8 ms·

The US government can't legally pay the salaries required to find AI experts right now (just look at the salaries on the open positions). So all this work will

The US government can't legally pay the salaries required to find AI experts right now (just look at the salaries on the open positions). So all this work will end up being done by contractors who _can_ pay those salaries while the contractors skim off the top. The US really needs a special schedule for software developers if they want technical expertise in house (and they really should have technical expertise in house)

mattnewton 3y ago
Can’t they just use the contractor loophole?
- hiddencost 3y ago
  Career bureaucrats have historically been more loyal to the government, IME, and have their incentives better aligned.
- acdha 3y ago
  They can but that has several widespread problems: the biggest is that you need skilled people in the government who can review bids and oversee work – otherwise you get a $10M contact form because someone at BigCo1 wanted to play buzzword bingo and keep their buddies employed, and the “overseer” from BigCo2 wasn’t going to saying the team, often former coworkers, wasn’t performing well because that’d invite honest grading on other contracts where the relations are reversed. Beyond that, you get various efficiency issues: contractor turnover tends to be higher, which is terrible when institutional memory matters or when you’re relying on their best people to keep a larger group productive, and there are often various contractual or policy issues making work less efficient. One challenge is that managing contracts is a lot of work, so there’s a temptation to go with the big companies who offer many services but that often means that they’re only good at one area and you get mediocre results in the other ones which were lumped into the same contract - often banking on the difficulty of taking corrective action meaning that people will accept low performance in non-core areas to avoid jeopardizing a core project by the same company. None of that’s intractable but it adds up and fixing it is hard.
  - wslack 3y ago
    > the biggest is that you need skilled people in the government who can review bids and oversee work It's also very isolating to be the "one person" who knows the subject area. Having teams of folks is much more sustainable.
godelski 3y ago
Some of the jobs go up to $180k/yr. That's not too bad considering it doesn't look like it requires DC location. But others like the grad assistant for $22/hr is laughable. Still, I agree there's a problem with salaries not being competitive. I've interned at DOE labs before and lots of people don't stay for long.
- ryanklee 3y ago
  $180k/yr isn't bad in the context of Sr SWE in medium-cost-of-living areas. But in the context of ML/AI, it's pretty rotten no matter where. Hard to get the right people if that's the top.
- j45 3y ago
  Some say wage suppression is one way the much better paid bureaucratic class can keep their well educated potential replacements where they are, far away.
- tomohelix 3y ago
  I mean the grad assistant is basically a post-doc position isn't it? Considering the stability and perks from a government job, I think that is not too far from market rate. But go up and those 180k jobs would be easily 300-400K in the industry. I doubt anyone who can do that work would be interested in a measly 180K. Unless it comes with a visa...
  - dheera 3y ago
    Almost all those 300K+ industry jobs are willing to sponsor visas. Most of these jobs on the other hand seem to be US citizens only. Some require a security clearance which a lot of AI experts would probably rather NOT have.
  - dvngnt_ 3y ago
    "a measly 180k" is a wild statement on an absolute scale, but i do see what you're saying relatively.
  - kjkjadksj 3y ago
aduffy 3y ago
Recoding America[1] is an excellent book about this problem, but it extends a lot further than just contractors charging inflated rates. The government itself has a culture that is so averse to trying new things and compliance with existing policy that it's actually dangerous to the careers of civil servants to do anything innovative, because that is one of the only things they can actually be fired for. The solution she proposes effectively would be an expansion of the US Digital Service (which the author spearheaded) into something that approximates the British civil service, which is actually highly functional and effective. If you actually care about this problem and have a spare Audible credit it's worth a listen. [1] https://us.macmillan.com/books/9781250266774/recodingamerica https://us.macmillan.com/books/9781250266774/recodingamerica
- j45 3y ago
  Bureaucrats are so risk averse to such an extent that the culture can have an allergy to trying to do things different or better too quickly (including innovation).
  - fkyoureadthedoc 3y ago
    Did you just strip the context and rephrase exactly what the parent comment says?
    - j45 3y ago
      Not at all. Sincerely asking, does your question intend to pre-suppose that everyone must have and express a unique perspective and no one can independently have anything in common? Or some other faux pas such a comment would run a-foul of? When I saw original post, I had this comment. There was kinship with the parent post. Sometimes echoing your support with another comment is a valid comment too. If you may want more novelty and context and details: My professional experience includes working across multiple industries of bureaucrats, and that was my visceral reaction to the piece. Change usually means studying change, analysis, design and ultimately decision fatigue by committee. There's many hard working and great bureaucrats. My statement tends to apply to the leadership and management layers. In terms of keepign things the same or very slow and tiny changes incrementaly, there's a lot of neat reading around institutionalization as it relates to not only government but bureaucracy here. An exciting part is how the pandemic challenged and made it OK for bureaucracies to try things and to roll out programs that weren't fully thought out or known and update them regularly. In environments many people could not handle navigating. Most of all I'm flattered by the downvotes.. happy to take those for the parent too. Perhaps I phrased it in a way that might have triggered some bureaucrats since they exist in enterprise, government, education, healthcare and more. Or maybe it bothered the dopamine seekers in only novel and unique comments have worth.
mmedellin 3y ago
This is most acutely felt in acquisitions where those responsible for deciding what to build and buy have very limited understanding of software development. The failed rollout of healthcare.gov is the billboard for this problem. There's an effort to train civilian acquisition professionals on how to buy and build better software, but they are pretty quickly hired by contractors or the commercial sector since their new skillset fetches a larger salary than what the government can pay (even with promotion). I agree there needs to be a special schedule. I do think an overhaul of the civilian service is worth exploring.
