I get the general skepticism and how this gives anti freedom vibes, but wouldn't this also prevent some actual rootkit like sideloaded apps stealing credentials?
Not deep into rooting scene but seems plausible to me that this has some merit if you squint at it from the right angle
Here's the fundamental problem:
Trusted agents are useful. And I'm using legal meanings, not technical meanings here - so a "trusted agent" is someone or something that is legally acting on your behalf, to perform actions as though you were performing them.
The whole fucking promise of "general purpose computing" is that citizens should be able to delegate repetitive and tedious tasks to a computer. And they should have the full freedom to pick both which tasks are delegated, as well as which agent (program) is performing them.
Instead - what we're seeing is that companies are closing off as many avenues of automation for the average citizen as possible, under the guise of security.
The problem is that selecting a neutral (trusted!) agent is really, REALLY important, and companies are absolutely not neutral. They don't want the best results for "average Joe customer", they want the best results for themselves: the company.
They will make decisions that are contrary to your best interests all the time. They have exactly zero fiduciary duty to you, and boy do they know it.
In a decent world - in a decent CAPITALIST SOCIETY (which we can already debate the decency of in the first place) you allow space in the market for modification. Ex - don't like your desk? Change it. Don't like your car radio? Change it. Don't like that tool handle? Change it. Pick a different one, even one from a totally different company. Replace it.
This allows new ideas, new growth, and prevents stagnation.
In the digital world... there are a few companies that are trying as hard as possible to prevent you from being able to change anything.
---
Want a new browser? Fuck you.
Want a different UI for your banking needs? Fuck you.
Want to watch something without the ads? Fuck you.
Want to watch something with the ads, but in a less miserable ui? Fuck you too.
Want to automate something? Fuck you.
Want to export your data? Fuck you.
Want to sell software without us taking our rent money? Fuck you.
Want to shop in a different store? Fuck you.
Can't be letting our users make decisions that might cost us money.
---
So we're seeing an absolutely insane number of "digital locks" being employed not to protect users. No - instead they're getting deployed to protect revenue at the expense of users.
The only possible outcome is that service quality degrades to the point where you literally are better off without. Because that's what happens to incentives when you let companies operate in this manner.
If the consumer has no choice - the market has no power, and what little value there is in capitalism goes right into the trash bin.
So sure - if you squint, this maybe prevents someone from making a bad decision on which agent they trust.
But the problem is that now they HAVE to trust an agent they know is going to make bad decisions for them. Hope you like the biggest ad company in the world owning you digitally... Serfdom here we come.