Comment by crystal_revenge
1 day ago
I'm never sure why people assume that Palantir is magically unlike the overwhelming majority of tech startups/companies I've worked at: vastly over promising what is possible to create hype and value while offering things engineering knows will never really quite work like they're advertised.
To your point, but on a larger scale, over hyping Palantir has the added benefit of providing a chilling effect on public resistance.
As a former government employee I had the same reaction to the Snowden leaks: sure the government might be collecting all of this (which I don't support), but I've never seen the government efficiently action on any data they have collected.
Incompetence might be the greatest safety we have against a true dystopia.
Because Snowden, agree with him or not, showed us that reality blew away our imagination.
It may feel normal now, but back then, serious people, professionals, told us that the claims just were not possible.
Until we learned that they were.
Until that moment, the general sentiment about the government and the internet is that they are too incompetent to do anything about it, companies like Microsoft/Apple/Google/Snapchat are actually secure so lax data/opsec is okay, etc.
Meanwhile, the whole time, communications and tech companies were working hand in hand with the government siphoning up any and all data they could to successfully implement their LifeLog[1] pipe dream.
[1] https://en.wikipedia.org/wiki/DARPA_LifeLog
> Until that moment, the general sentiment about the government and the internet is that they are too incompetent to do anything about it
In 2008 I worked with a retired NSA guy who had retired from the agency 5 years prior. He refused to have a cellphone. He refused to have a home ISP. Did not have cable tv, Just OTA. He would only use the internet as needed for the work we were doing and would not use it for anything else (news, etc). He eventually moved to the mountains to live off grid. He left the agency ten years before Snowden disclosed anything.
An example like that in my life and here I sit making comments on the internet.
3 replies →
That was not the sentiment, at least not in my experience. There was a far more pervasive and effective argument - if somebody believed that the government is spying on you in everything and everywhere then they're simply crazy, a weirdo, a conspiracy theorist. Think about something like the X-Files and the portrayal of the Lone Gunmen [1] hacking group. Three borderline nutso, socially incompetent, and weird unemployed guys living together and driving around in a scooby-doo van. That was more in line with the typical sentiment.
People don't want to be seen as crazy or on the fringes so it creates a far greater chilling effect than claims that e.g. the government is too incompetent to do something which could lead to casual debate and discussion. Same thing with the event that is the namesake of that group. The argument quickly shifted from viability to simply trying to negatively frame anybody who might even discuss such things.
[1] - https://en.wikipedia.org/wiki/The_Lone_Gunmen
dont worry lifelog was cancelled in 2004 according to that wiki. Phew!
1 reply →
> Snowden, agree with him or not, showed us that reality blew away our imagination.
pretty much everything Snowden released had been documented (with NSA / CIA approval) in the early 80s in James Bamford's book The Puzzle Palace.
the irony of snowden is that the audience ten years ago mostly had not read the book, so echo chambers of shock form about what was re-confirming decades old capabilities, being misused at the time however.
Considering the US military has historically had capabilities a decade ahead of what people publicly knew about, anyone who said it just wasn't possible probably wasn't a serious professional.
Which claims? HN around that time was taking anything and everything and declaring it conclusively proved everything else.
I honestly have no god damn clue what was actually revealed by the Snowden documents - people just say "they revealed things".
Why are you asking here, versus going to Google and reading the original article from The Guardian? Or the numerous Wikipedia links that are on this page?
5 replies →
You know how it's considered a kind of low-effort disrespect to answer someone's question by pasting back a response from an LLM? I think equivalently if you ask a question where the best response is what you'd get from an LLM, then you're the one showing a disrespectful lack of effort. It's kind of the 2026 version of LMGTFY.
If you still want a copy-paste response to your question, just let me know – I'm happy to help!
Incompetence could also be incredibly dangerous given enough destructive willpower.
https://www.thenation.com/article/world/nsa-palantir-israel-...
They're not trying to use the data to act efficiently (or in the public good for that matter), and they sure as fuck don't want you to see it. They're trying to make sure that they have dirt on anyone who becomes their enemy in the future.
I've often said we're recreating Brazil [1] instead of 1984. It's an excellent film if you haven't seen it btw, and in many ways rather more prophetic and insightful than 1984. But the ending emphasizes that incompetence just leads to a comedy of absurdity, but absurdity is no less dangerous.
As for PRISM, it's regularly used - but we engage in parallel construction since it's probably illegal and if anybody could prove legal standing to challenge it, it would be able to be legally dismantled. Basically information is collected using PRISM, and then we find some legal reason of obtaining a warrant or otherwise 'coincidentally' bumping into the targets, preventing its usage from being challenged, or even acknowledged, in court. There's a good writeup here. [2]
[1] - https://www.youtube.com/watch?v=aJCxVkllxZw
[2] - https://theintercept.com/2018/01/09/dark-side-fbi-dea-illega...
> I've never seen the government efficiently action on any data they have collected.
Someone else on HN said it would be nice if the NSA published statistics or something, data so aggregate you couldn't determine much from it, but still tells you "holy shit they prevented something crazy" levels of information, harder said than done without revealing too much.
The NSA tried to do this during the Snowden leaks!
There were stories like "look at how we stopped this thing using all this data we've been scooping up"... but often the details lead to somewhat underwhelming realities, to say the least.
It might be that this stuff is very useful, but only in very illegal ways.
Secrecy enables several things, including:
- abuse
- incompetence
- getting away with breaking rules and laws
Sometimes, those are desirable or necessary for national security/pragmatic reasons.
