I don't get the hype around palantir. It's a regular consulting shop like any other. Same as SAP, IBM, HP, etc. but a lot cheaper because their labor is a lot cheaper.
So what does it mean it no longer works for the NSA? Like they stopped taking government contracts? Of course not. They still work for big banks and big corporations and they still have government contracts. Whether the contract is with NSA or CIA or FBI I don't think makes a difference.
> Unless there's something else non-American like me cannot see :p
there is -- palantir is a privately held company. in his interview with charlie rose, alex karp mentions that being a private company is essential to palantir, though he doesn't specify why. at a minimum, it means they have no obligation to shareholders and they can take whatever contract they want without public scrutiny.
I wouldn't automatically jump on the conspiracy-theory bus: SOX and other regs for public companies are a great deal of work (while they provide some assurance to investors). Private companies get to keep more things secret, choose what/whom to reveal certain details and don't have to do as much reporting/audit, although it's still a good idea.
>> Serving shareholders’ “best interests” is not the same thing as either maximizing profits, or maximizing shareholder value. "Shareholder value," for one thing, is a vague objective: No single “shareholder value” can exist, because different shareholders have different values. Some are long-term investors planning to hold stock for years or decades; others are short-term speculators.
>> Also, most investors care not only about their portfolios, but also about their jobs, their tax burdens, the products they buy and the air they breathe. Which is to say, companies that maximize profits by firing employees, avoiding taxes, selling shoddy products or polluting the environment can harm their shareholders more than helping them.
>Going public does mean that your company is obliged to maximise shareholder profit, rather than achieve whatever goals the founders envisioned.
and
> And what exactly is the interest of the shareholders, rather than maximising share value / dividend payout?
Wait. By your own logic why would anyone invest in anything that wasn't public if you were just throwing money at some dude that wanted to do something that wasn't in your interest? Your logic makes no sense. It sounds like you're saying, it's sad that founders don't get to spend money anyway they want without carrying about the wishes of who gave them the money.
> By your own logic why would anyone invest in anything that wasn't public if you were just throwing money at some dude that wanted to do something that wasn't in your interest?
Because usually you get preferred stock, that means that all the 'crazy' non-profit-oriented things that the founders might be doing are still done with their money, rather than yours.
There's just a difference in both incentives and philosophy, isn't there? If the company is controlled by a private person, they're not obliged to make decisions that are 'best for the company'; and anyone choosing to go along for the ride is expressing trust into that persons ideas and goals. (And any contractual terms they set in their shareholder agreement; but the owners can refuse terms that misalign with their vision)
>It sounds like you're saying, it's sad that founders don't get to spend money anyway they want without carrying about the wishes of who gave them the money.
No; I'm saying that if as a founder you want to be able to spend money any way you want, you do not want to go public.
It's not overblown. We're averaging ~125 IPOs per year since the act was passed. That's not even equal to the worst IPO years in the 1980s during recessions. Meanwhile the US economy is at least twice as large as it was in 1985 on an inflation adjusted basis.
2016 saw 106 IPOs. That's below horrible recession years in the early 1980s. And that was with the stock market at very high levels.
The stock market has hit new all-time highs, the Nasdaq is on the moon, and very few companies want to go public. The entire vast shift to the private market occurred solely due to Sarbanes Oxley. The simultaneous extreme growth of the private capital market for ventures, the drastic reduction in IPOs vs the prior 30 years (while there has been an explosion of billion dollar start-ups), and the fact that the best companies are choosing to wait a lot longer to IPO vs the prior 30 years, makes it overwhelmingly obvious what has happened.
IPOs are businesses last-ditch, later-round financing after sovereign wealth, institutional investors, investment bankers, superangels and everyone else says "no."
They're becoming rarer, except for ventures which require an enormous amount of working capital, because angels and other investors typically have more cash than pre-Sequoia times to lend out to ventures which are smart about their cash & business models.
A cynic might suggest that it's to keep the army of consultants on board - the carrot in the distance of their stock options finally paying out. (How long is their exercise provision after you leave the company?)
It's not an everyday platform. It was created to compete with/replace the terrible systems FBI and others have spent billions on without results. It's used by large corporations, nonprofits and domestic and foreign governments to make sense of intelligence. It's like Google for intelligence... can be used to find human traffickers, LAPD infiltrating political groups or other govts assassinating journalists.
Palantir might not be "Google for Intelligence", but there is some unfavorable bias on HN against them, so the "Accenture for Intelligence" nickname might stem from wanting to be condescending.
