Can tech end poverty?

Today I had a fascinating talk with someone who spends a lot of time thinking about how to get technology to the poor that will actually be useful to them (she's on the board of a fairly prestigious institution and therefore has to speak carefully in public so she asked me not to use her name casually).
I'm still digesting this, but our topic was basically: can technology help the poor?
Her answer: Yes, but it's tricky.
And it's tricky in some interesting ways.

The main problem is that - when it comes to tech - the market doesn't work for the poor in the way we normally think it should. Ideally, things start off with an innovator seeing some need, developing a product and marketing it. And then it competes with all the other products, and if it really does help the poor, they buy it. So the poor, the people who understand best what technology they need have a hand in drawing it in. But, my source says, it never works out that way. Instead of a pull from the poor, products must be pushed by the rich. Problem #1 There's almost never competition between technologies - at best you have one alternative to the traditional single choice. This is because the only people who are in the business of developing expensive technology for the poor are governments and charitable organizations. They are not going to pay all the development costs to get a competing vaccine out there if someone else has developed a perfectly good one. Problem #2: If you are really poor you can't afford risk. Experimenting with new things doesn't make sense when the a failed experiment means the death of a family member. Problem #3: Market surveys don't work. My source says that attempts at asking subsistence farmers what they need have failed. People focused on staving off hunger are not in a position to imagine how technology might improve their life, but they are often good at figuring out what you want to hear and telling you exactly that.

If all this is true, it means that any attempt to bring technology to the aid of the poor is paternalistic. That's kind of a bad word, and for good reason: Sometimes when we tell the poor what they should do we end up making things worse. But there are many cases where the Pater of paternalism has proved to be a wise and loving father. Spend enough time on the ground working with people and tinkering with the technology and you can make a difference.

I left that initial conversation with the feeling that the whole enterprise of trying to develop good technology and push it to the poor is incredibly sticky. First - there are the difficult Kipling overtones. Second - whenever you are successful at introducing a new technology you are taking a bit of control away from the non-human world and bestowing it upon humanity. (You could define technology as that which gives humans more control) The question is, which humans seize that control and wield it for themselves? Hopefully it's the people you are trying to help. But what if it's the corporation manufacturing the product? There is a real danger of inadvertently furthering corporate colonialism. If the devil is in the details, so is the answer. I hope to learn the details of some examples that have worked.