Big News: GPT-4o Boosts Coming to SpeechToNote! 🚀

Enhancing the experience of Note-Taking with OpenAI's GPT-4o

I hope this email finds you in great spirits!

We have some thrilling news to share with you guys. OpenAI recently announced the launch of GPT-4o, and its performance has truly exceeded expectations. It's been an exciting 72 hours since the announcement, and we've had the opportunity to test the new GPT-4o Text Models. The results are remarkable!

GPT-4o is here… 🤩

Efficiency Improvements Across Languages

One of the standout features of GPT-4o is its enhanced efficiency, especially in languages other than English. For instance, generating a 2-minute summary from a transcript in Marathi used to take about one to two minutes but now takes only 30-45 seconds. This is a significant leap forward in performance!

Here are some of the top languages with marked improvements (efficiency and token utilisation from GPT4-Turbo):

  • Gujarati: 4.4x

  • Telugu: 3.5x

  • Tamil: 3.3x

  • Marathi: 2.9x

  • Hindi: 2.9x

  • Urdu: 2.5x

  • Arabic: 2.0x

  • Persian: 1.9x

  • Russian: 1.7x

  • Korean: 1.7x

  • Vietnamese: 1.5x

  • Chinese: 1.4x

  • Japanese: 1.4x

  • Turkish: 1.3x

  • Italian: 1.2x

  • German: 1.2x

  • Spanish: 1.1x

  • Portuguese: 1.1x

  • French: 1.1x

  • English: 1.1x

While the efficiency in English, Portuguese, and Spanish has also seen improvements, even if slightly, it’s truly remarkable to witness OpenAI continually push the boundaries of performance and output quality. We couldn't be prouder to integrate these advancements into STN.

Rolling Out GPT-4o Enhancements

We’re excited to announce that Team Speech to Note is rolling out these enhancements today. The upgraded summarization features and reduced processing times will be integrated into your STN account, making your experience even more seamless and efficient.

While we haven’t yet received the audio models from GPT-4o, we’re eagerly awaiting their release. Compared to the previous Whisper V3 model, the new audio models promise significant improvements, and we’ll roll them out to you as soon as they become available.

Post of the Week

This one is from one of our community members, Cam. He has always backed us and this was a kind of message which made us value our community even further. Thanks for this Cam.

Your Thoughts on Other LLMs?

In addition to GPT-4o, we're exploring other LLMs to make STN multimodal. This means you'll soon be able to choose from multiple LLMs for your transcript or summary generation.

We’d love to hear from you! What other LLMs are you excited about?

Have you tried any models besides OpenAI's?

Do you think we should integrate them into Speech to Note?

Reply to this email and let us know your thoughts.

By the way, as a STN user, if you aren’t part of the community yet, then that’s the place where you can be to be one of first in the action.

Check out our community on telegram

OR

Join our private Facebook group

Thank you so much for being with us and for your continued support. We're committed to bringing new and exciting features to SpeechToNote with the same passion and enthusiasm.

Drop a comment on this newsletter or simply reply to this email—your feedback is invaluable!

Cheers,

Abhishek.

P.S. Stay tuned for more exciting updates. We’re just getting started! (At least that's how we always feel)

Reply

or to participate.