Hacker News Reader

Comment by gregdoesit | Hacker News Reader

It’s not just Latin America. It’s also Canada, most of Europe, it’s India, Asia, and I’m hearing Africa also experiencing a similar pull.

There’s a global talent shortage for experienced people in software engineering, and it’s spilling over everywhere.

Remote work becoming the norm thanks to the pandemic, plus the rise of services like Remote.com, Deel and similar ones is making it much easier to hire remotely in most countries - and hiring outside the US is easier and cheaper: especially when you pay above the local market (but we’ll below the US one).

I’ve been covering this trend from mid 2021 both in my newsletter (The Pragmatic Engineer) and my blog. From all evidence I gathered, we are in the most heated tech hiring market of all time, one that is hotter than during the Dotcom Boom (details in [1]).

Having talked with closer to a hundred tech hiring managers the last six months across all geographies, the consensus is that it will get worse in Q1 2022 than before - and, obviously, this means better for many experienced engineers. And H2 2021 was hot enough with out-of-cycle compensation increases of 5-30% on top of annual raises at many tech companies, across all geographies [2].

[1] https://blog.pragmaticengineer.com/advice-for-tech-workers-t...

[2] https://newsletter.pragmaticengineer.com/p/more-follow-up-hi...

View on HN

This is exactly what university students in the U.S. and abroad have been faced with. Most engineers with a paycheck simply don't get the problem. Competition is higher now than ever for entry, companies do not want to hire graduates, they do not want to invest resources into them. They want the maximum amount of short term profit possible. Build things fast, sell them, next project. We are racing to the bottom and the natural result is the U.S's software dominance will be gone sooner than later. Just like manufacturing we are seeing an exportation of work en mass.

We're not really exporting work as much as hiring people to come over here. Or, it's not as clear cut as just exporting manufacturing.

Universities aren't reliably producing CS graduates that can code.

If a student enters a collegiate program with some interest or experience in programming, then they're likely to come out of it with solid skills and find some opportunities. If an employer has the resources to select from the best that universities have to offer, then they can find great candidates.

But for us that aren't working for FAANG, a university degree doesn't really tell us much, and certainly tells us a lot less than a portfolio of projects or work experience.

This is still fundamentally a hiring problem. There's no way to sort candidates by skill that doesn't involve a ton of labor. A CS degree sure ain't it.

This is the sweet spot that products like LeetCode or BinarySearch could be used to solve - people who have the academic background, and should know how to code if they're coming out of a good program, but don't have a portfolio of work to show or a catalogue of experience to draw examples and answers from for a typical interview. And all the DSA stuff that is irrelevant for 90% of dev jobs 90% of the time is still fresh in their mind.

Ask them 2-3 LC easies/mediums in the language of their choice for them to prove they can actually write code, and that's really all you need. Unfortunately it somehow became "let's have a 5 hour long two-part panel interview where we ask you half a dozen LC hards and oh yeah don't google anything" as a way to hire experienced people who have a decade of work they can talk about the discuss ad nauseum.

Not going to lie, as soon as you mentioned LC I was disinterested. But your suggestion sounds good. A few easy/medium questions and maybe the ability to Google sounds fair. I sort of enjoy basic fizz bang code screenings.

And just to be clear I'm only suggesting it for junior positions. Most of the time I think you can suss out whether someone can code by discussing their direct contributions to previous projects. If not, you can always do one LC problem.

I see it as a great way to just run a sanity check that this person who graduated from Random State six months ago can actually code, and as a great way to ensure high quality very senior people refuse to go through your interview (unless you're paying FAANG wages).

I have about a decade of real professional coding experience. Not going to say I'm excellent, or near the caliber of developer FAANG are looking for, but I can write code. I can count on one finger the number of interviews I've got in the past couple years. Zilch. There is a massive disconnect from what you hear on the news, and the reality, where somebody like me is a pariah and the deafening silence of _any_ interest.

This is just outsourcing 2.0, this time under the guise of a lack of qualified candidates.

You either have a location problem or more likely a CV/keyword problem.

If you're still breathing and have more than 2-3 years of dev experience you should at least be contacted.

What do you mean by "someone like me?" Your reality is not any reality I've experienced.

I work for a company nobody here has heard of, working on a boring tech stack. Absolutely nothing I do day to day would make into any blog post, let alone anything on HN. I, and my ~dozen or so coworkers, get at least one cold recruiter contact a week. Obviously most of them are garbage. But we've all gotten the random contact from an Amazon or Microsoft or Facebook recruiter. The interviews are available.

YC runs workatastartup.com - I submitted my resume and a two-sentence intro to 5 or 6 companies and I think I got an interview at all but one. The interview is the easy part, and if you can't even get that with a decade of programming experience, you're getting caught in some arbitrary filter. Which is to say, respectfully, you're doing something wrong because that's simply not what the market is right now. Or, you're not quite as qualified as you think you are.

