As a person that had the opportunity to work with them and their products, seeing a post like this makes me angry and sad. I personally find the company extremely incompetent, that hires young postgrade students without any prior experience to solve big problems for the biggest companies around the world. How it ends up is technical depts and gigantic costs. Not to mention few shady things that seems normal for them - full access to their customers business data and source code. Epidemics should not be a time for business opportunities. It's not like they have any closed patented solution that will suddenly solve any problems... It's definitely not a equivalent of openai in data engineering world. They are just experts in marketing bs.
If a platform involves tracking huge amounts of personal data and then making decisions off of that data, incompetence makes it more dangerous, not less.
>The LAPD uses Palantir’s Gotham product for Operation Laser, a program to identify and deter people likely to commit crimes. Information from rap sheets, parole reports, police interviews, and other sources is fed into the system to generate a list of people the department defines as chronic offenders, says Craig Uchida, whose consulting firm, Justice & Security Strategies Inc., designed the Laser system. The list is distributed to patrolmen, with orders to monitor and stop the pre-crime suspects as often as possible, using excuses such as jaywalking or fix-it tickets. At each contact, officers fill out a field interview card with names, addresses, vehicles, physical descriptions, any neighborhood intelligence the person offers, and the officer’s own observations on the subject.
>The platform is supplemented with what sociologist Sarah Brayne calls the secondary surveillance network: the web of who is related to, friends with, or sleeping with whom. One woman in the system, for example, who wasn’t suspected of committing any crime, was identified as having multiple boyfriends within the same network of associates, says Brayne, who spent two and a half years embedded with the LAPD while researching her dissertation on big-data policing at Princeton University and who’s now an associate professor at the University of Texas at Austin. “Anybody who logs into the system can see all these intimate ties,” she says. To widen the scope of possible connections, she adds, the LAPD has also explored purchasing private data, including social media, foreclosure, and toll road information, camera feeds from hospitals, parking lots, and universities, and delivery information from Papa John’s International Inc. and Pizza Hut LLC.
The LAPD have been careful not to cite information from this software in court documents so that its constitutionality cannot be challenged.
Yeah this is some dystopian stuff. Guarantee that people on here will (rightly) criticize non-Western countries for doing stuff like this, but will ardently defend Palantir doing similar stuff because they’re “innovative” or a “tech company”.
> big bad all-present panopticon, or are they incompetent, inexperienced muppets?
Sadly both are possible at the same time in a large well-funded company. If they succeed in getting these contracts from scared governments, they'll be better positioned in the future for more such work. Privacy and freedom "mishaps" will be chalked up to "oh, it was an emergency, we had to do what we had to do."
No inside knowledge, but by all accounts they seem to be both, and there’s no conflict.
An incompetent customer facing layer can enable a competent layer to act with full data and less visibility. The incompetent layer also helps extract finances.
Yes. They are Accenture or CGI, but just marketed in a different way.
They have no 'special sauce', it's in many ways kind of a scam.
When you grow up, you assume the authorities around you have integrity and competence, in many ways they really do, but on others, they really don't. It's weird.
It's really hard to even grasp the level of bureaucratic incompetence, which is a form of fraud, because it doesn't seem possible, but it's real.
This stuff is 'Developed World Corruption', it's a more sophisticated version of the same scams run in the developing world that we more obviously identify as fraud.
It's hard to explain how billions magically disappear on ridiculous, overly complicated, poorly designed systems that don't work and FYI had ridiculous requirements to begin with - but it happens all the time.
Semi-literate people building a power plant in a poor country don't know either, or rather, it may not be apparent when working on the project that 10% was taken off the top.
But I have to feel a lot of devs making a $1B simple app have a clue that something is off.
I think the power plant builders just by default assume they are being taken advantage of.
The devs are busy gathering requirements, going to meetings, being shuffled between projects etc etc. Consultants brought in. Stuff rescheduled. Maybe, maybe not.
