I’ve known Vladimir for over one decade. We first came in contact through his interest in WebRTC. Back in TADHack Paris 2015 Vladimir showcased his hack Thermal RTC. He linked a thermal camera to his phone using WebRTC, and showed a neat thermal image of the room.
He also did a great hack at TADHack Chicago in 2015 mashing up Matrix.org + WebRTC + Robotics. It was a small robot with telepresence functions controlled over WebRTC data channel peer to peer. Then in 2016 at TADHack Global Vladimir did an accessibility hack, Accessible video colors, using technology to help help color-blind people see more colors.
We’ve been sharing insights on voice assistants and generative AI recently. For example, I sent him the Vapi.ai information to get his view.
Last year Vladimir shared a white paper co-written by Amazon Alexa and Panasonic Automative on “Multiple Voice Assistant Dialogs and Arbitration” apart of the Voice Interoperability Initiative Architecture Series Whitepapers. The VII is focused on improving the interoperability between different voice services on a single device, with the goal of providing consumers with more choice and flexibility in how they use voice assistants:
- Customer Choice — Building voice-enabled devices that promote customer choice and flexibility through multiple, simultaneously available wake words.
- Secure Interoperability — Developing voice services that work seamlessly with others while protecting the privacy and security of customers.
- Technology Solutions — Releasing technologies and solutions that make it easier to integrate multiple voice services on a single product.
- Research and Development — Accelerating machine learning and conversational AI research to improve the breadth, quality and interoperability of voice services.
We then geeked out on all the work taking place on interoperability of agents.
Vladimir introduced the Open Voice Interoperability (here’s the github) initiative letting multiple GenAI-based agents potentially, accessible over a speech-based UI (User interface) interop.
Vladimir was just back from Google Cloud Next ’24 here are links to the 2 sessions he highlighted:
- Local government genAI use cases: https://cloud.withgoogle.com/next/session-library?session=AIML116#day_2.
- Use of genAI in software engineering https://cloud.withgoogle.com/next/session-library?session=AIML203#day_3
- In addition here’s a session on New Hampshire Department of Employment Security, is leveraging AI to improve efficiency and accuracy in unemployment benefit decisions
Overall, we’re in a period of rapid technology development, however, realizing the benefits of these technologies will take time and is use case dependent, e.g. accelerating software development requires work across the workflow. While removing a backlog of applications using agents is doable today. For more on Vladimir’s findings from Google Next check out this post.