I can't do LeetCode to save my life. I'm not that kind of natural Google-tier genius.

But I have a relevant PhD, demonstrable industry experience leading a recognisable project, many papers, lots of experience developing juniors, influential keynotes, blog posts, etc.

If you used LeetCode you wouldn't hire me. I don't know if you think that's a loss or not? But it's a data point.

Nothing is perfect.

At this point in the game hiring based on LC isn't optimizing for hiring people who can code (I think that may have been the case 3-4 years ago), so we don't use it for anyone. Now it's just optimizing for people who have A) heard of LC (who cares); B) have a basic understanding of DSA (good); C) have taken the necessary 0.5-3 months to memorize the handful of patterns that show up (who cares). I can understand why companies like FAANG that get thousands of resumes use it as a filter, though.

But I will say in my experience a PhD is a red flag for a developer, especially if you're highlighting your academic experience on your resume. It's just that the skills that make you successful in getting a PhD don't necessarily translate to day-to-day software development and in some cases will hinder you/the team.

Universities today seem to be focusing on quantity of education not quality

There goal is to get as much student loan money that they can, they do not seem to care about the quality of education the students are getting

This goes for all levels of universities, and all degree programs.

Universities were never job training programs, their product has always been the right to engage in class signaling. By paying the university a pile of money you signified to potential employers (and everyone else) that you were a member of at least the upper middle class, with the financial resources (and sometimes family connections) to support making those payments of time and money. Not having that signifier was a signal that you lacked the time and money to dump into that effort, given that you had to spend so much of it surviving.

It's the same human impulse that drives people to bind their feet, value bleached skin, engage in conspicuous consumption, etc. It's all an elaborate signal game designed to convince people of your social status.

The problem is that we looked at that system and instead of trying to build something better we dumped more money into it in the form of student loans and expected that now more people will be given access to those class signifiers and thereby raise their social status and standard of living. In actual practice, of course, what we did was raise the bar on what qualifies as a class signifier, forcing a generation into wage slavery with little real benefit to them or to society as a whole (other than those institutions who siphon off those extra dollars and use them to metastasize extra layers of administration and management to little effect)

What we need is for education to be more job skills training and less social positioning. Funding for adult education should be linked to the success rate of students leaving those programs. If you have the money to burn studying topics that will indicate to your peers how little you need the money, then great. That's apparently the way we've decided to structure things. For the rest of us though let's try to encourage study of topics that will help society work better instead of vainly trying to convince the rich kids club to let us in.

I agree with all of that....

"Universities aren't reliably producing CS graduates that can code."

And companies stereotype all graduates as worth nothing to them.

I had a couple simple Android apps when I graduated. Even though they were simple, it would show that I could follow best practices, code, test, and deliver something. I had a decent GPA (3.5), clubs, etc. I still had a hard time finding companies that would even give me an interview.

So sure, a degree doesn't mean too much (my masters has done nothing for me). But it seems companies have simply given up and are exacerbating the very problem they are creating.

How are you on leetcode?

I believe LC wasn't a thing back then, or at least not mainstream.

I don't waste my time on LC now. If I get free time, I'd rather work on a personal project or hobby.

Granted, I'm actually thinking of moving into some sort of corporate strategy analyst role since I don't really get to code anymore. The past 2 years has been very little coding or business problem solving. It's mostly been config, infrastructure, prod support, and meetings/paperwork. I'm tired of it. I want to solve problems and build substantial things. I assume I'm rusty when it comes to coding now.

Oh, I didn't realize you were talking about a time long ago.

Yeah, I'm old. That was 10+ years ago.

Nah, you're a youngster.

You're only as old as you feel. I feel like I'm ancient.

A friend of mine who died recently worked on Stretch at IBM. Later, he was one of the early employees of a hot cloud computing startup, in 01968.

Yeah, but there's increasingly no place for me in the working world. I'm decent at a lot of things, but not an expert. I'm no longer relevant.

You may enjoy https://www.smbc-comics.com/comic/2012-09-02

I've interviewed a lot of software engineer candidates. It's always surprising how often people with impressive resumes, including computer science degrees from good-to-great universities, can't code at all.

I'm not talking about trick "do you remember A* search" questions. I'm talking about the ability to write a basic program and to reason about what it will do.

I've seen this across the gamut, from new grads to staff engineers.

Part of this is selection bias: those folks probably apply to many companies before they slip through somewhere, so they're overrepresented as interviewees.

My sense is that it's becoming more common. Undergrad CS has ever more people who are in it for reasons unrelated to enjoyment or curiosity.

Plenty of people here on HN will tell you they choke in interview situations. A lot of people just aren't comfortable and it gets to them.

