I think the fundamental issue is there's too much stuff, if item A costs money, items B-Z are also available for free.
Bandwidth costs appear to be 0 now, you can pay to listen to music without also downloading a video, or download the video to listen to it for free.
It's been a while since I've read it but Shirky's original piece most focused on the mental transaction costs with deciding to pay a nickel (or whatever) to read an article. Assuming you accept that argument--which I'm largely inclined to--nothing has really changed. The arguments against micropayments have never been primarily about technology. Sure, you need a way to make low-cost transactions very low friction and very low cost overhead but that's the easy part.
Subscriptions seem to be the way paid content access has mostly headed. The problem with making subscriptions more broadly apply to online text content is that you need to get a critical mass for a platform and you need to deal in some way with the standard issues of paywalls.
Meanwhile, the less money you have, the more you pay! Smaller bundles cost more per GB, and many of the Kenyans I knew would regularly buy 150MB top-ups due to lack of upfront funds.
On AirTel in India, it looks like they charge about $3.88 for a 1GB bundle good for 28 days (there's a cheaper 1GB bundle good for 1 day, however, and some other deals available). There are also a few articles floating around on the _increasing_ cost of data in India, suggesting that companies started with cheaper data while establishing themselves in the marketplace, and over time have been more obliged to turn a profit instead of burning investor money on subsidized plans. Cambodia does look pretty cheap, though, I agree.
We can see this with the App store where people complain bitterly about $0.99 or $1.99 prices.