Yes, this is probably it. And it's scary how intelligent professionals can exist in a system of total dysfunction and in a way not recognize it. This is worth studying - how the perception reasonable authoritativeness (which is not actually authoritative) perpetuates itself in social systems, i.e. the blind leading the blind, unaware they are blind.
I worked in a Fortune 50 and recommended we 'find some smart young devs' to work on this project, they thought I was crazy, it hurt my credibility. A year later, one of our interns quit and did the same thing, almost alone, with much greater success. I shamefully admit I was looking for my 'I told you so' comeuppance, shockingly, it never happened. The system was too inept and lacking in self-awareness to see what had just happened, internalize and adapt. Nope. Just a mixture of unaware, don't care, ignored, avoid. Someone at lunch somewhere might have said 'Wow, look at what that intern did'. Of course, they almost went out of business but it's not the same for other conglomerates.
> I personally find the company extremely incompetent, that hires young postgrade students without any prior experience to solve big problems for the biggest companies around the world. How it ends up is technical depts and gigantic costs.
Isn't that literally any larger technical consulting company? I've had the same experience with Bechtle, CGI, BearingPoint and Accenture.
Worked for them as a Product SWE for 2 years. For the record, I would not recommend it if you have good alternatives as a SWE.
That being said, I want to give a more nuanced comment because, in my opinion, crisis require some pragmatism which I find to be surprisingly lacking from the usually reasonable HN crowd.
I understand why Palantir create this epidermic reaction and in most cases I would agree that they are justified. There is an issue of sovereignty, countries should aim be self-sufficient in critical areas like Defense, Healthcare, Counter-terrorism. It is a shame that it is not the case. Many Palantir employees are motivated to help their country but concluded that the most straight-forward path, if they did not want to spend many years in a bureaucracy hell, was to join a startup/software style company, that would respect their value-added as young/bright engineers and make some of them (the majority in commercial works on boring stuff) able to work on those topics.
Yes, compared to many cool tech companies, there is nothing great about their tech. It's a glorified Spark wrapper with a pretty UI and terrible UX. Nothing with significant tech depth is built-in house. However, THAT is compared to cool/good tech companies. Compared to a lot of what big corp and bureaucratic governments can readily provide TODAY, it's a significant improvement and it is ready, TODAY. It is also more secure compared to what an in-house gov team could build, as in, they actually encrypt things and there is no obvious security flaws. Now is the time to use the tool that can reduce the death tools without compromising security, tomorrow is the time to ask our politics to invest in their own technology sector and promote young/bright engineers to build state-of-the-art data platform that does not involve Palantir.
Why Palantir is doing such action is not to gather data (they do not), or to earn money (it's likely pro bono) but to establish a relationship with governments, a trojan horse if you will, so that they can sell their services to more departments. The pragmatic action would be to accept the help now in crisis time and leverage their tools and somewhat competent Forward Deployed Engineer workforce, but decline their expansion when we are past the crisis.
Also, their big principle is human insight over algorithms and machine learning. This was derived from experience in Paypal, where human in the loop approaches worked much better for fraud detection than algorithmic approaches.
In some positive light, 3-4 years ago Google or Facebook would've jumped on and published a contact tracing app without thinking twice. I wonder what's going on inside in terms of discussions. Some positive side effect of the tech slash to step back and think twice of side effects or maybe they've learn to avoid possible bad PR.
> I meant help in any way that would benefit everyone.
And I'm asking how that would help. So far, every single proponent of "surveillance because covid" I asked this avoided the question.
It's also pretty incredible how you split hairs about "everybody" in the first sentence, and then talk about something that will help "everyone" in the second, without even saying how.
How can you say that? The Netherlands still has people going to cafes. The UK is talking about banning outside exercise. Even if you can point to a figure saying people are moving less and adhering to their government's guidelines I don't think it follows that you can say the majority of people are acting like they are infected.
