There's many problems in OLPC right now, but scaling successful deployments is a biggie. Basta botón para muestra (a button will suffice for a sample), since together with this one specific problem most of what's wrong with OLPC has one central point:

The exclusion of the community from policy making is what is keeping OLPC from really taking off.

A major source of discontent around those who want to have OLPC become real in their own neighbohood is the minimum number of XOs that OLPC will take a donation for (it appears OLPC does not "sell" them, IRS technicality). The canonical example is Chile, where 900 was not enough.

Not a sustainable deployer

The Sur list (Pilar Saenz from Colombia in particular, kudos) seems to have found why it is so. OLPC decision-makers believe the only way to start a deployment is to send a Cambridge-designated expert in, or supervise things from CC1. Thus, obviously and quite sensibly, you would not want deployments that are so small that they would not pay your expert, who not only is an expert, but an American expert (more $$$)

No matter this is totally inconsistent with Nicholas' sola machina view that the XO itself will bypass "experts" (aka teachers), providing a better learning experience through the Internet than they would get through their teachers, even though, for example in Uruguay, those are rather highly trained. And a teacher-centered approach is a key to success.

It all bottom lines into that, apparently, the community cannot be trusted, and we need someone else (assigned by OLPC headquarters) to teach us how to do things right

Which explains how hard it has been for me to convince Rodrigo Arboleda, thanks but no thanks, against his constant insistence that I need to make the deployments in Bolivia larger so that he can justify sending his people in.

Now, I am sure some of those people are swell, I might even have met a couple of them, though not getting response to my emails has been my experience in trying to connect with some. And I am sure Mr. Arboleda sincerely believes this is the best path, not out of any evil intent but just because he has not been given the right information.

The real problem remains: even if that were a good idea (and it is not) such a deployment model is expensive, and it cannot be scaled up or down. Imagine the wildest dreams of Dr. Negroponte, and 4 or 5 governments "get the vision" and want 10 thousand XOs tomorrow, to get to a million each by the end of the year. Who would you send, if the only right way is experts?

Pilar, responsibly as the best of us, when she finds a problem, she offers a solution. In her own words:

Habrá que mostrarles que mantener centralizado el proyecto es lo que más daño le esta haciendo y que nunca podrán responder de forma adecuada si no construyen capacidad de soporte y respuesta local. Algo que perfectamente pueden asumir los grupos de la comunidad si hubiera la voluntad y la comunicación necesaria.

'We will need to show them that to keep the project centralized is what has been doing it the most harm, and they will never be able to respond the right way if they do not build local support and local response. This is something that perfectly can be done
by community groups if the necessary willingness and communication were there'

OK, to make it simple so it can be agreed upon by all'uns, when Dr. Negroponte comes in on Wednesday to hear the community proposals, please pass him the following:

1. community-based solutions are scalable, up or down
   Once you get going, you have options that will serve well the G1G1 1-family deployment of one XO, and the 200.000 Uruguayan one, and anything in-between mostly anywhere (I agree some artisanal work might be needed for some out-of-the-way localization, but the community is often better prepared to do that too. You may even hire a community-connected expert or consultant (Hi Michael!), who will help you connect to the community as his first act)
2. community-based solutions are better
   By drawing in the expertize of the many, no one is left alone to his own wits, and unless it is a proud and silly person, he will seek help when he needs it, and get it, because it is there, offered freely by the OLPC community.
3. community-based solutions are cheaper
   it eventually becomes a self sustaining process, no need to host training meetings for local experts to send out, no need to bring over to Boston local people who will then be the patronizing experts in their own countries. Yet you do need to provide some basic services, like a wiki, listserv, conference calls, and a few actual employees who are there 24/7 to serve the community. This is totally peanuts when comparing costs, and we already paid for that zillions of times over for all eternity with what the community already has contributed through G1G1. And the local community is right there to help things along, growing in its own skills and abilities, able to better troubleshoot than anything else, and having more buy-in than anyone coming from abroad, however good those might be
   

Yama Ploskonka is a representative for Bolivia and Chile with Open Learning Exchange, but who started as just one more member of the OLPC News community

Postscript

In honor and apology to the gentleman from Costa Rica, connected with several nonprofits, who tried to visit 1CC a couple days ago and was not allowed in the building, the Ploskonka household is having Costa Rican Tarrazu for their morning coffee today.

I am sure he will make a great community member, by the way, and he was not left too sore by the experience, which is a good thing, of course after a volunteer spend 25 minutes on the phone with him.