Others have addressed the obvious privacy issues, so let me address your logic on the business side of things.
I apologize in advance for the tone, but your move with exposing profiles made me angry. Good thing my TB profile is fake (and performs worse than my actual, real life resume despite all the embellishments)
1) There is no lock in - I can move on and off LI whenever I want, and have. I've exported my data and used it to create my own resume site with analytics that I send out to companies. I can see who viewed my CV, when, and whether or not they actually read through it or bounced immediately.
I've also learned to track the progression of my candidacy through the organization using this trick (recruiters tend to view my CV on their Windows desktop during work hours, hiring managers tend to check out resumes in the evening on their iPhones or Macbooks, engineers/tech leads tend to use Macbooks, desktop Macs or Android phones in the morning or during lunch time. Usually when I've hit the engineering lead I tend to get invited to interview).
It's extremely easy to create your own CV website for free (github/lab pages) that's versioned by git and deployed automatically using a CI script.
2) You're attacking the tech hiring problem from the wrong angle, like everyone else. There is no issue with discovery of candidates and employers. LI and stackoverflow, etc do a great job of approximating this O(N^2) exposure process, the filtering and sifting.
The ACTUAL problem is on the hiring end - companies won't take a chance on non-traditional candidates (not talking about race and gender here, more about credentials).
You have to start by chipping away at the costs of showing competence for a candidate (the traditional way to do this is to get a three- or four-year degree that's either expensive in terms of time and money, or useless, and if you get a degree with a low score, doubly so, even though you might be a better programmer than the people who scored over 90%).
This will only happen by convincing hiring orgs to hire non-traditional candidates, and this requires establishing a very strong signal/noise ratio for candidates coming from your hiring channel.
Before you start PRing me about how great TB is at this - no it isn't. Not any better than leetcode etc, and those are terrible at predicting engineering competence.