Currently in beta for💲Plus & Teams only, but will be available to free users after beta ends
Available now in beta for Plus and Teams customers
Works as a co-writer, much like working with another person in a Google Doc!
A preview- helps show how users can work WITH AI rather than overdependence on AI.
Getting Started
Select 'ChatGPT 4o with canvas' from the model selector- add files if needed, type prompt
Splits screen- chat on the left, preview on the right.
Users can directly edit the font, heading, etc, and the text in the preview and ask for help and revisions from the chatbot
Suggested Edits & Revision History
Edit text inline
Edit text inline as well as ask for feedback and revisions from the chatbot
The user can then choose whether to accept revisions or not.
Look out Grammarly!
Examples: This feature does not allow sharing of chats (yet...), so I am including a Google Doc showing various versions of a blog post using the different capabilities
More expressive voices, can emulate human emotions, user can interrupt chatbot.
A new feature lets you switch between chat, web search, image generation, and advanced reasoning in one chat.
Step By Step Directions:
Start a new chat in ChatGPT and type "/" in the message line.
Choose from three options: Picture (DALL-E), Search (web), or Reason (GPT-o1).
For images, use "/picture [description]" (e.g., "/picture quantum computer").
For web searches, use "/search [query]" (e.g., "/search quantum computer").
For complex reasoning, use "/reason [task]" (e.g., "/reason Explain quantum computing").
Pro tip: When using the /search command, try adding "latest" or a specific year to your prompt.
OpenAI has been teasing a model that will be much more powerful than ChatGPT 4.0 and it is here!
They released a preview of ChatGPT 01 to Plus and Teams subscribers only on Sept. 12, 2024. Read the Release notes HERE for more information on the capabilities and limitations of this new model.
New AI Reasoning Models: OpenAI has introduced a new series of AI models, starting with the "o1-preview," designed to spend more time 'thinking' before responding, improving their ability to solve complex problems in science, math, and coding.
Enhanced Problem-Solving: These models excel in complex reasoning tasks, outperforming previous versions and achieving high scores in challenging benchmarks, such as the International Mathematics Olympiad and coding competitions.
Safety and Alignment: OpenAI has implemented a new safety training approach for these models, significantly improving their adherence to safety guidelines, with advanced measures to prevent "jailbreaking" attempts.
Education and Research Applications: The o1 series, including the cost-effective "o1-mini," is particularly useful for educators and researchers in fields requiring complex reasoning, such as healthcare, physics, and coding.
Access and Usage: Educators and developers can access these models through ChatGPT and the API, with specific rate limits initially, and more features like browsing and file uploads planned for future updates.
Commitment to AI Safety: OpenAI has formalized partnerships with AI Safety Institutes in the U.S. and U.K., ensuring rigorous testing and evaluation of these models before and after public release.
These models perform best with straightforward prompts. Some prompt engineering techniques, like few-shot prompting or instructing the model to "think step by step," may not enhance performance and can sometimes hinder it. Here are some best practices:
Keep prompts simple and direct: The models excel at understanding and responding to brief, clear instructions without the need for extensive guidance.
Avoid chain-of-thought prompts: Since these models perform reasoning internally, prompting them to "think step by step" or "explain your reasoning" is unnecessary.
Use delimiters for clarity: Use delimiters like triple quotation marks, XML tags, or section titles to clearly indicate distinct parts of the input, helping the model interpret different sections appropriately.
Limit additional context in retrieval-augmented generation (RAG): When providing additional context or documents, include only the most relevant information to prevent the model from overcomplicating its response.
from OpenAI release notes HERE
All users, including free users will now have access to GPT4o mini, as part of OpenAI's quest to provide the most powerful technology to all users.
Users who have exceeded the 16 allowed prompts of ChatGPT 4-o for the day will now be using 4o mini rather than 3.5. This is a significant upgrade!
