Augmend

It has been an honor 🫡

RPA is dead. Long live RPA(gent) Everyone is talking about agents. I think of them as simply AI + tools + a for loop but wanted to get my notes down to share. Calling out in particular a need for more human/agent collaboration tools starting with WatchMe. Read more at the Augmend blog. https://lnkd.in/ger2zmfQ

TreeSeg on the orange site front page!

🔥

Our new video understanding API is live! With a few lines of python code, you can send a video to our service and generate documents, tags, title, and more. Great for technical videos, it uses what people say *and* what's on screen together! Or use Zapier with no code needed! Check out more info on our blog: https://lnkd.in/gtDVrkcB

At Augmend we believe it's important to share what we learn, and it's always nice to sit down and writeup something novel others might find useful. Preprint of a new meeting/video segmentation technique we worked on a few months ago that we're calling TreeSeg: https://lnkd.in/gJ68fvKv. Dimitrios and team will be sharing a blog post and repo in the next week for those interested. The chapters created from this improved segmentation will also be available as part of our upcoming API for any video you submit. - Abstract - From organizing recorded videos and meetings into chapters, to breaking down large inputs in order to fit them into the context window of commoditized Large Language Models (LLMs), topic segmentation of large transcripts emerges as a task of increasing significance. Still, accurate segmentation presents many challenges, including (a) the noisy nature of the Automatic Speech Recognition (ASR) software typically used to obtain the transcripts, (b) the lack of diverse labeled data and (c) the difficulty in pin-pointing the ground-truth number of segments. In this work we present TreeSeg, an approach that combines off-the-shelf embedding models with divisive clustering, to generate hierarchical, structured segmentations of transcripts in the form of binary trees. Our approach is robust to noise and can handle large transcripts efficiently. We evaluate TreeSeg on the ICSI and AMI corpora, demonstrating that it outperforms all baselines. Finally, we introduce TinyRec, a small-scale corpus of manually annotated transcripts, obtained from self-recorded video sessions.

In honor of our upcoming API beta at Augmend that lets anyone add automated video to document generation to their apps and agents we've revamped the homepage, check out augmend.com. There was an intervention to get rid of my "excessive" neon glow drop shadows (can you really have too much cyberpunk neon?), leave a comment if you of course want it back 😂

"A computer for every AI Agent and any task". Not quite as catchy as the Microsoft PC revolution, but possibly more important. I recently gave a small presentation on where the Human/AI Agent collaboration space is lacking and what we're doing to build the Agent PC Revolution. TLDR: Agents/Assistants need their own VMs or PCs, not to be time sharing with you on your desktop. We should be able to share screens with each other, just like we were working with a human. Watch the full talk in our intelligent video player to get all the detail and ask questions: https://lnkd.in/gH-BMizw or on youtube if you're on mobile: https://lnkd.in/gjXgfhaj

Big update! Turn any video into structured, queryable, docs in 1 click, with the Augmend AI assistant. Comment with a youtube or loom or just a gdocs link to a video and I'll turn it into a doc + intelligent Q&A link right now. Drop one below and let's try it out together. Have an existing loom from a brownbag that needs to be written out for your team's wiki or technical docs? Want step-by-step instructions from a youtube tutorial? All done for you with no extra work, automatically handled by Augmend. An AI assistant that takes video and turns it into multi-modal documents is a game changer for learning (and not having to write docs 😅). Everyone is building summaries and "chat with" experiences off of the voice transcript only, our new system has eyes making it way smarter!

📣 Calling all AI enthusiasts and investors! 📣 Google for Startups Accelerator: AI FIRST DEMO DAY is your front-row seat to the future of AI innovation. 🚀 June 13th – Join in as 15 game-changing AI startups pitch their groundbreaking solutions in healthcare, agent tech, cybersecurity, and beyond. We'll be LIVESTREAMING to you around the world – live from Google HQ. ➡ Register now: https://lnkd.in/gQ5w9DrK ⬅ Please share & repost w/ your networks! Aptori, Augmend, Backpack Healthcare, BrainLogic AI, Cicerai, CLIKA Inc., Easel AI, Inc., Findly (YC S22), Glass Health, KODIF, Liminal, mbue, Modulo Bio, Rocket Doctor, Sibli #GoogleforStartups #AIFirst #DemoDay #AI #Startups #Innovation #VentureCapital #acceleratedwithgoogle