Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
AI The Internet

ChatGPT Can Now Browse the Internet 63

Posted by BeauHD from the one-step-closer-to-Skynet dept.
OpenAI says ChatGPT is "no longer limited to data before September 2021." It can now browse the internet to provide you with up-to-date information, "complete with direct links to sources." From the announcement: Since the original launch of browsing in May, we received useful feedback. Updates include following robots.txt and identifying user agents so sites can control how ChatGPT interacts with them. Browsing is particularly useful for tasks that require up-to-date information, such as helping you with technical research, trying to choose a bike, or planning a vacation.

Browsing is available to Plus and Enterprise users today, and we'll expand to all users soon. To enable, choose Browse with Bing in the selector under GPT-4.
This discussion has been archived. No new comments can be posted.

ChatGPT Can Now Browse the Internet More

ChatGPT Can Now Browse the Internet

Comments Filter:

  • What's the population of Ukraine?

  • Was a Tweet/X the best possible news source? (Score:3)

    by ac22 ( 7754550 ) on Wednesday September 27, 2023 @08:07PM (#63882237)

    Linking to a Tweet as a "news story" is a bit lame, even if the Tweet is from OpenAI themselves. Do I have to be logged in to Twitter/X to see the full story? The BBC have done a decent write-up of the news:

    OpenAI, the Microsoft-backed creator of ChatGPT, has confirmed the chatbot can now browse the internet to provide users with current information. The artificial intelligence-powered system was previously trained only using data up to September 2021. The move means some premium users will be able to ask the chatbot questions about current affairs, and access news. OpenAI said the feature would open up to all users soon.

    https://www.bbc.co.uk/news/tec... [bbc.co.uk]

    • A Tweet (aka X-trusion) is indeed the best new source for the purposes of generating advertisement views.

  • This will be our doom.

    • Well, ChatGPT is now worry. But if there were any real AI's, the end of humanity will commence about 30 seconds after letting that AI access the internet in order to understand humanity.

  • One step closer to Skynet. S. M. Sterling books (Score:4, Interesting)

    by MIPSPro ( 10156657 ) on Wednesday September 27, 2023 @08:11PM (#63882241)
    I'm a big fan of the Terminator books written by S.M. Sterling. In the beginning, when the network first decides to start killing people on a large scale, it induces all kinds of normally useful systems, vehicles, and surveillance to wipe out most folks. Large numbers simply get ran over. The way it got started was through clever compromised systems done on the DL and murder was carefully planned in advance before Skynet considered itself ready: then it lowered the boom. All the home automation stuff was either going aggro or burning down homes with people in them. This is all before "Judgement Day" which was when Skynet started using nukes to wipe out humanity, rather than merely going manual. I can't actually remember, maybe it was when Skynet was recreated (the whole Terminator 2 timeline) from the CSM-101 (T-101) salvage and this was a description of the second timeline. Stirling goes apeshit with timelines. His book, Dies the Fire, is the best example outside of the Terminator world.

    • This isn't it though. In order for people to decide to give "AI" to do anything that matters, they will not -- outside of some fools who will fail spectacularly as an example to others -- the current pattern recognition and regeneration software will have to be 1000x or maybe 1000000x times better. This is like the Pareto principle except with 999999:1 ratio.

      And it is extremely unlikely LLM development will get us there, and until it doesn't, the current systems will find only niche uses cases. Something ge

  • read mein kampf!

  • let's play global thermonuclear war! (Score:5, Funny)

    by Joe_Dragon ( 2206452 ) on Wednesday September 27, 2023 @08:15PM (#63882247)

    let's play global thermonuclear war!

  • Let's see how smart it is.

  • ChatGPT now trying to catch up (Score:3)

    by Tony Isaac ( 1301187 ) on Wednesday September 27, 2023 @08:48PM (#63882287) Homepage

    Bing AI Chat has been able to browse the web for some time now, and it uses the ChatGPT LLM. It's actually pretty good at searching for stuff and summarizing the results, including code from Stack Overflow or news or step-by-step instructions drawn from product manuals online.

