> It's also a bit presumptuous to assume that a future authoritarian regime wouldn't have expert technologists at their disposal...
I think your thinking here is too binary. Clearly there exist possible futures where authoritarian governments are hyper-efficient and have brilliant technologists advising them. In those futures we're clearly doomed and thus we might as well give up now. The thing is, those futures are not inevitable. There are also many (IMHO more realistic) futures where governments are messy and inefficient (like they are today), where their authority is blunted by organizational and jurisdictional issues (as it is today), where their technologists are not hyper-competent (believe me, as they are today.)
In those worlds there is a huge difference between a scenario where the full weight of FAANGM's resources is pushing to build massive data repositories, and one where they've taken firm technological steps to limit it. We are in the first world and we should be in the second one.
> Issuing an edict to "do things like you did them in 2022" doesn't exactly require genius-level administrative talent...
I think that this will actually be more difficult than you think: this is why governments are spending ~millions right now to slow down the deployment of new encryption technology [1]. But stop worrying about 2022: what you should be worried about is 2036. Look at what we've done to privacy since 2007 -- the year the first iPhone launched. Now imagine someone from ~14 years in the future coming back to to explain what Silicon Valley has done with even better technology like wearables and powerful ML tooling. When you're in a hole, the most important step is to stop digging.
I think you massively under-estimate the banality of authoritarianism.
If an authoritarian movement takes over the US government, at least a majority of the Johns Hopkins CS faculty will continue taking grants from the NSF/DoD. Many of those grants will be more-or-less aligned with the objectives of that authoritarian movement. Non-authoritarian students will grind away on those projects.
Something similar would happen at FAANGM. No iCloud backups? NBD; lean on those companies to collect whatever data the state wants. You don't need super competent loyal technologists, because FAANGM and their employees will most of the time just do what you tell them. You don't need existing troves of data, because you can start collecting at any point and still get a huge amount of utility.
Could the authoritarian world be marginally better if big tech makes an about-face and stops collecting data? Sure. Is that difference enough to make any sort of significant difference in the lived experience of people or the trajectory of the authoritarian regime? Probably not.
I don't think you're wrong, per se, about the risks. But I don't think you have a compelling solution. And, anyways, there are much stronger arguments for reigning in data collection at big tech than the risk of impending authoritarianism.
Right? This comment thread reads like this sort of thing hasn’t already happened here. But it has. The US may not be the most authoritarian regime, but I think that its recent actions scream authoritarianism louder than any words claiming that it’s not.
I think your thinking here is too binary. Clearly there exist possible futures where authoritarian governments are hyper-efficient and have brilliant technologists advising them. In those futures we're clearly doomed and thus we might as well give up now. The thing is, those futures are not inevitable. There are also many (IMHO more realistic) futures where governments are messy and inefficient (like they are today), where their authority is blunted by organizational and jurisdictional issues (as it is today), where their technologists are not hyper-competent (believe me, as they are today.)
In those worlds there is a huge difference between a scenario where the full weight of FAANGM's resources is pushing to build massive data repositories, and one where they've taken firm technological steps to limit it. We are in the first world and we should be in the second one.
> Issuing an edict to "do things like you did them in 2022" doesn't exactly require genius-level administrative talent...
I think that this will actually be more difficult than you think: this is why governments are spending ~millions right now to slow down the deployment of new encryption technology [1]. But stop worrying about 2022: what you should be worried about is 2036. Look at what we've done to privacy since 2007 -- the year the first iPhone launched. Now imagine someone from ~14 years in the future coming back to to explain what Silicon Valley has done with even better technology like wearables and powerful ML tooling. When you're in a hole, the most important step is to stop digging.
[1] https://www.eff.org/deeplinks/2022/01/uk-paid-724000-creepy-...