Tom Lindfalk’s Post

View profile for Tom Lindfalk, graphic

Automate your way to growth | Ex MD @ Flight Story

Spend 5 minutes on Hugging Face and you'll realise that the future of AI is already here 👇 ChatGPT (& other well known models) are general purpose and designed to be used by the masses. They are very good at a wide range of tasks - but in some cases they may fall short of your specific use case. Hugging Face is an open source library containing thousands of different models, some of which have been trained to be much more competent at specific use cases than models aimed at the general population. 🗣 Evaluating sales calls and want to know what is working vs what isn’t? MIT have developed a model that can classify audio interactions - detecting emotions, satisfaction levels or urgency in the customer voice. 🤖 Building a chatbot that requires faster responses than GPT-4o? Microsoft have their own mini-instruct model which can be far more efficient. There are almost 1 million models on the Hugging Face platform right now. More often than not, GPT-4o will be more than sufficient for your own use case, but a quick scan of the platform may lead you to new areas and new possibilities.

To view or add a comment, sign in

Explore topics