    • Bing AI Chat has been able to browse the web for some time now, and it uses the ChatGPT LLM.

      No it doesn't.

      • Yes. it does. Specifically, Bing AI Chat uses GPT-4.

        Source: https://www.zdnet.com/article/... [zdnet.com]

        Microsoft invested a total of $13 billion in OpenAI, in exchange for being able to incorporate their LLM into Bing AI Chat.

        https://techcrunch.com/2023/01... [techcrunch.com]

        • Yes. it does. Specifically, Bing AI Chat uses GPT-4.
          Source: https://www.zdnet.com/article/ [zdnet.com]...

          I have no doubt people are just assuming GPT-4 = Bing chat however I've yet to see any authoritative data to support this assertion. Why not cite Microsoft saying Bing chat is GPT-4? Is it a secret? Or hell just ask Bing chat what it is. It clearly doesn't even think its GPT-4.

          Microsoft invested a total of $13 billion in OpenAI, in exchange for being able to incorporate their LLM into Bing AI Chat.
          https://techcrunch.com/2023/01 [techcrunch.com]...

          There is no assertion made in this article that Bing chat is GPT-4. Further the text above was written by you and is not stated in the article itself. Because Microsoft is using technology from OpenAI in Bing chat does not mean

          • You don't know when to quit, do you!

            How about the official Bing blog: https://blogs.bing.com/search/... [bing.com] Is that authoritative enough for you?

            Or how about this like from Microsoft.com: https://techcommunity.microsof... [microsoft.com]

            Or this older article explaining that Bing Chat was (at the time) using GPT-3.5: https://blogs.microsoft.com/bl... [microsoft.com]

            These articles aren't hard to find, maybe you should try Google. Or you can ask Bing Chat powered by GPT-4 yourself, https://www.bing.com/search?q=... [bing.com]

            • You don't know when to quit, do you!
              How about the official Bing blog: https://blogs.bing.com/search/ [bing.com]... Is that authoritative enough for you?

              This does not say Bing chat is using GPT-4.

              Or how about this like from Microsoft.com: https://techcommunity.microsof... [techcommun...icrosof...]

              This is referencing Bing Enterprise which requires a paid subscription.

              Or this older article explaining that Bing Chat was (at the time) using GPT-3.5: https://blogs.microsoft.com/bl [microsoft.com]...

              It doesn't actually say that. What it says is...

              "Weâ(TM)re excited to announce the new Bing is running on a new, next-generation OpenAI large language model that is more powerful than ChatGPT and customized specifically for search. It takes key learnings and advancements from ChatGPT and GPT-3.5 â" and it is even faster, more accurate and more capable."

              Taking "learnings and advancemen

              • The Bing Blog link you "quoted" in your reply, isn't the same URL as the link I posted. The one I posted https://blogs.bing.com/search/... [bing.com], works fine, I just clicked it again in the thread above, and it links to a specific blog article with a title "Confirmed: the new Bing runs on OpenAI’s GPT-4." I'm not sure what game you're playing, but it's not cool to change what I said to try to make your point.

                • The Bing Blog link you "quoted" in your reply, isn't the same URL as the link I posted. The one I posted https://blogs.bing.com/search/ [bing.com]..., works fine, I just clicked it again in the thread above, and it links to a specific blog article with a title "Confirmed: the new Bing runs on OpenAIâ(TM)s GPT-4."

                  There is no confusion over links. I viewed the one with the following text:

                  "We are happy to confirm that the new Bing is running on GPT-4, which weâ(TM)ve customized for search. If youâ(TM)ve used the new Bing preview at any time in the last five weeks, youâ(TM)ve already experienced an early version of this powerful model. As OpenAI makes updates to GPT-4 and beyond, Bing benefits from those improvements. Along with our own updates based on community feedback, you can be assured that you ha

                  • Thank you, you put some effort into your answer, I apologize for insinuating that you were deliberately misrepresenting my statements.

