>But I have to wonder whether the internet connection can be maintained during cloudy days and what the expected upload/download speeds will be and finally what the expected costs will be.
Depends on the size of the constellation. Ultimately I think they want to have 3 sats visible at any given time which should be enough. The bands used can penetrate though anyway. I've seen numbers quoted of speeds up to 1gbps, so basically "good enough". I'm currently using a 4 mbps connection just fine and the fastest I have access to is 8 mbps down 0.5mbps up. And I'm getting my browsing done just fine. Costs are up to discussion, the main selling point of Starlink is backbone and the receiver was according to Shotwell one of the main research points to drive the cost down( 1k$ at the time but an obscure tech, they want it down to 300$).
>Affordable and globally available internet could be a game-changer. If viable, couldn't it challenge wireless carriers and ISPs?
No because of density limitations. They can't support enough bandwidth for an entire city. Also direct LoS is required so big buildings will limit you.
>Also, aren't there geopolitical ramifications. Would China, Russia, EU, etc allow their citizens to access the starlink system? Or will starlink have to be censored, filtered and monitored in these regions?
Current sat internet providers just don't sell their receivers in China or any country that doesn't allow them.