ChatGPT'S New 'BROWSING UPDATE' Surprises Everyone! (Access Internet, Plugins + Examples)

Chat TPT Update: Web Browsing and Plugins

In this video, the speaker discusses the recent update to Chat TPT that allows users to access web browsing and third-party plugins. The speaker demonstrates how the new browsing feature works and compares it to other search engines like Bing and Bard.

New Features of Chat TPT

• Chat TPT has released a major update that includes web browsing and access to over 70 third-party plugins.

• There are now two different icons for Chat TPT - GPT 3.5 for free users and GPT 4 for those with access to the new features.

• The browsing tab is located in a drop-down menu under GPT 4.

• When using the browsing feature, you can see Chat TPT's entire thought process as it reads information, clicks on links, and extracts data.

Comparison with Other Search Engines

• The speaker compares Chat TPT's browsing feature to other search engines like Bing and Bard.

• The speaker shows examples of how different search engines respond to the same prompt about Google's I/O event in 2023.

• While Bing and Bard give immediate responses, Chat TPT takes time to read through information before giving a summary. However, in some cases, Chat TPT was able to find more information than the other search engines.

Overall, this video provides an overview of the new features of Chat TPT's recent update and demonstrates how its browsing feature works compared to other search engines.

Citing Reliable Sources

The speaker discusses the importance of citing reliable sources when doing research for a paper or argument. They mention that there are different sources of articles that can be included as hyperlinks, which is effective. However, they note that GPT-4 is slow and only allows 25 messages every couple of hours.

Importance of Citing Reliable Sources

• It's important to reference information with reliable sources when writing a research paper or argument.

• Different sources of articles can be included as hyperlinks, which is an effective way to cite sources.

• GPT-4 is slow and only allows 25 messages every couple of hours.

Comparing AI Softwares

The speaker compares GPT-4 to other AI software like Bot and Bard. They note that while Bard was faster at generating responses, it missed some key information related to AI. They also highlight the importance of using the best AI software available for research purposes.

Comparing GPT-4 to Other AI Software

• Bard was faster than GPT-4 at generating responses but missed some key information related to AI.

• It's important to use the best available AI software for research purposes.

Recap Google's AI Announcement at Google's IO in 2023

The speaker demonstrates how quickly Bard can generate a response by asking it to recap Google's AI announcement at Google's IO in 2023. They compare this with using GPT-4's browsing feature, which takes longer but may generate a better response.

Using Bard vs. GPT-4 Browsing Feature

• Bard was able to quickly generate a response when asked to recap Google's AI announcement at Google's IO in 2023.

• GPT-4's browsing feature takes longer but may generate a better response.

• Plagiarism is a concern when referencing information, so it's important to give reference back to the official article depending on the project.

Accuracy vs Speed: Chat GPT-3 vs. OpenAI's GPT-4

In this section, the speaker compares the accuracy and speed of two AI models, Chat GPT-3 and OpenAI's GPT-4. The speaker notes that while Chat GPT-3 is more accurate in providing research sources, OpenAI's GPT-4 is faster and has fewer limitations.

Accuracy

• Chat GPT-3 provides research sources for citations.

• OpenAI's GPT-4 is more accurate in detecting misinformation.

Speed

• OpenAI's GPT-4 is faster than Chat GPT-3.

• OpenAI's GPT-4 has fewer limitations compared to Chat GPT-3.

Potential Applications of Browsing Feature

In this section, the speaker discusses potential applications of a new browsing feature. The speaker notes that Chachi Boutique can now summarize news articles and cite their sources using this feature.

News Summarization

• Chachi Boutique can use the browsing feature to summarize news articles and cite their sources.

Table Formatting

• The speaker tests whether both Bard and Chatbot can list 10 things that happened in AI during the week and put them into a table with links to sources.

• While neither Bard nor Chatbot were able to complete the task effectively, Bing was able to do so in a table format.

Specific Prompts and Information Depth

In this section, the speaker discusses how specific prompts can affect the depth of information provided by AI models. The speaker notes that while Chat GPT-3 can work with simple prompts, it needs a very specific prompt to provide comprehensive information.

Specific Prompts

• Bard needs a very specific prompt to provide comprehensive information.

• Chat GPT-3 can work with simple prompts but needs a very specific prompt to provide comprehensive information.

Information Depth

• Bard provides speed and depth of information for those who know the exact kind of prompt they want.

• Bing is able to list 10 things that happened in AI during the week and put them into a table format effectively.

Fact Checking and Data Sources

The speaker discusses the importance of fact-checking and being able to trace data sources in AI.

Importance of Fact-Checking

• In AI, it is important to fact-check statements.

• If there is something untrue in a statement, it should be possible to verify it.

Tracing Data Sources

• It is important to be able to trace where data has come from.

• This allows for verification of information even if it is still true.

Browsing vs. Bing and Bot

The speaker compares browsing with Bing and Bot, highlighting the strengths and weaknesses of each.

Browsing

• Browsing is not a complete failure.

• However, it can be slow compared to other options.

Bing and Bot

• Bing and Bot are faster than browsing.

• They have more prompts available for obtaining data.

• They are better suited for tight schedules or when more data is needed.

Research Test Results

The speaker conducts a research test on the effectiveness of different AI software options.

Test Prompt

• Find the top three places to get coffee in Vancouver based on the best reviews.

Results - Chat TPT

• After extensive research, Chat TPT was unable to find the top three coffee places in Vancouver.

Results - Bard

• Bard was able to provide results within seconds that were accurate upon cross-checking.

Results - Bing

• Bing was able to provide results within 20 seconds that were accurate upon cross-checking.

Chat TPT in Beta

The speaker discusses the current state of Chat TPT and its potential for improvement.

Current State

• Chat TPT is currently in beta.

• It aims to provide the most accurate information, which can sometimes lead to inefficiency.

Potential for Improvement

• Bard takes prompts better in terms of speed and efficiency.

• It will be interesting to see if Chat TPT can improve its effectiveness.