That said there are people who really can't code, from my experience working with such folks.

Currently in a Canadian college for cybersecurity and the content of the courses are being nerfed and the quality of the graduates are churning out are largely dogshit.

I suppose I was once part of the problem back when I marked assignments in grad school. We're heavily incentivized to not fail students. There's virtually no difference in grades between a 10x programmer and someone who can barely code fizzbuzz in their 4th year.

This is in part due to blue politics in universities pushing for more and more genericism in education, suiciding their skill output. The other part is the companies that should be involved in education and funding educational programs simply don't care to. This is the one single way to know if the "shortage" is real: is the company willing to invest in education? The answer has been no. Fewest scholarships, fewest interns, they don't fund programs, and they treat students as cash cows.

It makes sense when looking at the numbers. No use in putting money into U.S. education if they don't have to. The law allows them to import workers from elsewhere, parasiting off of their social educational programs and never having to pay a dime to them.

The solution is clear, we have to make it more economically viable to invest in the U.S, and that is by removing or severely limiting the mechanisms by which tech companies employ foreign workers. Or alternatively we start actually taxing these companies.

> Universities aren't reliably producing CS graduates that can code

Have they ever been?

A cs graduate isn't somebody that can code; those are only loosely related to each other

> There's no way to sort candidates by skill that doesn't involve a ton of labor.

This is the exact concept of an IQ test. They aren't labor intensive.

My first job was at a startup and they simply didn't have any money to hire experienced people, so it was petty much only people fresh from uni.

You see the same at other smaller companies who will rather hire a student or a fresh graduate because it's much cheaper (and more available).

Racing to the bottom by paying talent more and more and more?

I feel like companies are investing even more in the long term than ever before, building increasingly ambitious infrastructure projects.

I don't see what you see at all.

I haven't seen any company do that in a very long time. What industry are you working in?

I work in e-commerce at Shopify. My company is building some great developer experience stuff for the long-term like some truly excellent cloud-development environments.

My previous company, Oracle, was developing an entirely new kind of language virtual-machine to run the programming languages of the future as well.

I see this in many places.

That's just long term product planning. The rest of the people here are talking about long term employee development - making your employees more valuable so they produce better products rather than treating them as costs that must be reduced.

> rather than treating them as costs that must be reduced

But employers are paying more and more than ever before?

Paying more money doesn't mean you're investing into entry level employees necessarily.

The goalposts are all over the place in this thread.

These people must be working for terrible companies. Come and work for Shopify! We even partner with a university to offer year-round internships while you study for your degree. I regularly mentor juniors to build them up to research-level engineers.

> The goalposts are all over the place in this thread.

yeah, they really are. no one said anything about terrible companies, smh.

Not most of the employers I deal with. Wages are exactly where they've been for decades.

If I wanted to get a tech masters degree, my company would pay for it. If I want to learn a new technology, my manager will absolutely allow me to spend time doing that. I’m constantly working on projects that force me to learn new technologies. I feel like my company had no problems with my becoming more valuable. It’s mostly up to me.

I dont agree with your rather pesimistic view on the matter. I know many companies having fresh new graduates in their hire strategy because that just works in the long rong. Proven fact. Also most companies have smart people not shortsighted and they come up with proper long term strategies. There might be foolish companies as you describe but not the majority.

This is not true. I work in a FAANG-like company. Talent shortage is real - we are having to hire entry level grads where we originally wanted experienced candidates. We are even going out of the way to hire candidates from non-traditional backgrounds and train them.

Maybe your company is the exception.

The fact that there is a talent shortage seems to point to a long-term issue regarding barriers to entry across the industry.

That's news to me.

In December, I applied to a number of blue chip companies for frontend positions and only got callbacks to three of them despite a lot of work experience writing JS for real applications including a YC company. Some YC companies also said to me they wanted someone more experienced in Vue/React instead of potentially allowing me time and space to ramp up my knowledge of it. So clearly there were other applicants who had both a lot of work experience AND the precise tech knowledge they needed, so they didn't have to take a risk on someone who didn't perfectly fit the position.

I eventually landed a dream position, but a huge reason why they looked at my application in the first place was because I knew of a long time employee. Obviously I had to pass the technical and behavior interviews, but they had a deluge of applicants and my application would've been lost to the ether had it not been for that connection.

Seems weird. If they're willing to wait many more months to hire their perfect candidate, couldn't they do it faster by hiring somebody good and allowing for some ramp up?

The "talent shortage" is self caused by FAANG. Your company does not invest in education, yet expects people to be experienced. Accreditation for compsci and others is sorely out of touch with the tech industry, and university programs are not funded. We are literally strangling the life out of young adults with course materials that don't matter, expenses they have to work manual labor to pay, classes that don't teach, and making them have to sort through dumpsters to find an Arduino. It's that bad. What do tech companies do when asked to sponsor a program? No response. They don't care unless it's from a FAANG employee themselves, if that.

