Implementing new OpenAI Features
Jonathan Kressaty
Groupthink is the place to chat with VI, a GPT-powered AI that reads the internet and connects to your email, browser history, and Slack history to provide personalized responses. Invite your friends and family, collaborate on projects, and make use of Groups and Threads to organize conversations.
Welcome new (and returning) users! For those who may be seeing this email for the first time, we are the founders of Groupthink and we send this daily email with our progress.
Improving Groupthink with new OpenAI Capabilities
When you talk to Groupthink, we perform a number of steps before the bot responds, many of which use services from OpenAI. These steps include everything from summarizing your conversation to answering a question using text results from the web.
OpenAI limits the amount of text that can be used to generate a response. This means that you can’t simply ask OpenAI to summarize an entire book for you – it can’t handle that much text at one time. We’ve built software that allows us to find the most relevant parts and return a high-quality response, but it’s still limited by the maximum “memory” of the language model.
This week, OpenAI announced a new language model based on the system they use to power ChatGPT. One of the key improvements is the ability to understand twice as much text as it can today.
We believe that the entire industry will see these systems change and grow at a rapid pace, and we’ve built Groupthink to make it easy and fast for us to implement these improvements.
We’re excited to implement these changes into Groupthink next week! You’ll see improved responses, especially when asking questions that require research, reading the web, or understanding the results of a long conversation.