A couple of weeks in the past, I returned to sunny Honolulu for TeleGeography’s annual telecom tendencies workshop at PTC.
As standard, Brianna Boudreau explored world pricing tendencies, and Jon Hjembo shared insights from the info middle world.
A bit uncommon was the title of my phase: Mom Earth, Motherboard. This title was really impressed by an article written in 1996:
This clip explains the imaginative and prescient.
Working with this theme, I additionally used AI to create a picture for every part of my presentation:
World Local weather Change appears significantly intimidating, so let’s bounce there. (You’ll discover my slides from Cycles of Renewal and A Resilient System over right here.)
World Local weather Change
Evolution
How AI goes to impression worldwide networks is a scorching, scorching subject nowadays.
And the reality is, nobody actually is aware of what AI will imply for long-haul networks. Nobody can inform you it’ll result in X % extra bandwidth development over X years—that’s unknown at this stage.
What we can do is focus on how AI works and what may affect the tempo of bandwidth development, similar to:
- Mannequin coaching places
- Inference places
- Distributed coaching and federated studying
- AI-based knowledge compression and visitors routing enhancements
- Spatial temporal load shifting
- Authorized points
To be taught extra concerning the components shaping long-haul demand from AI, obtain my full presentation.
