Connect ChatGPT to Internet

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
chat GPC gpt4 or in General open AI models are not connected to Internet that means the data set that they use to get trained is based on historical data available on web maybe based on couple of years ago data and if you want to ask questions that is relevant to the data available real time and live on internet they cannot answer because they're simply not a search engine they're just large language models question is can we connect these models like chat GPT to live real-time internet data so we can start chatting with live and update the data on internet if you want to figure it out then let's go hello everyone this is mg and welcome to another open AI video and in this video we're gonna talk about how you can connect your open air model like chat GPT to Internet that means next time when you start asking questions or any question from these models they will be able to answer you with the most recent live updated data on internet and the trick here is again these models are not search engines and we cannot leverage them as a search engine but what we can do for any relevant prompt and question you're going to ask from these models we can bring search results from internet relevant to that prompt and then let Chachi BTC them interpret them and start answering you in a conversational manner based on the updated data that we grab from internet although there are some plugins you can do it as of now but what we're gonna do we're gonna code this process so you can include this solution that means connecting charity to internet within your own project and application that can be a piece of your steps that you do in your projects that can certainly open up a new era of opportunities and possibilities for your use cases let's check it out before we start make sure you subscribe and hit the Bell icon so you will get notified for the next video right so we recorded couple of videos that how you can start using chat GPT gpt4 and all these opening upgrade models to chat with your own data I mean how can I use chat gbt on my own large data set on my own project data set I may own big company data set so I will add the link of all those videos in the video description and also I will add it to the top right of this screen that will tell you how you can let's say use chat DVD to start creating a chatbot that you can ask questions from your own data set because chatgpt and these models they have not seen your data they have got trained with internet data or web data anything available online so if you have something priority just for yourself your company these models they don't have any knowledge about them so there are some tricks on working on that how you can start using them ask questions about your own data so we talked about first text data or tabular sorry unstructured data we had some PDF files from our own company and we started to sort of ask question against those and then on the top we created another video on how we can start chatting with a let's say SQL live SQL database instead of you writing on SQL query ask your query as a human language based conversational Manner and the result will be back so on backhand chat GPT will create the queries for you execute it the line chain will execute it on your database server and then the result is going to be back without you writing any single lineup code and what we're gonna talk about today how we can chat with internet data or live internet data these models again they do not have access to live intranet they have been trained with historical data years ago on the internet and if something happened let's say yesterday or today these models they don't have any context about it so you cannot asking questions about live data now we're going to focus on what's going to be the workaround that we can let using this chart GPT and opener models to ask questions relevant to the live internet data so in order to show you a little bit further what's going on when we talk about live data for example here I am opening up the guardian football live scores so the time that I'm recording these videos today Saturday April 15th and you can see for example Newcastle and Aston Villa match theater three versus zero that was sort of the match results and now if I ask chat GPT that give me the result information for that specific match look what's gonna happen I actually have the question ready so let me copy and paste in my chat GPT UI what is the match result between Aston Villa and Newcastle on April 15 2023 which is today as an AI language I do not have access to real-time information so it cannot really help me with that queries that was just a simple query I don't need to ask this question from chat GPT I can just open up my browser and ask it right but you can have more complex questions that is beyond just a search that needs of live data and they cannot answer but take this question as an example to show you that they do not have access to real data but then what Microsoft did with Bing and you might have already heard about it which is called Bing chat they started to have these models like charity Beauty but connected to internet now I can ask questions from internet using um chat version of Bing and we're gonna see how we can create our own like chat being version with our own code to using our own project so if I ask the same question in bing let's see what's going to be the answer so it is searching for the The Prompt that I generated and it says that yeah based on the state the match let's see yeah we discussed three zero this is exactly what we just saw in Guardian website and that sort of works right now here's the question how we can create our own Bing chat in a code that I can write down my question and it will ask my question based on live data on internet using chat GPT so if you have it implemented automated encoded you can use in a variety of different type of applications and use cases you might have been thinking about it okay so let's do it what I actually created I need to first when I ask a question let's say for example what is the match result between Austinville and Newcastle so what will happen before being called chat GPT to