For instance, good luck running an effective covert operation without being abusive to someone or breaking rules and laws somewhere!
Usually (80/20 rule) it’s just used to be shitty and make a mess, or be incompetent while pretending to be hot shit.
In a real war, these things usually get sorted out quickly because the results matter (existentially).
Less so when no one can figure out who the actual enemy is, or what we’re even fighting (if anything).
In addition to terrorist stuff, they are probably passing of bunch of stuff to the military or defense industry to do things like fine tune their radar to cutting edge military secrets.
Would be nice if we had some form of statistics in a way that wouldnt endanger any of the intel that just tells the general public "we dont just sit here collecting PB of data daily"
1 reply →
One of the problems is the fundamentals of their tech works "just enough".
IE; just looking at their puff piece demo for https://www.youtube.com/watch?v=rxKghrZU5w8
- semantic data integration/triplestores/linking facts in a database.
- feature extraction from imagery / AI detection of objects as an alarm
- push to human operators
You or I might expect this to be held to a high standard - chaining facts together like this better be darned right before action is taken!
But what if the question their software solves isn't we look at a chain of evidence and act on it in a legal/just/ethical manner but we have decided to act and need a plausible pretext; akin to parallel construction?
When you assess it by that criteria, it works fantastically - you can just dump in loads and loads of data; get some wonky correlations out and go do whatever you like. Who cares if its wrong - double checking is hard work; someone else will "fix" it if you make a mistake; by lying, by giving you immunity from prosecution, by flying you out of state or going on the TV, or uh, well, that's a future you problem.
To take a non US example: https://en.wikipedia.org/wiki/Robodebt_scheme
Debt calculations were flat out wrong
The unstated goal/dogwhistle at the time was to punish the poor (cost more than it would ever recover)
It was partially stopped after public outcry with a few ministerial decisions.
It took years; people dying; a royal commission and a change of political party to put a complete stop to it.
No real consequences for the senior political figures who directly enacted this
Limited consequences for 12 of 16 public servants - no arrests, no official job losses, some minor demotions.
If the goal of the machine is to displace responsibility; the above example did its job.
You've never seen it because when it's efficient you won't see it.
If they throw out things like due process and reasonable doubt they can do a whole lot with the data they've collected.
That may sound hyperbolic but I hope it's obvious to most people by now that it's not.
They can do parallel construction or use "undercover" informants etc.
Fuzzy Dunlop (it's from The Wire). Their CI was a tennis ball (with an unauthorized camera inside).
> ... I've never seen the government efficiently action on any data they have collected.
It isn't usually a question of efficiency, it is a question of damage. Technically there is an argument that something like the holocaust was inefficiently executed, but still a good reason to actively prevent governments having ready-to-use data on hand about people's ethnic origin.
A lot of the same observations probably apply to the ICE situation too. One of the big problems with the mass-migration programs has always been that there is no reasonable way to undo that sort of thing because it is far too risky for the government to be primed to identify and deport large groups of people. For all the fire and thunder the Trump administration probably isn't going to accomplish very much, but at great cost.
>I've never seen the government efficiently action on any data they have collected.
As a former intelligence officer with combat time I promise you there are A LOT of actions happening based on that data.
doing Bad Things poorly is still doing Bad Things.
I see palntir as a techno whitewashing Mckinsey consultant. But the tech is there to make a much bigger problem than prior art, halucinations et al.
They are still dangerous even if theyre over promising because even placebos are dangerous when the displace real medical interventions.
Because palantirs selling proposition is: you can’t find the answers in your own data, but we can.
It sure would be convenient if they were always ineffective. Sadly there have been periods in history where governments have set themselves to brutality with incredible effectiveness.
No, incompetence is terrifying. No one wants to get caught in a machine driven by imbeciles who don't care about truth or honoring the Constitution.
Competence is also terrifying, but for different reasons.
Except you don't need to solve any remotely hard technical problems for the capabilities to be terrifying here.
I honestly tempt fate for fun to see how good police surveillance tech is the last few years.
I let one of my cars expire the registration a few months Everytime, because I'm lazy and because I want to see if I get flagged by a popup system Everytime a police officer passes near me. My commute car is out of registration 3 months right now. And old cop friend told me they basically never tow unless it's 6 months. I pay the $50 late fee once a year and keep doing my experiment for the last 6-7 years. Still no real signs they care.
My fun car has out of state plates for 10 years now. Ive been pulled over once for speeding, and told the officer I just bought it. I've never registered it since I bought it from a friend a decade ago. They let me go. It makes me wonder if one day they'll say "sir, we have plate scanners of this vehicle driving around this state for a year straight.. pay a fine." Not yet.
Cops use those systems to make easy arrests for things like active warrants, stolen vehicles and they feed into systems that keep track of where licensed vehicles are and when.
In a way that's worse, because the systems aren't looking up your car or to target your vehicle for fines, but to look up and target you for arrest.
Same systems can be used to identify, track and arrest undesirables.
[dead]
[flagged]
The algorithm was sorting punch cards and then putting the cards in different stacks on a table.
We can only hope that the surveillance state is still working with the same algorithm…
[flagged]
The nazi transformation didn't happen over the course of half an hour. Or one election cycle, even. The history is rhyming pretty hard right now.
2 replies →
Yeah if deportation is now Nazism, then the Allies after WW2 were Nazis too for the millions of mass displaced persons to match new borders.
lol. came here to say pretty much the same thing.
I've generally held this position, but assume a sufficient combination of models could do a lot more than was possible before.