Roling out now: Free users can create two images per day on free version
Democratization of the best AI models by making ChatGPT 4-O Omni available to free users
with daily limits- when usage exceeds limits, will revert to 3.5 until the next day.
When using GPT-4o, ChatGPT Free users will now have access to features such as:
Experience GPT-4 level intelligence
Get responses(opens in a new window) from both the model and the web
Now, Open AI's context window is up to 128K tokens, a significant increase that allows for much larger file input and AI output and deeper analysis.
Free users may also now upload a variety of file types (all common file types, spreadsheets, presentations, and documents), Connect with Google Drive or OneDrive including multiple files at once.
Create charts, tables, graphs
Analyze Data and Create Charts, Graphs, and Visualizations
The new upload feature allows users to upload data files such as csv, spreadsheet etc and create graphs, charts and visualizations from them or to upload graphs charts and visualizations for analysis
Chat about photos you take
Upload files(opens in a new window) for assistance summarizing, writing or analyzing
Build a more helpful experience with Memory
Up to 16 prompts per day
Example 1: Literature Analysis
Demonstration: Upoad a story like "The Fun They Had" (or even an entire novel, like "To Kill a Mockingbird," into the context window).
Prompt: "Analyze the character development of [character] throughout the novel, providing specific examples from the beginning, middle, and end."
Educational Value: Students can see how GPT-4o maintains coherence and draws connections across the entire text, enhancing their understanding of long-form literature analysis.
Example 2: Historical Document Comparison
Demonstration: Input the full texts of the U.S. Constitution, the Declaration of Independence, Common Sense by Thomas Payne.
Prompt: "Compare and contrast the principles of governance outlined in these documents, highlighting their influence on modern American democracy."
Educational Value: This showcases GPT-4o's ability to handle multiple lengthy documents simultaneously, promoting critical thinking about historical texts.
Example 3: Scientific Research Paper Analysis
Demonstration: Upload a comprehensive research paper such as the one on 'Evaluating AI Detection Tools", including all data tables and figures.
Prompt: "Summarize the key findings of this research, explain the methodology used, and discuss potential implications for educators."
Educational Value: Students learn how to extract and synthesize information from complex scientific documents, preparing them for advanced academic research.
Example 4: Analyze Data
Use some of the data sets, charts, or images in the folder or some of your own and try it out.
Image- Ask ChatGPT to write a caption, explain what is going on in the image, explain the historical significance etc
Data Set- Ask ChatGPT to create a visualization that represents the data
Graph or Chart- Ask ChatGPT to explain it
Limit of 512MB per file, 20MB per image.
💡 Upload your Standard Course of Study & NC Portrait of a Graduate to request help incorporating the POG into your SCOS etc.
💡Available on the ChatGPT app for Android and IoS and on the Mac Book app
ChatGPT App for ANDROID
ChatGPT App for APPLE I PHONE
CHatGPT App for Mac Book
Multimodal Interactions- Engage with ChatGPT with your voice and hear responses in natural voice- no need to type.
Chat GPT can also 'see' and interact with your surroundings via the camera or uploaded images and files.
Live Translation:
Example prompt
"I only speak [language] and my friend speaks only [language]. Act as a live translator. Translate [language 1] to [language2] and translate [language2] to [language1]."
Advanced Voice- with Seemless Real Time Conversational mode and ability to emulate emotional abilities, ability to interrupt, change pace, style
Rolling out now to Plus users, All Plus users should have access by early October
DEMO ADVANCED VOICE: Ethan Mollick (LinkedIn)
AI as a Tutor- coming soon
Combining the new multimodal and vision capabilities with ChatGPT has the potential to work as a tutor.
Free users may now use GPTs (but may not create them)
💡GPTs are user generated AI Chatbots trained for a specific task or purpose- think of them as a preview to agents, which will likely become autonymous.
Explore the GPT Store for a GPT that sounds interesting and try it out. 👇https://gptstore.ai/gpts/categories/education