The 10x factor
Just listened to this podcast episode on the tinyML ecosystem. In it, Evgeni Gousev, chairman of then tinyML Foundation (now Edge AI foundation) talks about the state and future of edge machine learning.
He mentions three key learnings, drawing from his extensive career in both industry and academia:
- There is always a solution
- Aim for the 10x factor
- Look at your solutions holistically
All three resonate with me, especially "Aiming for the 10x factor". Evgeni explains that working with disruptive technology is difficult and takes time, and the state of the art improves as you work. In order for your solution to be relevant when you're done, it's important to aim high. When you finish, you're lucky if you end up with 2x or 3x compared to state of the art.
My recent lithium project would have benefited from appropriate metrics, to determine to what degree digital lithium blood sample referrals improved on the old ways.