New AI can condense anything into a bullet-point Summary.

The days of sitting through a lecture, Reading a 500-page book, and Joining a 4-hour zoom meeting, all to hear/read what could’ve been “AN EMAIL”, are coming to an end.

Thanks to Assembly AI, AI will now be able to

Condense a video down to the key takeaways.

Create podcast summaries.

You can customize the summaries to your preferred structure

Summarize any downloaded file or youtube video

Assembly AI already has an API that transcribes videos and audio, which allows for summarization, as the summarization model simply summarizes a transcribed/text version versus audio.

They don’t give a very clear technical explanation for how their model works, but this is what they said:

Our newest Summarization models are built on the same AI technology (Transformers) that are behind most of these advances [generate high quality, realistic images and art with a single text prompt. Transformers have taken the world by storm and opened up the opportunity for Large Language Models (LLMs) that can generate code and write articles.] We’ve purpose-built our Summarization models to work extremely well on conversational data (phone calls, zoom meetings, screen recordings, videos, etc.).

Now, don’t expect to be using this tool tomorrow.

It has been taking me a few hours to run even the most basic version of this API.

It is very promising, nontheless.