As for the second part, by definition, if those companies actions benefit everyone.. it would help. I didn't say and I don't know how they would do that.
Since this conversation started, it looks like Google has used their great engineers, great infrastructure and lots of money to publish movement reports. This might feed into Government decisions and hopefully will be a benefit for "everyone". Check it out here: https://www.google.com/covid19/mobility/
> How can you say that? The Netherlands still has people going to cafes.
Because vast majority isn't, certainly not in Germany. I can say that because I have been experiencing it for weeks now.
> Even if you can point to a figure saying people are moving less and adhering to their government's guidelines I don't think it follows that you can say the majority of people are acting like they are infected.
Those guidelines to after entail acting like people oneself might be infected. That's why people are to keep distance from anybody, never ever touch their face while out and about, and wash their hands thoroughly when they come home (and I might add, also before they leave home).
> As for the second part, by definition, if those companies actions benefit everyone.. it would help.
I ask how surveillance help, and you say "if it helps, it helps by definition". That is hardly an answer.
Let's say 10% simply don't take it seriously. If they don't keep distance from anyone out of the 90%, those people will know. In Bavaria, someone already went to jail for 8 days because of exactly that, for not paying heed despite repeated warnings.
There is no surveillance necessary for that because -- by definition -- the only way for those 10% to "secretly mingle" would be with other people of those 10%. Movement tracking also doesn't tell us anything about if people cough right at other people or into their elbow, or if they wash their hands when coming home. Mere distance itself isn't saying all that much.
As for the Google thing: super general, anonymized "where do people move around how much" is something else entirely than the location data of all indiviuals.
> What data is included in the calculationdepends on user settings, connectivity, andwhether it meets our privacy threshold. If theprivacy threshold isn’t met (when somewhereisn’t busy enough to ensure anonymity) wedon’t show a change for the day.
That isn't the kind of surveillance we are talking about, like the kind of German justice ministry previously rejected, for example.
Information that harms nobody, violates no rights, is 100% anonymous, doesn't even have to be useful for me to have no problem with the collection of.
Palantir on the other hand "declined to comment", so you give me something completely different from Google. If what Palantir have in mind is so benign and useful, why don't they proudly explain it? It's not like CoV19 could overhear that and adjust, like criminals or terrorists might.
And from a PR perspective, declining to comment is the ONE predictably bad move here. If they want to help, if they have something helpful to bring to the table, why aren't they open about it? With something like this, they should have had their story straight before even anyone asked them, before they even finalized their offers to those governments. This is nowhere good enough. That isn't "the second part", that is the main issue.
To me it's just as creepy to let any big tech company into healthcare such as Google or Facebook. To play the devils advocate, maybe people here have this reaction to Palantir because they disagree with Peter Theils political viewpoints whereas others would have the same issues with Google/FB.
Current administration supports Peter Thiel's Palantir and getting them contracts: FDA, CDC, NIH. Apparently Mr. Thiel was one of the few Silicon Valley supporters of Mr. Trump.
Also Trump's (and other leaders around the world) initial downplaying of the pandemic risks paved the way for extreme measures like the employment of pervasive surveillance. Looks like someone isn't letting this crisis go to waste.
To the victors belong the spoils. Previous administrations benefited their friends and future administrations will likely benefit theirs. Don't like it? Vote for good government.
This is true but in a uninteresting, shallow way. Yes, Team Blue is also bought. It's the same people doing the buying regardless of which side holds the majority. The wheels need turning so the wheels need oiling.
However, general elections and the say of the people is still an important part of the political machinery. And this shows:
Team Red is going all in for a mythical Core American Christian Constituency Who Values Hard Work and Won't Let Anyone Else Steal Their Hard Earned Dollars.
(This rings as much true whether you are a rust belter seeing the coal industry shrinking or you are Thiel, they share the same sentiment.) The Thiels of the world with enough money to buy Congressmembers will have a much easier time buying Team Red, because they are already aligned ideologically.