- mistrial9 3y ago
  second hand knowledge in California is - the State is large and wealthy enough to have some echoes of the actual Federal govt.. and repeated words from direct software project participants was .. that mid-level bureaucrats from the "controlling" contract originator, would spuriously and without warning change the requirements.. sometimes so much so that it would invalidate large'ish portions of completed work.. and this happened all the way into the final month of a supposedly serious deadline, with failure to perform clauses and all that, at stake. A first-hand voice from California State software projects said - that the required prime contractor had 10+ person teams and $1m+ budget, and failed to build useful results. The final result from the project(s) more than once, was "oh oops, doesn't fit with -other framework- cannot be deployed". In other words, literally making things that are abandoned on Day 1 of deployment, for large money, with plenty of apparent pressure on "workers" to perform. etc..
- wslack 3y ago
  > I do think an overhaul of the civilian service is worth exploring. There are decades of reports and proposals about this. Unfortunately, we keep learning the same lessons over and over.
altdataseller 3y ago
A lot of those jobs aren’t really AI/ML related too just reading the descriptions. They seem like generic software/IT positions
- vwcx 3y ago
  This is how government hiring works. The job buckets are incredibly unspecific. Related: Open jobs at the CIA with the title "analyst".
dmead 3y ago
The US digital service tries to do this, but it just ends up maxing out the government payscale very quickly.
jlpom 3y ago
Same story in France and I suspect other countries
b8 3y ago
Yep, the US government gives contracts for a lot of stuff like cybersecurity etc. because the GS pay scale can't compete with the private industry standard.
curiousllama 3y ago
> The US really needs a special schedule for software developers There's actually precedent for this - SEC lawyers are on the SK pay scale, not GS. It's exactly the same problem: the alternatives in finance were too lucrative to retain competent enough attorneys to regulate wall street. Congress should create a separate tech pay scale, given the alternatives are too lucrative to retain competent enough engineers.
akira2501 3y ago
> and they really should have technical expertise in house Why? Moreover.. how would any of this expertise actually get applied top down in a federal system? How would any department making purchases or technology decisions actually benefit from such an apparatus? Wouldn't it be more sensible for them to manage those purchasing and operational decisions the way they typically do? Having the expertise being more local to the actual deployments? There's this "federal czar" ideology that comes from the 1950s, but I think the spread of technology and social development has reached a peak where this centralized expertise model actually holds us back. If we are /actually/ at a point where something like "ai.gov" should credibly exist, then we're definitely at a point where centralized bureaucracy isn't an adequate approach to manage outcomes.
boringg 3y ago
Working for the government should be in the same ballpark but typically work at the government pays a less as there is to be the perceived benefit of social good imbedded in your salary as well as there are power considerations (in this case impacting AI policy). That and your job is generally considered to be safer than working in private interest. The incentives aren't the same. For example judges take on their roll even though they would be paid significantly more in the private sector as lawyers.
- rlayton2 3y ago
  I'm not sure about the USA, but government jobs in Australia also are very stable with strong contracts and benefits (in general, every department is different).
spandextwins 3y ago
It's not the government's job to do that.
vinhboy 3y ago
As someone who used to work in government contracting, this comment is so freaking true. The whole thing about contractors is so spot on. The only slight difference I would add is that the contractors will NOT hire the best talent, they will hire whoever they can to fill the position, for the cheapest price, so they can claim to have fulfilled their RFP. It's a dead end.
- donmcronald 3y ago
  I haven't done much, but anything I ever deal with that involves the government and contractors is the same. The contractor has an astronomical margin and, in addition to not hiring the best talent, they don't care about their employees and have a high churn rate. These are also the same ones that are super eager to have AI replace their (unappreciated) employees. They don't care about quality at all and as long as they fulfill the contract with something barely working that's all that matters to them. I predict things are going to worse and worse when systems don't work. Having failures in critical institutions that are life altering for the people that are unlucky enough to be the outliers is going to become the norm IMO.
orochimaaru 3y ago
Not to mention that you will have your livelihood at risk because the Congress and President can’t agree on a budget. I’ve considered working for the government. But I’m not willing to put my paycheck in the hands of incompetent elected representatives.
meowtimemania 3y ago
I wonder if there's some way for the government to partner with one of the major tech co's (e.g. microsoft/google/apple) similar to how the military seems to partner with private co's (e.g. lockheed martin).
bumby 3y ago
This is probably an unpopular opinion in tech circles, but perhaps the goal of top tech talent shouldn't always be centered on amassing the most money? I get that the pay scales don't hold a candle to private industry, but they aren't pauper wages either. They're in the top 10% of the US as a whole. I suspect if that isn't sufficient, it might indicate a difference in value systems.
- Frost1x 3y ago
  We've abstracted so much in life as a transactional situation through which our system of exchange, money, holds power to do so much. It shouldn't be surprising that the some of most useful resources to accomplish tasks in life or even necessary to survive are so sought after. If we adjust the incentives of life a bit so people aren't constantly looking for another buck to make sure they can actually retire or pay their mortgage off, assure they have access to healthcare when they inevitably get sick, and so on we might find a lot more people would invest their time towards other contributions in society rather than those that they find that compensate them the most.
  - bumby 3y ago
    I agree that the transactional mindset is part of the problem. The irony is that the research seems to imply that an awful lot of our happiness can be improved from taking a different mindset.