Palantir and Accenture are consulting company. They make money by reselling meat by the hours, as we say in business. They don't profit from what is created at the client. They profit more by buying cheaper meat (graduates) and reselling with higher margin to clients.
Google is a software company. It makes money on the software service they provide. They get the returns for everything you do. They get more money by you making "better" software infrastructure and services.
Software companies benefit from code re-use and quality by reselling the same software to every client, then getting all the returns.
Consulting companies don't. The reason they are called in the first place is to make highly customized software.
Most consulting companies will pretend to re-use things across clients. In practise however, it cannot be achieved because it's going against their business model (yet they'll still pretend they do).
Consulting companies very often do not own the software they build at a clients, and so can't really reuse it. That's not what Palantir does.
I'm not sure what you think the business model is that would go against code reuse across sites. If you can afford to negotiate for ownership of what you build at a client site, why wouldn't you reuse things that are common?
Like, the reuse thing isn't binary. There's a whole spectrum between 'give everyone exactly the same thing', and 'give ever y single client something built for them from scratch'. And if you can build something from scratch at one client site, take half of that to another client, take 3/4 of that to yet another, and eventually end up only having to build 2-3% of the thing at every place, that's building a product while already getting paid for it.
>>> I'm not sure what you think the business model is that would go against code reuse across sites. If you can afford to negotiate for ownership of what you build at a client site, why wouldn't you reuse things that are common?
Because there isn't much that is common. You're called to craft customised software, with special customization for the client needs and tight integration with the client systems.
Anyway, none of this matters. At the end of the day. A consulting shop doesn't bill software, it bills the hours*men.
Whatever spectrum you think you can achieve, your actual business model always push back to its extreme side.
I've worked at consulting shops trying to bill software as software. I've never seen it goes very well.
Palantir is more consultingish-convert-to-subscription/product model where possible... they make money anyway they can and keep looking to build new business opportunities with consultingish svcs.
In general, strict consultancies are inherently unscalable and expensive, whereas SaaS has a unit cost of nearly $0. The benefits of consultingish, unscalable behaviors are known, in order to get to, and create new, scalable products.
I'm asking because I don't know, does Accenture typically create the software they sell? Off topic, does TCS/Infosys typically make the thing they sell?
Google's bet is that technology will replace humans. Palantir's bet is that technology works best when combined with human intuition. So yes, they do horde a lot of people but the idea is to bet on the intersection between man and machine rather than trying to replace one with the other!
at first glance, yes i think this is the case, especially people straight out of college
a similar trend happened with dell, hp, oracls and ibm- gone are the older, whiter males making over 100k base to sell to corporations as "trusted advisors"
in are the thousands of young kids fresh out of college dialing for dollars and making 45k
Depends on the audience but the senior management at a lot of places is often disproportionately older white men of a certain background (MBA from one of a few schools, former frat / current country club member, etc.). This is relevant to consulting and sales because a lot of clients like to think someone like them is in charge.
I've known a few salesmen who were quite open about their role as the senior guy on the contract was literally to take the client it for a round of golf and reassure them that the 25 year-olds weren't going to suggest too much change too quickly. Sharing the same cultural background, wearing a good suit, etc. gave them instant credibility on that front so they could spend lunch talking about which colleges their kids/grandkids were looking at, which pro golf/baseball/football team was going to do well, etc. and get the deal because the client now felt confident that they'd get what they asked for.
It does depend on the audience, and really isn't unique to white golf-playing males. I've seen plenty of that sort of client management in a few types of work, ranging from suits and golf to hoodies and pubs, black tie events to nightclubs.
Oh, definitely – I just mentioned that because there was an era where that particular one was just a gravy train for salespeople targeting the Fortune 500, government agencies, etc.
This is definitely a generational thing and I'm sure there's more diversification as the Boomer executives retire. There's must be some salesperson cleaning up with craft beers and SXSW or Tour de France trips now.
It has nothing to do with sales attitude and everything to do with empirical evidence of who large corporations (and their employees) decided to hire for these jobs.
They are allegedly better at exploiting young people.
They've got a better image, they get graduates for cheaper, oversell the future IPO and make them work longer hours.
It's no secret that Palantir is a sweat shop filled with "Forward Deployed Engineer", a bullshit title to make people feel important and impress graduates (it does work!).
The sad reality of consulting shop: You are sent from client to client anywhere anytime (the "forward deployed"). Your company takes a big cut on your salary, and they benefit by paying you less while charging more for you.
Trilogy Software in the late 90's had the exact same model with cultish hype aimed at graduating students at top universities who mostly didn't know any better and bought into the hype.