As long as general voting is still an important part of politics, Team Blue can't accept every Thielesque bill, because their constituency is Everyone Else, a much more disparate group than what Team Red is aiming for.
Outside of the minority in the Twittersphere bubble most engineers are apolitical. Palantir offers the opportunity to work on interesting software solving interesting problems and that gets them very far.
Defense contractors building bombs and the NSA don't really struggle to hire engineers, either. Assuming everyone has your same belief system is a critical mistake.
Are you sure about that last point? I remember reading somewhere that U.S. government agencies were having a hard time hiring tech talent, as they did not pay as well as Silicon Valley nor had the same growth opportunities. Moreover, they had issues with the patriotic argument, as many millenials don't share the same ardent patriotism that previous generations had. It's been a while since I read that article, so things might have changed and I might be wrong.
You can uphold your morals if your morals are different. Not like morals are encoded in law or need to adhere to progressive leftist political views.
For instance: algorithm copyrights and patents are BS. It is immoral to patent a mathematical recipe. Wikileaks turned out to be a dangerous threat. It is immoral to launder Russian intelligence and hide your non-redacted Tor exit node stolen data dumps as journalism. It is nonsense to require a company to be unable to hire for culture fit, and immoral to claim your genetic make-up made all the difference for you not getting a job. It is silly not to make allowed use of Facebook data to help your campaigning client, and it is immoral to attack this, just because it resulted in a President you do not like, while ignoring the fact that social media "engagement" was what won your favorite pick a presidency before that. It is immoral to oppose creating technology for the police to combat human trafficking, illegal immigration, gang violence, and drug problems, under the banner of woke anti-racism.
> I wonder who responds positively to Palantir recruiters when I dislike this company so much?
> Is the money so good that it is easy to suspend my morals?
Maybe those who were violently 'other'-ed or marginalized in our dysfunctional technological society? Those who developed a subsequent unconscious (euro-centric) savior-complex? The harmful ubiquity of the lone genius myth doesn't help.
Alexander Karp talks about 'terrorists', and seems to be very patriotically committed to the American empire:
I actually really get the general reaction here to Palantir. What I don't get is why the FOSS community hasn't just done the right thing and produce a competitor yet. It's not like link analysis [1] with shared backends isn't an interesting problem and big business producing billions in revenue for various companies every year.
Gephi (and other graph tools) aren't quite the same thing.
The front-end user interface, and the workflows are different. It's a little closer to how maybe yED works with dragging and dropping of entities, connections, groups, but also a robust meta-data system attached to all of those things + some tools that make bringing in new information from free-text (or wherever) simple. Then all that stuff should be stored in some kind of centralized semantic database that any user using that system can search and find stuff somebody else already put in there. Maybe a document search as well so you can see what raw documents have been ETL'd.
Then take various attributes off of some of the entities and relationships (like dates, or geos, or IPs) and builds various ways of looking at, filtering, and search that stuff quickly.
It's a great opportunity for Palantir to do some corona-washing and get their hooks into EU Govt.
On the plus side, Palantir has a reputation for delivering results. Govt needs to do location and contact tracking on a mass scale, integrated with the test regime, and it needs rapid analysis of this data.
Negatives include what they will do for governments post the immediate Corona crisis. I'm expecting forcible deportation on a mass scale to deal with the long tail of hard to manage infected populations, and general xenophobia.
I sure hope governments will be wise enough to not cooperate with Palantir, which is the embodiment of the worst excesses of surveillance capitalism. Managing the spread of the virus is a critical issue and if some surveillance of the population is required, it necessitates a lot of trust and transparency on part of the government. Working with Palantir would undermine that from the start.
In principle, I'd have been open voluntarily to installing such a privacy-preserving bluetooth contact tracing app (provided it's open source and auditable). But knowing that Palantir may or may not be involved in analyzing that data, I'll probably pass.