mattnewton 3y ago

Can’t they just use the contractor loophole?

hiddencost 3y ago
Career bureaucrats have historically been more loyal to the government, IME, and have their incentives better aligned.
acdha 3y ago
They can but that has several widespread problems: the biggest is that you need skilled people in the government who can review bids and oversee work – otherwise you get a $10M contact form because someone at BigCo1 wanted to play buzzword bingo and keep their buddies employed, and the “overseer” from BigCo2 wasn’t going to saying the team, often former coworkers, wasn’t performing well because that’d invite honest grading on other contracts where the relations are reversed. Beyond that, you get various efficiency issues: contractor turnover tends to be higher, which is terrible when institutional memory matters or when you’re relying on their best people to keep a larger group productive, and there are often various contractual or policy issues making work less efficient. One challenge is that managing contracts is a lot of work, so there’s a temptation to go with the big companies who offer many services but that often means that they’re only good at one area and you get mediocre results in the other ones which were lumped into the same contract - often banking on the difficulty of taking corrective action meaning that people will accept low performance in non-core areas to avoid jeopardizing a core project by the same company. None of that’s intractable but it adds up and fixing it is hard.
- wslack 3y ago
  > the biggest is that you need skilled people in the government who can review bids and oversee work It's also very isolating to be the "one person" who knows the subject area. Having teams of folks is much more sustainable.

godelski 3y ago

Some of the jobs go up to $180k/yr. That's not too bad considering it doesn't look like it requires DC location. But others like the grad assistant for $22/hr is laughable. Still, I agree there's a problem with salaries not being competitive. I've interned at DOE labs before and lots of people don't stay for long.

