Wordtune Read Offers Enterprise Solution for TL;DR
Linguistic AI startup AI21 Labs has launched a tool that will read, comprehend, and summarize long documents in just a few seconds. AI21 designed the new Wordtune Read product for commercial and enterprise use, offering a more sophisticated solution to the common digital complaint of ‘too long, didn’t read’ (tl;dr).
Wordtune Read
Wordtune Read analyzes text submitted directly, in a pdf, or from a link. The AI then starts identifying important sentences and phrases, building up a summary of themes and simplifying language where relevant. The tool also comes with a Spotlight feature that will prompt Wordtune Read to reexamine the text and come up with a new summary with a different emphasis.
The company pitches Wordtune Read as an ideal solution for saving time when going through lots of long articles and documents. Wordtune has a free online demo that a couple of experiments found to be quite good and concisely summarizing long articles, even adjusting to different angles with Spotlight. It starts to struggle with more specialized terminology, but the company claims repeated use will help teach the AI what is important to the user.
Summarize AI
“Our mission at AI21 Labs is to fundamentally reimagine the way people write and read, so we thought it was high time to share our vision for the reading part of that equation. Underpinned by our sophisticated language models, Wordtune Read was specifically designed to help professionals across a range of industries and academia navigate the issue of information overload,” the startup explained in a blog post.” We built Wordtune Read specifically to make it easier for people to consume information more easily – so whether you’re doing a competitive analysis, market research, or writing an award winning scientific study, Wordtune Read has got you covered.”
AI21 created Wordtune Read as part of its larger work developing language models, including a potential GPT-3 rival called Jurassic. The Tel Aviv-based startup has raised $54.5 million, including $20 million at the beginning of November. Understanding and summarizing text is one of the many potential uses of this kind of AI, one that OpenAI likely expects to see now that the waitlist for GPT-3 licensing has ended. A growing number of companies are already working toward similar goals. For example, IBM upgraded Watson in September with new virtual agent skills for call centers. The AI can now perform short-answer retrievals that will summarize information pulled from databases, among other improvements. Alexa used to do the same with email inboxes until a recent decision to end the feature enabling the voice assistant to read out emails.
Follow @voicebotai Follow @erichschwartz
