Yes, that, accuracy, speed, and single computer-use.
I find those to be the limiting factors to speed.
I have extensive rules, I do extensive planning. Yet at implementation, the rules are not respected, errors are introduced, etc...
I spend more time fixing than writing code.
Then speed... Because of the fixes and bad code quality even with frontiers model speed makes a very big difference. I (agents) spend hours daily doing reviews and fixes. 5x speed boost would make me much more productive.
And when working super fast with agents, having only one computer is limiting. Even worktrees don't solve problems because I use things like convex, chrome use, etc... and it conflicts with each others all the time.
Still many problems to solve. It's already evolved so much in the last two years.
Also, I believe that if Apple is working on blood glucose estimation, it's probably not the way described in this article.
I could be wrong that Apple will do it this way, but the current hardware supports different wavelengths PPG, (as well as the Samsung Watch Galaxy 8), which might be able to use to estimate blood glucose variations. But it's pretty complex, especially to generalize.
I wouldn't be surprised if they start like blood pressure, just estimations as to how it's going up and down.
I know that because I'm literally working on this with https://trackourhearts.com, using an AI model I created to estimate blood glucose, though I'm doing it in a personalized way so I can provide more absolute estimates but it's still incredibly hard from PPG as the wavelengths available aren't the most optimized for this.
Oh yes, I've been actually looking at integrating this and other mmwave research into my startup https://trackourhearts.com
Non-intrusive technology which can work at home to monitor people' vitals is a game changer, there are so many applications to this. Research is at the beginning.
Indeed there are privacy issues with big providers doing this, but then this really opens up so many possibilities if done well.
I find those to be the limiting factors to speed.
I have extensive rules, I do extensive planning. Yet at implementation, the rules are not respected, errors are introduced, etc...
I spend more time fixing than writing code.
Then speed... Because of the fixes and bad code quality even with frontiers model speed makes a very big difference. I (agents) spend hours daily doing reviews and fixes. 5x speed boost would make me much more productive.
And when working super fast with agents, having only one computer is limiting. Even worktrees don't solve problems because I use things like convex, chrome use, etc... and it conflicts with each others all the time.
Still many problems to solve. It's already evolved so much in the last two years.