answer this Bing will search for this question with search engine it will find the references you know it's it's a simple search that you have seen in Google browser or even ping it will start searching all the references to find answers relevant to that and then when has all these references then it will ship all these references as a prompt to charge GPT and will ask chargpt that hey answer this user question or you know you don't have access to real data but I have brought it for you with Bing search engine this is exactly what I gonna code and Bing search is an engine is just an example you can use I don't know Google search API or whatever to bring relevant sources to chatgpt and now chat GPT can see real-time live data to answer your question so in order to do so I need to have openai API definitely use your Azure Azure open AI key or the the openai.com API link you don't need to have Azure open AI you can have just a simple opening I keep by yourself then I need to have a search engine API to start getting the references and then ask your chat GPT answer the question based on the references so where did I grab my search engine API this is my being search cognitive service I created on Azure you just type Bing search on your Azure portal it will give you a key API key that you can use search engine of Bing on code so ask question and get the results back for example here's a test try me if I gonna search machine learning there we got some results for me machine learning search result in the Json format and both I can see the reference some content so on and so forth so if you go to the key and endpoint you need to copy and paste copy your both key and endpoint because you need to have them in the code so grab them this is of course confidential grab those and here's the code and it will again add the code notebook in the video description as well you can check it out it's a very simple and a set forward code okay let me go back to my code so as you can see I'm here using Azure ml to run my python notebook but I'm it's not a must just open up your vs code or Jupiter Notebook on your laptop you don't need to be on cloud wherever you have you are as long as you have python installed and pip open AI is also installed you can run this code so don't care about my environment Azure mail whatever so I'm running the code it is asking me now ask a question so let's start asking the same question that we ask from Bing chat and chat GPT which is what is the match result between Australian Castle we're gonna see if it really works with the chat being similar to chat being or not paste there you go first it is started to grab the sources that talk about this question as you can see these are sources relevant to that match and chat gbtus and all the way here's my question sorry what is the match result of blah blah and here is answer Villa Biz Newcastle three zero in the match on April so I sort of recreated something like chat being put in my code and Chan being to the best of I know doesn't have any API so I'm not using the chat being API I don't know even it has it but I created such an environment a code that can query the the data from internet based on the question and ask chat GPT to answer based on those sources so chat saw these sources but let's go through go through the code and see how I implemented that actually so at the beginning first you need to have your open AI key so I'm using in Azure open AI but you can just use the opening out of azure it doesn't matter and you can put these credentials inside in OS operating system environment to have it secured also beside open AI you need to have your Bing search API key and endpoint as well I told you how we can copy that again you can use Google search engine API whatever as long as you can't have the the search result get back to you and then in this function here I am copying actually Azure search bin gave me this code example in Python I just modified a little bit that it will start asking finding the results of search based on your given query so when I have this function then I ask user give me your question and then I call this search query that calls Bing to get the results and then I will get the result of Bing search pars over this and then when I want to ask chat GPT I say that you use these sources which are coming from Bing search results to answer the user question and then lastly I'm just calling Azure opening a model and you can see I'm using GPT 2.5 turbo which is chat GPT and then it will give me the results back as you saw so that's it that was a very simple code to just show you the art of possibility that how you can even bring live internet data to chat GPT in an automated encoded manner for your applications Bing was just an example any search engine that gives you an API to grab the research results use it grab the search result give it to the charitivity and then ask question it will start answering to you hope you enjoyed the video and that's it do you know what is the hardest things to listen The Whisper of your dreams The Whisper of your instincts your human personal intuition they do not shout sometimes they just whisper a slowly inside you so it's very important to listen carefully of those Whispers that talks about your human personal intuition and when you listen to that carefully and if you see there is something that is tickling your heart that's your destiny and that's what you're gonna do for the rest of your life we all gonna benefit from it dreaming my friends believe in yourself and take action [Music]
Info
Channel: MG
Views: 8,399
Rating: undefined out of 5
Keywords: open ai, gpt 3 ai, gpt 3, azure ai, artificial intelligence, machine learning, openai chatbot, gpt 3 fine tuning, openai gpt 3, gpt 3 prompt engineering, openai chatbot gpt writing code, openai chatbot demo, chat gpt explained technical, open ai chat gtp, ChatGPT in Azure, Open AI in Azure, chat gpt, advanced tutorial chatgpt, chatgpt advanced guide, GPT 4, SQL Data with ChatGPT, connect chatgpt to internet, Connect ChatGPT to bing, Azure Bing Search, Bing Search API
Id: u943E5PeDJU
Channel Id: undefined
Length: 13min 49sec (829 seconds)
Published: Tue Apr 18 2023
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.