Oh Jesus. The shenanigans that went down at Trilogy. The stories and forwarded emails that went around Austin I still talk about.
The guy who replied all seeking clarification on what exactly a "serious" relationship was (regarding who could be your plus one in the company retreat). He got quite graphic.
The guy who got hired, showed up one day to do HR paperwork and got paid for 3 months before anyone figured out he never showed up again.
I was at a client site that used Trilogy (an insurance company using their commissions engine.) One day, i'm told that Trilogy HR [unintentionally] implemented an exponential bonus scheme for referral hires. $1000 for one referral, $2000 for two, $4000 for three, and so on. The Trilogy consultants at our client site were laughing at the matter and joking about pooling together all the referrals so they could get $1000 * 2^10 in referral bonuses. Needless to say the bonus scheme was revised.
Because they have a product base that lets them leverage a smaller amount of people into a better end result for a deployment. Less labour required = lower labour cost.
At least in the UK, FDE's are definitely not underpaid; I don't know how the pay compares in the SV.
The hours are a red herring. Sure, there's people who work very long hours, because they have nothing else to do. But in 2 years as an FDE at Palantir I averaged maybe 50h a week, and only remember one crunch week in all two years where I chose to work >80h (i.e. work on the weekend).
You might not realize but 50 hours average is a lot.
The 42k figure is very low. That's likely for all IT roles including tech support and help desk.
If you are good enough to enter Palantir (they have some standards despite the bad PR), you are good enough to enter a real tech company.
Note that if you are a consultant going left and right to a new client regularly, you should make a comparison against being an independent contractor too.
50h average is highish but not overwhelming. (Also I probably overestimated, I probably have as many ~50h weeks as 35-40h weeks; but I figured better be conservative) You get people on hackernews talking about how they work 60-80 hours all the time.
42k is not very low for an effecitvely entry level software engineer position. My first job in london after (a v. good) college was ~30k (5 years ago though, so sure inflation probably makes that ~32-33k now); finance was paying ~38-40k to grads back then.
Of course independent contracting is a thing; but you can't really be an independent contractor right after college (unless you built a great network, but then you're a ridiculous outlier), and even with job experience it's not a clear tradeoff.
But generally, it feels like you've moved the goalpost quite a bit here. Originally you claim the company exploits young graduates to build a sweatshop. Now we're talking about how they compare to the best tech companies around, and independent contracting.
That's 25% over a "standard" work week in the US. If that 42k is based on 40h/week, you'd need to earn 53k+ to "break even" on a 50+h/week (which Palantir appears to do).
FWIW, there are plenty of 40h-45h/week software jobs in the US that pay well into six figures.
There is a wide range of companies, usually body shops at the bottom and tech companies at the top. Of course there is a wide variance, a great consulting shop is better than a terrible software company.
Palantir London is decent in compensation from what I have heard (that doesn't speak for their other offices). They are however known for pushing 60h+ hours weeks very very bad until your whole life is your work, screwing people with shares that will never be paid, and sending people here and there around various locations.
Of course, there is also variance within a single company. It's possible that you've got a good position at Palantir.
Is there still hype around Palantir? It's 13 years old. I think the hype has been gone for years now, similar to Airbnb or Uber. They're just regular big companies now like an eBay or Priceline. Certainly there's some curiosity focused on Palantir due to their somewhat secretive nature and who they've worked with historically.
Except Palantir doesn't do work that's valuable to anyone except the government. Palantir doesn't have advanced tech. They had a profitable government rent to exploit. The government is not allowed to spy on citizens as freely as the private sector is, so the former administration leaned on Palantir to do the spying on behalf of it.
It makes absolutely no sense to /directly/ work for your client when facilitating intelligence laundry is you core business or simply a opportunity for exponential growth. How long must causality chains be in the shady data business for it to always be legal ?
Whether Palantir does not like Trump or, as other have said here, Palantir does not have any sort of special technology. In either case I'm happy that the Trump administration won't be given the tools to dominate others.
To me, Trump should not be allowed to use any more technology than he can scrap together himself: a rock, a stick, a pile of dirt. That's about it.
If you don't believe in science, you should not be allowed to benefit from it.
65 comments
[ 466 ms ] story [ 1748 ms ] threadSo what does it mean it no longer works for the NSA? Like they stopped taking government contracts? Of course not. They still work for big banks and big corporations and they still have government contracts. Whether the contract is with NSA or CIA or FBI I don't think makes a difference.
