 16.98% Windows 2000: 12.46% Windows ME: 5.45% Windows NT: 1.53% This means that Windows 98 was still the dominant operating system at that time, but Windows XP was gaining ground quickly. Over the next few years, Windows XP continued to gain market share, while Windows 98 and Windows ME declined. By 2007, Windows XP had a market share of over 76%, according to Net Applications. > Can you give me those figures for Janurary 2002? Sure! According to StatCounter Global Stats, the worldwide market share breakdown for desktop operating systems in January 2002 was as follows: Windows 98: 49.38% Windows 2000: 16.22% Windows ME: 13.15% Windows NT: 4.56% Windows XP: 1.73% This means that Windows 98 was still the dominant operating system by a wide margin in January 2002, followed by Windows 2000 and Windows ME. Windows XP had only been on the market for a few months at that point and had not yet gained significant traction. Over the course of the year, Windows XP's market share would gradually increase as more users upgraded to the new operating system. > OK, can you just give me the figures for Win98 and WinXP but from 2002 through 2007? (GPT now formats result as a nice table) Certainly! Here are the worldwide market share figures for Windows 98 and Windows XP, according to StatCounter Global Stats, from 2002 through 2007: Year Windows 98 Windows XP 2002 49.38% 1.73% 2003 35.02% 16.98% 2004 18.32% 47.26% 2005 8.11% 67.87% 2006 3.23% 77.26% 2007 1.54% 76.09% As you can see, Windows XP gained significant market share over this period, while Windows 98 declined rapidly. By 2005, Windows XP had become the dominant operating system, and its market share continued to increase until it peaked in 2006. From 2007 onward, Windows Vista began to replace Windows XP as the primary operating system for new computers, although Windows XP remained in use by many users for several years after its release. > So it was really 2003 to 2004 that saw the big migration to XP. Yes, that's correct. The period from 2003 to 2004 was when Windows XP really began to gain significant market share, largely at the expense of Windows 98 and Windows 2000. During this time, many home users and businesses began to upgrade to Windows XP due to its improved stability, security, and user interface compared to its predecessors. ------------------------------ Now imagine me having to troll through websites trying to get the same information, let alone formatting in such a nice way. I can also ask CGPT for references if I want to go and actually dig in to an article more deeply. But this really poses a question - why would you ever use a search engine again for any non-trivial research task?

hammyhavoc 3y ago
You typo'd the title. It's `ChatGPT`. `/imagine when you discover that ChatGPT hallucinates a lot and is terribly unreliable with facts, especially statistics` If you don't fact-check, you're going to look like someone who half-assed the assignment very soon.
- verdverm 3y ago
  OP's post is case in point, it took me less time to use traditional search to refute the output of ChatGPT https://news.ycombinator.com/item?id=35736072 https://news.ycombinator.com/item?id=35736072
verdverm 3y ago
LLMs cannot search, this is just responding with something it may or may not have read. Have you verified these numbers are in fact correct? Have you tried "searching" for anything outside of its training data? ChatGPT will happily hallucinate references. If you have not experienced this yet, you haven't seen the "true" side of LLMs. There will be plugins for search, but will still need all the crawling and scraping and serving to be available. LLMs will offer an alternative interface to search, but not always the best one. One thing the LLM does not give you is a spectrum of options. How do you know it didn't miss something. This is where having to "troll" through search results is actually worth the effort.
- tus666 3y ago
  My understanding is the ChatGPT basically swallowed up the entire internet, although it was a snapshot that might be out of date now. Of course accuracy and correctness are concerns, but that would apply to any kind of summarizer service, and even primary sources can be wrong. Extensive research would always be required to get near 100% confidence of correctness.
  - verdverm 3y ago
    It only had part of the internet, OpenAI is nowhere near as comprehensive at web scraping as Google, I don't think they actually scraped at all for this, using existing data like CommonCrawl. The other thing you are not understanding is that it did not memorized these things, it built representations for predicting the most likely next token. This is why it hallucinates and makes up numbers and web links or citations that do not exist.

2 points by tus666 3y ago

hammyhavoc 3y ago

You typo'd the title. It's `ChatGPT`. `/imagine when you discover that ChatGPT hallucinates a lot and is terribly unreliable with facts, especially statistics` If you don't fact-check, you're going to look like someone who half-assed the assignment very soon.

verdverm 3y ago
OP's post is case in point, it took me less time to use traditional search to refute the output of ChatGPT https://news.ycombinator.com/item?id=35736072 https://news.ycombinator.com/item?id=35736072

verdverm 3y ago

LLMs cannot search, this is just responding with something it may or may not have read. Have you verified these numbers are in fact correct? Have you tried "searching" for anything outside of its training data? ChatGPT will happily hallucinate references. If you have not experienced this yet, you haven't seen the "true" side of LLMs. There will be plugins for search, but will still need all the crawling and scraping and serving to be available. LLMs will offer an alternative interface to search, but not always the best one. One thing the LLM does not give you is a spectrum of options. How do you know it didn't miss something. This is where having to "troll" through search results is actually worth the effort.

tus666 3y ago
My understanding is the ChatGPT basically swallowed up the entire internet, although it was a snapshot that might be out of date now. Of course accuracy and correctness are concerns, but that would apply to any kind of summarizer service, and even primary sources can be wrong. Extensive research would always be required to get near 100% confidence of correctness.
- verdverm 3y ago
  It only had part of the internet, OpenAI is nowhere near as comprehensive at web scraping as Google, I don't think they actually scraped at all for this, using existing data like CommonCrawl. The other thing you are not understanding is that it did not memorized these things, it built representations for predicting the most likely next token. This is why it hallucinates and makes up numbers and web links or citations that do not exist.