These companies are meant to be heavily taxed for how much they are taking from society, with those taxes put back into education. Yet here we are.

> I work in a FAANG-like company. Talent shortage is real - we are having to hire entry level grads where we originally wanted experienced candidates. We are even going out of the way to hire candidates from non-traditional backgrounds and train them.

I am currently applying for an entry-level position at a FAANG company. I had "in-person" remote interviews a few weeks ago. I was told that I passed, and I should expect some "team fit" interviews the following week.

Except over the intervening weeks, I've had zero team fit interviews, because - as far as I can tell - the recruiter handling me hasn't been able to persuade even a single manager to agree to an interview. This is not a situation that screams "we are experiencing a worker shortage, and we will even lower ourselves to hire inexperienced workers that we have to train". It would tend to suggest that the company is swimming in far more applicants than it wants for every role.

They want the maximum amount of short term profit possible. Build things fast, sell them, next project.

Not a graduate but instead an industry veteran. If what you say is accurate then Im happy to have my bias of “maybe it’s time I go the mercenary route” confirmed.

Surely the market is there, right?

The talent abroad ends up working for US companies though, how will dominance dissipate. The market and financing is in the US.

Anecdata but in my experience it's definitely true in Africa, I did a project for a Cameroonian tech company recently and they couldn't afford to compete with major tech co's with remote roles which is where anyone decent and semi-senior went. They ended up with a mishmash tech team in Algeria and Ukraine mostly although they would have much preferred to hire local. I heard this story many times in the tech community there and in Kenya.

Same in Tunisia. Remote and France have swooped practically most of the tech talent. You are forced with either mediocre developers or to pay EU rates.

Shouldn't mediocre be generally acceptable? It seems unreal to expect the majority to be above average.

Problem is, if you are outsourcing remotely, you want someone who is quite competent for him to be able to work on his own and pick up on the lack of face to face meetings. Mediocre as a result will give less than mediocre results or no results.

Or there's a need for communication. It doesn't matter if the person is "quite competent" if they're on their own building something that doesn't fit the requirements. An average dev shouldn't have any issue being able to Google the occasional problem. The vast majority of the issue we have with outsourced work is subpar English communication.

> Or there's a need for communication.

> the issue we have with outsourced work is subpar English communication.

If you have good English command, you are already worth something on the Internet. Also, someone not from an English speaking country will need to pick up English by himself and practice frequently.

I'm not worth anything.

My experience is mediocre is insufficient in SW preojects. You NEED some good technical pillars, else it will all be shit.

Then 75% of software is shit?

Yes.

This feels like a gross over exaggeration.

Sturgeon's Law, 90% of everything is shit

Isn't that law basically saying that absolutes aren't 100% accurate?

No. It's saying exactly what it sounds like, that most of everything is at a very low level of quality.

Historically, it was a response to the argument that science fiction was a genre that didn't deserve to have any attention paid to it, because look how awful some science fiction books were. Sturgeon's response was that most science fiction books were terrible for the same reason that most work in every other genre, past, present, or future, is terrible.

https://en.m.wikipedia.org/wiki/Sturgeon's_law

Did you read your link?

> Sturgeon's law (or Sturgeon's revelation) is an adage stating "ninety percent of everything is crap."

> Sturgeon deemed Sturgeon's law to mean "nothing is always absolutely so" in the story "The Claustrophile" in a 1956 issue of Galaxy. The second adage, variously rendered as "ninety percent of everything is crud" or "ninety percent of everything is crap", was known as "Sturgeon's Revelation", formulated as such in his book review column for Venture in 1957. However, almost all modern uses of the term Sturgeon's law refer to the second, including the definition listed in the Oxford English Dictionary.

Yep.

"Sturgeon deemed Sturgeon's law to mean "nothing is always absolutely so" in the story "The Claustrophile" in a 1956 issue of Galaxy. The second adage, variously rendered as "ninety percent of everything is crud" or "ninety percent of everything is crap", was known as "Sturgeon's Revelation","

Not to mention, it's not a scientific law, but just an adage. Even then, when there's a higher barrier to entry the law is thought not to apply (such as technical skills required to create professional software).

On the ground, this seems to be demonstrably false. I have numerous current and former coworkers that would love opportunities elsewhere.

It’s the recruiters who are the problem. If you’re looking for a C# dev and you put “5 years Node.JS experience minimum”, then not only are you going to miss out on some great, if not the best, developers, but you’re much less likely to hire the man you actually want.

I'd go further and say it's not just the recruiters that are the problem, but the automation of recruiting that is the problem.

Hiring is a people skill, not an algorithm skill.

Preach!

