While that's undoubtedly true, is that really feasible?

Training programs are expensive, and i imagine difficult to conduct across potentially remote areas with underdeveloped infrastructure.

Internet access is maybe more doable now with starlink, but how practical was it at the time? I imagine this varries significantly with region, maybe in some cases all that was needed was LTE modem -> wifi, but if actually new infrastructure needed to be set up, that could be very pricey very fast.

Like everything its all about trade offs, if olpc did those things would they have budget for other things?

the only way to improve education is to train more and better teachers. that's even completely independent of projects like OLPC. asking if training teachers is feasible is simply the wrong question. arguing that it is expensive is the wrong argument.

education and teacher training is the only way to achieve progress in this world. and if training is expensive or difficult to achieve then that's a challenge we need to overcome, not an excuse not to do it.

There are viable alternatives to Internet access, notably offline resources as provided by projects such as Kiwix. I'm not sure to what extent the projects described in OP actively leveraged these efforts with any real effectiveness. If it didn't, those OLPC mini-laptops would've been functionally equivalent to glorified calculators, and the results would be quite unsurprising.