> In reality that data will sit there unreviewed in 99.9999% of cases.

I believe AI and LLMs will change that, unfortunately.

Believe it or not we’ve had the technology a long time to “look” or “track” everyone all the time. At this point it’s simply a political decison not to.

The leadership of the PRC with help from sensetime and other industrial leaders have a working and well tested comprehensive domestic individual tracking system in production. It’s been working well over a decade at this point.

Palantir is politically involved enough that I don't think we can raise any hackles about PRC that don't apply equally in practice to USA.

> Palantir is politically involved enough that I don't think we can raise any hackles about PRC that don't apply equally in practice to USA.

Absolutely absurd statement. Palantir doesn't run concentration camps.

The US governement kept that contract in house.

Please point out easily found reports of the US governemnt kidnapping citizens and then torturing them and daring them to say something about it like happens on the regular in China and Russia and N. Korea. So yeah sorry, not the same. Do I think TLAs spy on US citizens without warrants or any good reason at all. Absolutely, but we aren't quite up to CCP/Russisan standards just yet.

I joke, ruefully, what would happen if we had another Stalin or Mao now. Vacuuming up dissidents would be easy - for many of them it’s probably been on their Facebook at least once. And if it’s not, check what people bought from certain businesses. Then use AI to analyze and weed out the rest.

>> In reality that data will sit there unreviewed in 99.9999% of cases.

> I believe AI and LLMs will change that, unfortunately.

Maybe. It'll be a race between one set of AI to generate lots of content, and a different set of AI to analyze it all.

It's not obvious to me that the side doing the analyzing will win out.

Only now (or soon) you will be able to instruct an agent to read, comprehend, and intelligently categorize any information you find on any individuals, at scale. You can't even hope to be in the 99.99999%.

> Just hope you’re not in that 0.0001%

Who already is in that 0.0001% and how have they already been abusing that data?

> If you create data anywhere, assume some govt will eventually get a copy

I have a hunch that we've been giving these agencies too much credit this whole time

Every time there's an unmitigated act of terror makes me think: the intelligence community is either malicious (a crazy rabbit hole to go down) or incompetent

And also there's the problem of searching, like finding a needle in an ocean

no one is manually searching through petabytes of data. it's all ingested into ML systems which spit out inferences and predictions about individuals. repeat with enough people, combine datasets, now you're generating increasingly accurate predictions about society at large. less privacy = more data = better training = more accurate predictions.

Or It’s an impossible problem

You can’t win 100% of the time, so you’ll miss some and people will be rightfully mad you missed

You also can’t talk about most wins or - more frequently - events prevented/mitigated

Finally, it only takes a few bad actors in the govt to destroy long term public trust and we’ve had decades of lackluster at best leadership

So add all that up and there’s never going to be positive news about the IC