I think US salaries are more to do with it than anything else. It's just not sustainable anymore for most tech companies. So they hire where talent is cheaper. It was obvious it will happen with remote work.

>It's just not sustainable

I'm not sure if it's a sustainability issue or tactic to eek out every dollar possible given the landscape. I suspect it's the latter. I'll bet most don't do the leet code interviews for offshore developers though.

The US salaries force you to go in more competitive markets though and use your engineers better, because your costs are higher. Salaries in France have been 1/3 of US salaries forever, and yet people don't outsource to French engineers, more like French engineers move to the US.

If you outsourced to French engineers you would have to accommodate all of their rights and work styles which don’t exist so much in “developing economies”. A French engineer probably would not be online at 6am and 11pm on the same day to attend meetings and solve problems for their American managers. Eastern European, African, or SE Asian engineers would though.

I’ve hired French engineers in France and they’ve done great work. Hard workers too.

But the impression France has is probably a turn off for many. There are other problems too like it taking 3 months for a new French hire to start. Getting rid of them is extremely difficult as well. There are lots of additional taxes you have to pay the French government too. There’s just so much risk in hiring there that you’re better off going elsewhere with more modern employment law for skilled, high demand workers. Don’t remind me of the monthly paperwork we have to send the government to assure them the French employee isn’t working too much and is taking vacation. An absolute joke.

The French government doesn’t make French engineers very attractive unfortunately.

Haha, modern employment law :D You'd cause hilarity if you used that expression anywhere in the Scandinavian public discourse, or at least the Norwegian. You'd get tomatoes thrown at you and no one would listen to a single word.

Not that I necessarily disagree with you or miss your point, but this is a very strong ideological divide. It's not about modern vs. archaic unless you posit that employment law that protects employees is such a detriment to an economy's effectiveness that it's effectively obsolete.

Maybe at some point it will be a question of what economies actually manage to get things done and thrive and those who don't, but that's the kind of long-term, almost geopolitical shift that happens over decades at the least.

E.g. no one who wants a position of power would advocate for a pure planned economy today, rather than a market economy. I sort of doubt that employee protections are such a millstone around the neck that they will go the same way, but who knows.

It's an interesting question as you see a tendency of economies with less protections and higher salary luring away lots of really high performers.

What I mean by that is how worker protection laws from the Industrial Age and which apply to hourly wage employees are applied to professional workers that have autonomy and other attributes that make them very different from hourly workers.

The USA has a concept of an “exempt” and “nonexempt” worker and a series of questions that determine this status. In general, hourly workers (nonexempt) get many protections encoded into law that professionals don’t. And professionals don't want them except for a very small minority of oddballs that want to unionize.

So in essence, applying the same set of outdated rules to everyone. It even makes it hard to compete with colleagues for promotions if your hours are limited. Of course the French engineers I had lied to the government about hours worked as they wanted to maximize bonuses, stock grants, and promotions.

The legal protections on hourly workers aren't enforced in the US or are skirted around by making hourly employees contractors. Professionals would like some of the protections but only if they are actually enforced.

The law defines what a contractor is vs employee. However it can be ambiguous in some cases, especially modern “gig employee” work. Existing boxes don’t fit well. We need new law for this class of worker that wants freedom to choose hours, platforms (like a contractor) but has employer set rates and tasks (like an employee) where there are certain protections beyond a 1099 but also allow these businesses to work, as well as the workers having freedom.

I’ve been at and seen plenty of places where a growing layer of middle management was paying for itself by outsourcing engineers. Outsourced engineers very often need much more management and company directors either lose touch or decide to run the company in ways that justify raising their own salaries.

It’s not that the cost of engineers is unsustainable, it’s that aging companies tend to want to take power and decisions away from engineers towards management and accomplish this with outsourcing.

This has happened in my wife's rather staid manufacturing industry. Many plants in the US have no on-site engineers. The "engineering" design is performed by remote fresh-outs at the exurban headquarters, and a lot of that work is simply regurgitating the specs from the equipment manufacturers. Who quite rightly charge exorbitantly to have their own engineers perform the inevitable changes that would be nearly trivial for an in-house on-site engineer to make. It's not unique to her particular company; this is the way her industrial customers and suppliers work too. The backend software is managed the same way. Buy Oracle/MS/Google whatever integrated functionality, and hire consultants to make changes. No expertise in-house.

No in-house expertise means no management responsibility for failure: it's the supplier's fault. Yet those same vendors are more often than not locked in by prohibitive replacement economics. Gruesome for old skool highly competent engineers, like my wife.

Most of these companies are profitable, so who can argue? This is the present and inevitable future.

This is how startups succeed and take market share very quickly. A small number of highly paid highly competent people build something which is significantly cheaper and better than the incumbent because they aren’t weighed down by very large numbers of unnecessary people doing things poorly in the most expensive way possible.

