ChatGPT now supports bigger context and function calling!!!

Yes, that's correct! OpenAI has introduced an update to ChatGPT that allows for a larger context window and supports function calling. This enhancement enables more extensive and coherent conversations with the model, as it can now access and refer back to previous parts of the conversation. Additionally, the introduction of function calling allows users to invoke specific commands or functions within the conversation, making interactions with the model more interactive and dynamic. Feel free to explore these new capabilities and let me know if you have any questions or need assistance with anything! sehat card

It will greatly enhance the capabilities of the language model and enable more dynamic and interactive conversations. Great job to the team behind ChatGPT

Even though parsing structured outputs with, e.g., langchain is super easy, I suspect that this is going to take it to the next level. One particular use-case that I'm working with is disentangling structured from unstructured bits of a natural language query to an image database. I think this will be much easier when I can ask the model to provide summaries of each bit to separate parameters in a function. Very excited to get started with this 🤩

I have to admit, after playing around with the "weather function" as an example I'm excited. https://platform.openai.com/docs... Function calling allows you to more reliably get structured data back from the model. For example, you can: Create chatbots that answer questions by calling external APIs (e.g. like ChatGPT Plugins) e.g. define functions like send_email(to: string, body: string), or get_current_weather(location: string, unit: 'celsius' | 'fahrenheit') Convert natural language into API calls e.g. convert "Who are my top customers?" to get_customers(min_revenue: int, created_before: string, limit: int) and call your internal API Extract structured data from text e.g. define a function called extract_data(name: string, birthday: string), or sql_query(query: string) While the SQL example is a tad too dangerous for me, the natural language -> API might be interesting to explore further.

Interesting to have that. Do you have idea how large the context limit is?

Function calling will make it a LOT easier to standardise GPT output. This is valuable to us as we pass GPT results to the FE, and so require it always to be in the same JSON format. Currently, to ensure this we are doing some convoluted prompt engineering and cleaning up the data in the BE, but hopefully this flow can now be simplified.

Moving faster than I thought. Exciting wohoo

Super exciting - played around last night and built my first web app in React within 3 hours using ChatGPT. SUPER impressed with the team and how well ChatGPT handles long, complex inputs!

I have now upgraded to gpt-3.5-0603, and the response time has significantly improved. The model's ability to plan and extract function parameters has also noticeably increased.

Moving faster than I thought.