ryanklee 3y ago
$180k/yr isn't bad in the context of Sr SWE in medium-cost-of-living areas. But in the context of ML/AI, it's pretty rotten no matter where. Hard to get the right people if that's the top.
j45 3y ago
Some say wage suppression is one way the much better paid bureaucratic class can keep their well educated potential replacements where they are, far away.
tomohelix 3y ago
I mean the grad assistant is basically a post-doc position isn't it? Considering the stability and perks from a government job, I think that is not too far from market rate. But go up and those 180k jobs would be easily 300-400K in the industry. I doubt anyone who can do that work would be interested in a measly 180K. Unless it comes with a visa...
- dheera 3y ago
  Almost all those 300K+ industry jobs are willing to sponsor visas. Most of these jobs on the other hand seem to be US citizens only. Some require a security clearance which a lot of AI experts would probably rather NOT have.
- dvngnt_ 3y ago
  "a measly 180k" is a wild statement on an absolute scale, but i do see what you're saying relatively.
- kjkjadksj 3y ago

aduffy 3y ago

Recoding America[1] is an excellent book about this problem, but it extends a lot further than just contractors charging inflated rates. The government itself has a culture that is so averse to trying new things and compliance with existing policy that it's actually dangerous to the careers of civil servants to do anything innovative, because that is one of the only things they can actually be fired for. The solution she proposes effectively would be an expansion of the US Digital Service (which the author spearheaded) into something that approximates the British civil service, which is actually highly functional and effective. If you actually care about this problem and have a spare Audible credit it's worth a listen. [1] https://us.macmillan.com/books/9781250266774/recodingamerica https://us.macmillan.com/books/9781250266774/recodingamerica

j45 3y ago
Bureaucrats are so risk averse to such an extent that the culture can have an allergy to trying to do things different or better too quickly (including innovation).
- fkyoureadthedoc 3y ago
  Did you just strip the context and rephrase exactly what the parent comment says?
  - j45 3y ago
    Not at all. Sincerely asking, does your question intend to pre-suppose that everyone must have and express a unique perspective and no one can independently have anything in common? Or some other faux pas such a comment would run a-foul of? When I saw original post, I had this comment. There was kinship with the parent post. Sometimes echoing your support with another comment is a valid comment too. If you may want more novelty and context and details: My professional experience includes working across multiple industries of bureaucrats, and that was my visceral reaction to the piece. Change usually means studying change, analysis, design and ultimately decision fatigue by committee. There's many hard working and great bureaucrats. My statement tends to apply to the leadership and management layers. In terms of keepign things the same or very slow and tiny changes incrementaly, there's a lot of neat reading around institutionalization as it relates to not only government but bureaucracy here. An exciting part is how the pandemic challenged and made it OK for bureaucracies to try things and to roll out programs that weren't fully thought out or known and update them regularly. In environments many people could not handle navigating. Most of all I'm flattered by the downvotes.. happy to take those for the parent too. Perhaps I phrased it in a way that might have triggered some bureaucrats since they exist in enterprise, government, education, healthcare and more. Or maybe it bothered the dopamine seekers in only novel and unique comments have worth.

mmedellin 3y ago

This is most acutely felt in acquisitions where those responsible for deciding what to build and buy have very limited understanding of software development. The failed rollout of healthcare.gov is the billboard for this problem. There's an effort to train civilian acquisition professionals on how to buy and build better software, but they are pretty quickly hired by contractors or the commercial sector since their new skillset fetches a larger salary than what the government can pay (even with promotion). I agree there needs to be a special schedule. I do think an overhaul of the civilian service is worth exploring.

mistrial9 3y ago
second hand knowledge in California is - the State is large and wealthy enough to have some echoes of the actual Federal govt.. and repeated words from direct software project participants was .. that mid-level bureaucrats from the "controlling" contract originator, would spuriously and without warning change the requirements.. sometimes so much so that it would invalidate large'ish portions of completed work.. and this happened all the way into the final month of a supposedly serious deadline, with failure to perform clauses and all that, at stake. A first-hand voice from California State software projects said - that the required prime contractor had 10+ person teams and $1m+ budget, and failed to build useful results. The final result from the project(s) more than once, was "oh oops, doesn't fit with -other framework- cannot be deployed". In other words, literally making things that are abandoned on Day 1 of deployment, for large money, with plenty of apparent pressure on "workers" to perform. etc..
wslack 3y ago
> I do think an overhaul of the civilian service is worth exploring. There are decades of reports and proposals about this. Unfortunately, we keep learning the same lessons over and over.

