OpenAI’s DevDay event was a game-changer, rolling out some cool GPTs / Assistant API that’s set to shake up how we chat with clients. These tools are making complex data stuff way easier, using chat-style talks that feel super natural.

Key to this innovation are these three critical components: the knowledge retrieval tool, the code interpreter tool, and a versatile function_json interface allowing conversation words to trigger certain functions being called, which, when combined with a diverse range of APIs—encompassing functionalities like screening, time series analysis, Performance Attribution (PA), and optimization—lead to ultimately the desired output our clients seek.

This means clients can just chat with our systems like they’re talking to a friend. They can upload their raw files, ask for stuff in plain language, and skip the headache of figuring out complex formulas or messing with loads of PA settings.

Guided by this vision, I think our focus should be on collecting unique and valuable data and developing a range of effective and functional APIs tailored to our clients’ requirements.

At the end of Nov, 2023, my testing creation of GPTs and GPT with APIs are proved to be applicable. The obstacle is the authorization from the firm.

My next project should be robots now.