FYI, they are already working for a while with the police in some German states (Bundeslaender) [1]. So this is not as big news, as it might seem to some.
The goal is not to fight the epidemic, it’s probably to track: (1) current social networks talk about the no-go zones that don’t respect confinement, and (2) see if any Gilets Jaunes 2.0 or any revolutionary groups is forming. The risk is real for the government, which was already hugely unpopular and as handled the crisis catastrophically (see below).
Government mistake includes:
- doing nothing for two months
- lying about the dangerousness of the virus
- saying that masks are useless for citizens yet requesting all available
- sending masks to China at the beginning of epidemic
- not restocking masks despite the existence of a pandemic emergency stock
- neglecting hospitals for decades (staff was striking a few months ago)
- battling the potential treatment of Pr. Raoult because of a personal feud between him and the ministry of health’s husband
- putting chloroquine under toxic medicine list just as the epidemic starts
- changing the layout of the paper need to present to police everyday
- enforcing confinement on a part of the population but not another one
- ... I probably forget a lot more
Sources: French media, friends, family and social networks.
People are genuinely angry at the govt response and some are already talking in insurrectional terms.
Phew, I was getting worried there for a second for the US. If any Gilets Jaunes would form in the US, they wouldn't turn on the administration. And they'd be more of "Gilets Brune" anyway.
They'd be on the side of the administration ready to torch whatever group must bear the blame. Some Palantir-like intelligence can be very useful when you want to direct your s̵h̵o̵c̵k̵ ̵t̵r̵o̵o̵p̵s̵ supporters to peacefully protest.
Yes, that was my thought. Contact tracing is likely the top reason Singapore. S. Korea (and to a lesser extent Japan) were able to selectively & aggressively contain infections, along with active policing and enforcement. Palantir's core focus is on databases for tracking "assets" with a geographic component, so it would seem like a natural application.
If this reduces the need and application for nation-wide bans like contact tracing did in Singapore than whatever the cost it will be a rounding error.
> than whatever the cost it will be a rounding error.
"Whatever the cost" disregards the non-monentary costs of giving a player like Palantir access to this data. If you need to implement digital help for contact tracing, at least pretend to do it in a privacy-preserving manner (plenty of attempts out there). There's no need to give surveillance data to the security apparatus of another country just because they have a few spreadsheets already.
It’s also pretty decent at constructing social networks and interaction graphs form different types of data which is what they pitch for fraud detection and tracing.
Basically one of their only models that actually works seems to be suited for contact tracing nearly perfectly.
I understand the reservations of people here regarding the company itself tho, but I’m not against it as a blanket statement.
E.G. on prem installation on German government servers with German staff doesn’t seem like an overreach, especially considering that the German government is already a client.
> Contact tracing is likely the top reason Singapore. S. Korea (and to a lesser extent Japan) were able to selectively & aggressively contain infections, along with active policing and enforcement.
We can provide you a tracking solution, and after the pandemic is over, you can continue to use the product to track and spy on all your citizens. What a perfect way to incorporate our superior technology into your Gov't, all under the auspices of a public emergency !
It doesn’t affect me (because I don’t trust my government enough with this data to use that voluntary service, use of any such things tends to expand more and more and deleted data hasn’t been deleted), but when I first read that Palantir is one of the companies being talked to I had to laugh.
I strongly oppose the existence of companies like this and the sorts of mass surveillance activities that the NSA, GCHQ etc engage in. Having these programs makes core processes in our society like democracy and trial by jury uncertain.
But given we have let them in, I am also amazed how little we seem to demand of them. The NSA should easily be able to track everyone an infected person has been near. That includes people who have been to places shortly after the infected person was there.
So why aren't they doing that? The marginal cost to our society is nothing and this epidemic is already worse than all the terrorism ever in most western nations. These organisations are not just dangerous, they're dangerous and so far totally useless.