altdataseller 3y ago

A lot of those jobs aren’t really AI/ML related too just reading the descriptions. They seem like generic software/IT positions

vwcx 3y ago
This is how government hiring works. The job buckets are incredibly unspecific. Related: Open jobs at the CIA with the title "analyst".

dmead 3y ago

The US digital service tries to do this, but it just ends up maxing out the government payscale very quickly.

jlpom 3y ago

Same story in France and I suspect other countries

b8 3y ago

Yep, the US government gives contracts for a lot of stuff like cybersecurity etc. because the GS pay scale can't compete with the private industry standard.

curiousllama 3y ago

> The US really needs a special schedule for software developers There's actually precedent for this - SEC lawyers are on the SK pay scale, not GS. It's exactly the same problem: the alternatives in finance were too lucrative to retain competent enough attorneys to regulate wall street. Congress should create a separate tech pay scale, given the alternatives are too lucrative to retain competent enough engineers.

akira2501 3y ago

> and they really should have technical expertise in house Why? Moreover.. how would any of this expertise actually get applied top down in a federal system? How would any department making purchases or technology decisions actually benefit from such an apparatus? Wouldn't it be more sensible for them to manage those purchasing and operational decisions the way they typically do? Having the expertise being more local to the actual deployments? There's this "federal czar" ideology that comes from the 1950s, but I think the spread of technology and social development has reached a peak where this centralized expertise model actually holds us back. If we are /actually/ at a point where something like "ai.gov" should credibly exist, then we're definitely at a point where centralized bureaucracy isn't an adequate approach to manage outcomes.

boringg 3y ago

Working for the government should be in the same ballpark but typically work at the government pays a less as there is to be the perceived benefit of social good imbedded in your salary as well as there are power considerations (in this case impacting AI policy). That and your job is generally considered to be safer than working in private interest. The incentives aren't the same. For example judges take on their roll even though they would be paid significantly more in the private sector as lawyers.

rlayton2 3y ago
I'm not sure about the USA, but government jobs in Australia also are very stable with strong contracts and benefits (in general, every department is different).

spandextwins 3y ago

It's not the government's job to do that.

vinhboy 3y ago

As someone who used to work in government contracting, this comment is so freaking true. The whole thing about contractors is so spot on. The only slight difference I would add is that the contractors will NOT hire the best talent, they will hire whoever they can to fill the position, for the cheapest price, so they can claim to have fulfilled their RFP. It's a dead end.

donmcronald 3y ago
I haven't done much, but anything I ever deal with that involves the government and contractors is the same. The contractor has an astronomical margin and, in addition to not hiring the best talent, they don't care about their employees and have a high churn rate. These are also the same ones that are super eager to have AI replace their (unappreciated) employees. They don't care about quality at all and as long as they fulfill the contract with something barely working that's all that matters to them. I predict things are going to worse and worse when systems don't work. Having failures in critical institutions that are life altering for the people that are unlucky enough to be the outliers is going to become the norm IMO.

orochimaaru 3y ago

Not to mention that you will have your livelihood at risk because the Congress and President can’t agree on a budget. I’ve considered working for the government. But I’m not willing to put my paycheck in the hands of incompetent elected representatives.

meowtimemania 3y ago

I wonder if there's some way for the government to partner with one of the major tech co's (e.g. microsoft/google/apple) similar to how the military seems to partner with private co's (e.g. lockheed martin).

bumby 3y ago

This is probably an unpopular opinion in tech circles, but perhaps the goal of top tech talent shouldn't always be centered on amassing the most money? I get that the pay scales don't hold a candle to private industry, but they aren't pauper wages either. They're in the top 10% of the US as a whole. I suspect if that isn't sufficient, it might indicate a difference in value systems.

Frost1x 3y ago
We've abstracted so much in life as a transactional situation through which our system of exchange, money, holds power to do so much. It shouldn't be surprising that the some of most useful resources to accomplish tasks in life or even necessary to survive are so sought after. If we adjust the incentives of life a bit so people aren't constantly looking for another buck to make sure they can actually retire or pay their mortgage off, assure they have access to healthcare when they inevitably get sick, and so on we might find a lot more people would invest their time towards other contributions in society rather than those that they find that compensate them the most.
- bumby 3y ago
  I agree that the transactional mindset is part of the problem. The irony is that the research seems to imply that an awful lot of our happiness can be improved from taking a different mindset.