Then the startup either gets bought by the incumbent who has existed so long they just have piles of money or the startup grows into a similarly inefficient monster.

It is a sign that there is something wrong with the game created by the economic and legal environment which tends towards a large proportion of useless work and barely adequate quality. I.e this is why we can’t have nice things and it’s not exactly clear how to fix it.

> and it’s not exactly clear how to fix it.

Constantly creating new companies to outcompete the old corrupted ones. In other words capitalism, the solution is to ensure that competition never dies. No country in the world has found a better solution to this.

I guess the problem currently is that the incumbents aren't actually getting replaced. Doing a good job is just a means for a small number of people to make an enormous amount of money while everyone benefits from their quality for perhaps a handful of years before they're bought out to be eaten or destroyed by the existing big players.

There is a problem with competing with big players for more than a few years because they figure out ways to push you out of the market before you can beat them, with only a few exceptions.

That is the problem with the environment which needs to be solved.

The ERP that my company uses is proprietary and they charge large sums of money to make custom changes. I am an in house developer that has gotten to know the database backing the ERP and have extrapolated functionality based on stored procedures and table schemas. The work I do would cost my company much more than my salary and I build interfaces that connect to marketplaces or marketplace API aggregators such as ChannelAdvisor. I can attest to your statement that in house engineers can cost much less than the modification engineers who know the system but will charge an arm and a leg.

As a matter of fact I've carved out a niche in these matters over the years. Essentially reverse engineering systems and building out functionality instead of the original creators of the software.

EDIT: The ERP company never made it not possible to interface through their DB in their contracts or by encrypting their functionality in the DB.

The amount of value I bring to my company is greater than 100 fold of my salary.

Yeah, that's an obvious thing to do, right? However, the company armored themselves against this sort of competency attack against The Machine by having in-house IT by policy gate keep access to the ERP db tables, which are exactly as you describe, and can be reverse engineered. This does have the additional side effect of making it quite difficult (in practice impossible) to implement in-house statistical process control. Which, again, makes it much more difficult to, um, need to be diplomatic here... discover which processes could be improved. Much better to outsource to another corporate consulting parasite a multiyear/multimillion $$ effort indoctrinating the troops on the abstract importance of process.

However, I enthusiastically applaud your success! Seems it might be a tightrope act to balance "Ima worth a bunch of money to you" vs. management realizing "that nerd is a massive SPOF".

SPOF can be applied to the ERP company themselves. Services fail all of the time. DockerHub failed on Friday with their container API which you could argue made it impossible to know the outcome of your builds but you could still get by and pull your containers after pushing repositories. As someone who understands how I can be a single point of failure I make sure to ask for raises within reason and since I love what I do usually means at market rate. It's not like I am holding a company hostage by knowing what value I bring. Not everyone is of the persuasion of extorting their employer due to what their perceptive value is. My time is valuable but not so valuable that I price myself out of a market by my ego. To add to the backstory I have been an employee of the company for the last 4 years with never an intention of leaving nor an intention of taking advantage of any perceived leverage. Some people will disagree and say that I should advocate for more at all times. My perception is either I start my own company and build my ideas/imagination or I keep doing what I do. Why does everything have to be leverage in every situation? I live and work in the Seattle area and may not get paid the same amount as Amazon employees but I certainly get the freedom and encouragement to pursue ways of furthering the company and their future in revenue generating projects. I tend towards the I will give ample notificationt that I will be leaving and will help in making my own job automated eventually. I'd rather leave a trail of automated services than leave another set of people in the lurch because they took a chance on me. One day when I decide to leave there will be enough Go programmers around to maintain what I have built also. Either my services will run until the integrations change their API or I leave knowing that what I have built will maintain itself if I were to take sabbatical and in that case it is no longer my responsibility. The argument that a single developer is worse than a group of developer is old hat when it's clear that an entire team of developers can up and quit and be a single point of failure either in an individual company with a set of nerds or their outsourced company where all developers decide to unionize or the brain of the team leaves and the remainder of the people may be able to sustain the service or don't have the capacity to. Third party services are just as susceptible to these forces but fear should not be a driving factor to hinder the furtherance of your business.

Hmm, so 99% of the value you create is being skimmed off by the company's shareholders and management? Have you thought about trying to renegotiate to a more equitable split, like 10%/90% or 50%/50% instead of 1%/99%? Or is that impossible because you're in a very weak bargaining position, like a migrant cannery worker?

That seems like a major reason people might quit jobs like yours and go work for ERP vendors or other outsourced vendors: even if they create less value, they are in a better bargaining position and so they can capture maybe 5% or 10% or 30% of the value they create instead of less than 1%.