[0] https://en.m.wikipedia.org/wiki/Peter_Thiel
there is -- palantir is a privately held company. in his interview with charlie rose, alex karp mentions that being a private company is essential to palantir, though he doesn't specify why. at a minimum, it means they have no obligation to shareholders and they can take whatever contract they want without public scrutiny.
No. They have to act kb the interests of shareholders, whatever interest that is.
>> Serving shareholders’ “best interests” is not the same thing as either maximizing profits, or maximizing shareholder value. "Shareholder value," for one thing, is a vague objective: No single “shareholder value” can exist, because different shareholders have different values. Some are long-term investors planning to hold stock for years or decades; others are short-term speculators.
>> Also, most investors care not only about their portfolios, but also about their jobs, their tax burdens, the products they buy and the air they breathe. Which is to say, companies that maximize profits by firing employees, avoiding taxes, selling shoddy products or polluting the environment can harm their shareholders more than helping them.
and
> And what exactly is the interest of the shareholders, rather than maximising share value / dividend payout?
Wait. By your own logic why would anyone invest in anything that wasn't public if you were just throwing money at some dude that wanted to do something that wasn't in your interest? Your logic makes no sense. It sounds like you're saying, it's sad that founders don't get to spend money anyway they want without carrying about the wishes of who gave them the money.
Because usually you get preferred stock, that means that all the 'crazy' non-profit-oriented things that the founders might be doing are still done with their money, rather than yours.
There's just a difference in both incentives and philosophy, isn't there? If the company is controlled by a private person, they're not obliged to make decisions that are 'best for the company'; and anyone choosing to go along for the ride is expressing trust into that persons ideas and goals. (And any contractual terms they set in their shareholder agreement; but the owners can refuse terms that misalign with their vision)
>It sounds like you're saying, it's sad that founders don't get to spend money anyway they want without carrying about the wishes of who gave them the money.
No; I'm saying that if as a founder you want to be able to spend money any way you want, you do not want to go public.
2016 saw 106 IPOs. That's below horrible recession years in the early 1980s. And that was with the stock market at very high levels.
The stock market has hit new all-time highs, the Nasdaq is on the moon, and very few companies want to go public. The entire vast shift to the private market occurred solely due to Sarbanes Oxley. The simultaneous extreme growth of the private capital market for ventures, the drastic reduction in IPOs vs the prior 30 years (while there has been an explosion of billion dollar start-ups), and the fact that the best companies are choosing to wait a lot longer to IPO vs the prior 30 years, makes it overwhelmingly obvious what has happened.
They're becoming rarer, except for ventures which require an enormous amount of working capital, because angels and other investors typically have more cash than pre-Sequoia times to lend out to ventures which are smart about their cash & business models.
https://www.buzzfeed.com/nitashatiku/protest-palantir-muslim...
https://www.bloomberg.com/news/articles/2011-11-22/palantir-...
I.e. they don't seem to have scary innovative technology as much as a horde of people doing integration work (and some good reusable tools)
See discussions from 1 year ago
https://news.ycombinator.com/item?id=11646587
And 2 years ago
https://news.ycombinator.com/item?id=8872054
Google for Intelligence is what they have been selling, but it's unclear they have actually delivered.
Palantir and Accenture are consulting company. They make money by reselling meat by the hours, as we say in business. They don't profit from what is created at the client. They profit more by buying cheaper meat (graduates) and reselling with higher margin to clients.
Google is a software company. It makes money on the software service they provide. They get the returns for everything you do. They get more money by you making "better" software infrastructure and services.
Palantir does indirectly profit from what's created at the client, by reselling the code created at one client to other clients later.
Consulting companies don't. The reason they are called in the first place is to make highly customized software.
Most consulting companies will pretend to re-use things across clients. In practise however, it cannot be achieved because it's going against their business model (yet they'll still pretend they do).
I'm not sure what you think the business model is that would go against code reuse across sites. If you can afford to negotiate for ownership of what you build at a client site, why wouldn't you reuse things that are common?
Like, the reuse thing isn't binary. There's a whole spectrum between 'give everyone exactly the same thing', and 'give ever y single client something built for them from scratch'. And if you can build something from scratch at one client site, take half of that to another client, take 3/4 of that to yet another, and eventually end up only having to build 2-3% of the thing at every place, that's building a product while already getting paid for it.
Because there isn't much that is common. You're called to craft customised software, with special customization for the client needs and tight integration with the client systems.
Anyway, none of this matters. At the end of the day. A consulting shop doesn't bill software, it bills the hours*men.