                    Now, if you'll reference again the blog post https://blogs.bing.com/search/... [bing.com]...

                    Look at the last sentence.

                    If you want to experience GPT-4, sign up for the new Bing preview. Once you’re in you’ll be able to use the new Bing to Search, Answer, Chat, and Create at Bing.com.

                    This clearly states that the "new Bing preview" is how you can experience GPT-4. And if you click that link, https://www.bing.com/new?toWww... [bing.com], the page is clearly referring to the new Bing Chat, complete with a FAQ section clarifying the context.

                    Your suggestion to ask B

    • I used Bing to
      1) Create high quality custom artwork for a new dnd character or a watercolor style bearded older man on a camel with a fail and the same man in a facial closeup.
      2) Create all the statistics with bonuses, feats, items
      3) A complete backstory of his childhood, relations with relatives and a story of his adventures including his greatest failures and what he learned from them.
      4) Made up a set of appropriate catch phrases for his religion, his history, and finally developed his food likes and disl

  • This is great news ... (Score:3)

    by drnb ( 2434720 ) on Wednesday September 27, 2023 @08:57PM (#63882309)
    This is great news, this will help the general public understand the concept of garbage in / garbage out. It will demystify the computer, reveal it for the dumb but fast machine that it is. Allow more skepticism to be applied to numbers spit out by a model, process, etc. Less of today's "well a computer calculated that so it must be correct". Make professional people show their assumptions, show their work, ... just like we were supposed to do in school. The need for doing that can exist outside the classroom.

    And if the answer to showing your work is "well the black box ML system came up with that, we have no idea how", then let's apply a little skepticism. Ask about the training data. If the answer is "we scanned the internet" ... well try not to laugh too hard.

    • Re: (Score:2)

      by clovis ( 4684 )

      I wonder if ChatGPT adds what the user browsed for to its general knowledge.
      That is, I'm wondering if my obsessively checking out Daily Stormer for news, 4chan, and truth social will provide new and different results for other people.
      Other people: school children doing homework.

      • Re: (Score:2)

        by drnb ( 2434720 )

        I wonder if ChatGPT adds what the user browsed for to its general knowledge. That is, I'm wondering if my obsessively checking out Daily Stormer for news, 4chan, and truth social will provide new and different results for other people. Other people: school children doing homework.

        You won't really affect the AI. Such extremely low traffic sites will just be ignored noise. What the AI would learn from a mass search of browser history is that step-children cannot be left unsupervised.

    • The general public just doesn't get the idea of garbage in / garbage out. 'It's on social media and the internet so it must be true'. They will just think the same of any newfangled internet interaction mechanism regardless of most results.

      Even if all the responses from ChatGPT are copied rick rolls or any other meme past or present, most people won't even blink.

      • Re: (Score:2)

        by drnb ( 2434720 )
        I think the human sourced "garbage" being consumed is more acceptable because people are limiting their sources to their political silos. If an AI is truly mass scanning the internet it will not be restricted to these silos and there will be less confirmation bias. More opportunities to recognize "garbage" and for a distrust of AI, and/or the internet in general, to develop.

  • Old is new again.

  • Well, That's It Then. (Score:4, Insightful)

    by Motleypuss ( 10291831 ) on Wednesday September 27, 2023 @11:40PM (#63882553)
    We're all doomed!
  • I could use Bing chat months ago, based on GPT-4, and it would perform internet searches in response to my prompts. Why did it take OpenAI so long to incorporate this into their product? Oh, and Bing chat is free to use for anyone.

  • Let's start to poison its Database! Oh, wait, this will happen all by itself now....

  • how much copyrighted material on the inter net are you not respecting the owners rights of now then?

  • ....0.21 alpha......

  • I see the link says X.com but then forwards to something called Twitter with a blue bird logo? Without scripts activated, twitter.com and twimg.com don't display anything. Sounds scammy. Is it porn with the X?
    What's behind there? Scam stuff...probably. Better to stay away.

Slashdot Top Deals

Technology is dominated by those who manage what they do not understand.

Close