I would argue that if I went to work for the ERP company themselves that my "bargaining" power would actually decrease due to the revenues that they are garnering compared to the company that utilizes the ERP itself. It also assumes I am unhappy with such an outcome that my work is worth 100x my salary and somehow that in itself is reason enough to ask for more. I have ideas that I share with no one that are worth much more and one day soon I will build them myself. Along the way I will have made good contacts in business and investor potential all while not rocking the boat all of the time. I could be completely wrong and would love argument against without down-voting to convince me and potentially others that this in fact is the wrong way to think about things.

I don't think it's a wrong way to think about things. Multiple ways of thinking about things are valid.

It's possible that your bargaining position vis-a-vis the ERP company would be worse if you were an employee. But probably not if you were a founder. And surely the ERP company's bargaining position relative to your company is better than yours.

> It's possible that your bargaining position vis-a-vis the ERP company would be worse if you were an employee. But probably not if you were a founder. And surely the ERP company's bargaining position relative to your company is better than yours.

I don't think that this is the case as the value they add comes with opinion entrenched in the belief of the path that their product should take. This opinion actually has caused them to flat out refuse to do some work which in my opinion is an opportunity to make things better for the company that has originally taken on this ERP system as their central authority of information. I actually came in from a friend of mine who knew that I was familiar with Coldfusion. Ultimately in my exploration for knowledge I have become intimately entrapped by the ERP data model based on the simple fact that it is esoteric and inflexible which has in fact been proven by how the front-end interface is a kludge of C++ utilities, C#, and some shell scripts. Upgrades continue to happen but they give me 0 faith that the small company that they are (ERP) has a clear path to less of a kludge of a system. It gives me solace that as long as their opinions are strong contrary to the revenue model of my employer that in fact I will be useful for a long time. The moment the ERP company actually understands the niche they are pursuing is the moment I will have less of a potential for work. Until then they can fight their way upstream against their customers as much as they like and their customers will hedge their bets until they can migrate to something better. It is assumed I am less in line with the directive of the customer and have my own agenda against the ERP company when in fact I am elongating the runway for my company to operate until they have a better plan to migrate away to a solution that doesn't assume so much about their customers. Manufacturing knows they have to evolve yet they have made past decisions that they realize they can only mitigate by extending their systems without having to pay the high prices their vendors request. How is this not a good strategy? It's not like I am a contractor and trust me when I say that my company is better off with my SPOF than just the SPOF of their ERP company. Lose the ERP company and you not only lose the company and their regular updates but you lose the custom development of that company which will be 10 times more when the developers that were fired become independent contractors themselves.

It's just business.

> actually has caused them to flat out refuse to do some work which in my opinion is an opportunity to make things better for the company

What were the consequences for them of flat out refusing to do that work? What would the consequences be for you if you were to flat out refuse to do something similar? If they would be worse for you, as I suspect, that means their bargaining position is better than yours.

What does your management believe the consequences for them (the management) would be if the ERP company stopped working for them entirely, for example because they got out of the ERP market? What does your management believe the consequences for them would be if you quit? If they would be worse in the first case, as I suspect, that means their bargaining position is better than yours.

This is only loosely connected with how much value each of you actually contributes, which is what matters most in a moral sense. Unfortunately, how that value is distributed among stakeholders usually has more to do with how negotiations go than with how the value is created. Indeed, it's common for negotiations to happen in cases where the net value produced is actually negative, as in lawsuit settlement negotiations, and the question is not so much how to split the winnings as how to split the losses.

>It's just not sustainable anymore for most tech companies.

Translation: Most tech companies are addicted to cheap labor, either through exploiting new graduates or handing out potentially worthless stock options. They won't choose to reduce their profits regardless of what happens.

Very few startups "need" to hire less experienced people, mostly the ones that do just haven't got a viable business model. The ones that can't afford to hire the people they need at a reasonable salary for the work they expect shouldn't exist. They're just machines for turning venture capital into personal wealth for the founders as they exploit their employees then sell out.

I mean, if you were CEO, would you choose to reduce profit when you have investors on your back or want to grow staff quickly? Outside of VC driven cash cows or big tech, there's not that much money to go around at the bottom.

I think you really mean that in your experience there's not that much cash that makes it to the bottom... more cash COULD filter down if the corporation was run that way. Most corporations aren't.

>>It was obvious it will happen with remote work.

Not to many, Many still believe that remote work for programmers simply means lower cost of living for them while keeping their high salaries...

They are about the learn the lesson US steal workers, and other blue collar works in many American industries did when globalization hit them....

The truth is that most software isn’t all that special. For the average business that can accelerate their processes with CRUD apps you don’t need to do much in the name of bespoke work. I can have a standard run of the mill CRUD app up in like a week or maybe if sufficiently complex a month or so - with proper tests. That’s the kind of work I’m seeing outsourced in droves typically.