Whatever spectrum you think you can achieve, your actual business model always push back to its extreme side.
I've worked at consulting shops trying to bill software as software. I've never seen it goes very well.
In general, strict consultancies are inherently unscalable and expensive, whereas SaaS has a unit cost of nearly $0. The benefits of consultingish, unscalable behaviors are known, in order to get to, and create new, scalable products.
http://paulgraham.com/ds.html
Why is that? (genuinely curious here.) Do they employ younger people with hopes an IPO?
a similar trend happened with dell, hp, oracls and ibm- gone are the older, whiter males making over 100k base to sell to corporations as "trusted advisors"
in are the thousands of young kids fresh out of college dialing for dollars and making 45k
What does that have to do with sales aptitude?
I've known a few salesmen who were quite open about their role as the senior guy on the contract was literally to take the client it for a round of golf and reassure them that the 25 year-olds weren't going to suggest too much change too quickly. Sharing the same cultural background, wearing a good suit, etc. gave them instant credibility on that front so they could spend lunch talking about which colleges their kids/grandkids were looking at, which pro golf/baseball/football team was going to do well, etc. and get the deal because the client now felt confident that they'd get what they asked for.
This is definitely a generational thing and I'm sure there's more diversification as the Boomer executives retire. There's must be some salesperson cleaning up with craft beers and SXSW or Tour de France trips now.
They've got a better image, they get graduates for cheaper, oversell the future IPO and make them work longer hours.
It's no secret that Palantir is a sweat shop filled with "Forward Deployed Engineer", a bullshit title to make people feel important and impress graduates (it does work!).
The sad reality of consulting shop: You are sent from client to client anywhere anytime (the "forward deployed"). Your company takes a big cut on your salary, and they benefit by paying you less while charging more for you.
The guy who replied all seeking clarification on what exactly a "serious" relationship was (regarding who could be your plus one in the company retreat). He got quite graphic.
The guy who got hired, showed up one day to do HR paperwork and got paid for 3 months before anyone figured out he never showed up again.
At least in the UK, FDE's are definitely not underpaid; I don't know how the pay compares in the SV.
Disclamer: I work for Palantir
Tip: When evaluating a work, adjust the compensation for hours worked and don't be tricked by the shares that will never IPO.
The hours are a red herring. Sure, there's people who work very long hours, because they have nothing else to do. But in 2 years as an FDE at Palantir I averaged maybe 50h a week, and only remember one crunch week in all two years where I chose to work >80h (i.e. work on the weekend).
The 42k figure is very low. That's likely for all IT roles including tech support and help desk.
If you are good enough to enter Palantir (they have some standards despite the bad PR), you are good enough to enter a real tech company.
Note that if you are a consultant going left and right to a new client regularly, you should make a comparison against being an independent contractor too.
42k is not very low for an effecitvely entry level software engineer position. My first job in london after (a v. good) college was ~30k (5 years ago though, so sure inflation probably makes that ~32-33k now); finance was paying ~38-40k to grads back then.
Of course independent contracting is a thing; but you can't really be an independent contractor right after college (unless you built a great network, but then you're a ridiculous outlier), and even with job experience it's not a clear tradeoff.
But generally, it feels like you've moved the goalpost quite a bit here. Originally you claim the company exploits young graduates to build a sweatshop. Now we're talking about how they compare to the best tech companies around, and independent contracting.
That's 25% over a "standard" work week in the US. If that 42k is based on 40h/week, you'd need to earn 53k+ to "break even" on a 50+h/week (which Palantir appears to do).
FWIW, there are plenty of 40h-45h/week software jobs in the US that pay well into six figures.
Palantir London is decent in compensation from what I have heard (that doesn't speak for their other offices). They are however known for pushing 60h+ hours weeks very very bad until your whole life is your work, screwing people with shares that will never be paid, and sending people here and there around various locations.
Of course, there is also variance within a single company. It's possible that you've got a good position at Palantir.
Is there still hype around Palantir? It's 13 years old. I think the hype has been gone for years now, similar to Airbnb or Uber. They're just regular big companies now like an eBay or Priceline. Certainly there's some curiosity focused on Palantir due to their somewhat secretive nature and who they've worked with historically.
Wait.. what? Aren't they still on the top of the list for Stanford students?
To me, Trump should not be allowed to use any more technology than he can scrap together himself: a rock, a stick, a pile of dirt. That's about it.
If you don't believe in science, you should not be allowed to benefit from it.
I don't want these tools in the hands of any president when they don't respect the people. Whether it's Obama spying on us or Trump. Both are bad.