The kind of work I actually do nowadays I don’t see getting outsourced so easily. This is core architecture and fundamental differentiation that the business sees as key to product and core success. You’d be highly unwise to outsource that.

An experienced developer that can produce high quality work but can accept a lower salary (nominal to local markets of the employer not the developer) based on geographical concerns? You can make a lot of money with those CRUD contracts I imagine

Edit: this all assumes someone competent is overseeing the work and someone that understands the technical things involved in the overall project to steer it know what they’re doing too

While simple CRUD apps are not technically challenging; the hard part of doing these is figuring out the specs and communicating with the client. I can see how this can go wrong with offshore clients (time difference, cultural difference, language difference, only remote). There could be a market for local agencies outsourcing tech work (and even that has to be done carefully). But clients directly outsourcing to off-shore agencies, that spell trouble.

A company I worked for struggled for years trying to find the right CRUD system.

Off-the-shelf packages didn't work right because they couldn't be tailored to the existing workflow, and couldn't be integrated into other systems. Three attempts at offshoring all failed because of culture and communication differences.

In the end, it was done in-house. It took longer, but the application is exactly what was required, and new features can be added in days or weeks, not months or quarters. Last I heard, almost everyone was happy with the home-grown solution.

CRUD can be simple. But CRUD can also be hard. Anyone who thinks that they can spin up a generic CRUD and solve any problem is someone who doesn't really understand what the problem is.

That is not what I mean. I am talking about well…when they are generic applications that don’t require bespoke work to be done. Maybe HN is the wrong crowd for this as I think most of us work in more specialized capacity. There is a lot of software out there however being written that is say, wiring power BI applications with predefined requirements, or collecting information via surveys that need to be tied to a CMS. Stuff I’ve seen work pretty well when contracted assuming the person overseeing the venture was competent of course.

I’d say anything core to a business in terms of how it functions is not ripe for outsourcing, and anyone doing this is going to feel a lot of pain

>There’s a global talent shortage for experienced people in software engineering

No, there's not. There's a shortage of cheap experienced software engineers in the US.

Reminder, every “talent shortage” is actually a wage shortage. Double those salaries and you’ll find plenty of talent.

Yeah at this point, swe are more key to bringing in the future efficiency advancements per person than doctors or lawyers, I remember doctors making multiple hundreds of thousands a year decades ago, with surgeons into the millions. It seems reasonable that swe salaries surpass that as software eats everything and our ability to disseminate engineering Skill stagnate.

Not neccessarily. Say there's 3 million competent senior software engineers in the US, and 75% of them are already working as senior SWEs (the remaining 25% are in early retirement or have switched professions). The best that doubling wages in short term can accomplish is makes those 25% of people move back to the profession.

As for people who don't know how to code retraining to be SWEs - since becoming a "senior software engineer" takes at least 5 years, you'd have to wait 5 years to see a result, irrespective of how much you increase the wages.

Believe me you start paying those guys 800k a year they’ll wait a few more years before retiring.

That was part of my point...

This is not actually entirely true (although I've had similar sentiments in the past).

The question is, how inelastic is supply? "Labor of a senior software developer" isn't like "a widget" in that you can pretty simply create more of them. It takes years and special training to create senior software developers.

This means the supply is inelastic over the short term, although in the long term more people get the education to become one.

That means that doubling salaries might not have the effect you predict.

While, generally, that is a good solution, I think this is not true in this instance. Realistically speaking, SW people are paid very well relative to the average population. And you can't really just mint them on demand. Time to bring up to a SW dev is long. Since the pay is so high, I think most people who had the ability and interest are in SW already. So the supply is pretty much fixed. What would increasing pay do? Just heat up the market. They probably won't get more manpower in the market, but just poach from eachoter. Great for devs, bad for companies(stock holders? CEO? I don't know, modern companies are so nebulous and diffuse, I don't even know who is to gain from them) bottom line.

I posit that the "hot" market is the marker of inflation. The money running around leads to more competition for talents and competitive fields is where, I assume, inflation should show up most easily.

I know I might sound anti-worker, I swear I am not, I just attempt to get as close to the truth as my small mind allows me.

High wages do not manufacture talent. Spending more money won't allow every soccer team in the world to have a Lionel Messi on squad. There are a finite supply of most skill sets, which can be well below the total demand for those skill sets.

Companies don't want to hire generic software engineers, there is a huge demand for software engineering skill sets that are extremely difficult to learn, and which can only be acquired after years of diligent effort and experience. This is where the gap is and no amount of paying more will create a larger supply. Some software engineering expertise is already paying in excess of $1,000,000, and there are still severe shortages. That's not a "wage shortage" unless you expect me to believe that software engineers are choosing to grind away on ERP software for $150k instead.

The highest paying software engineering skill sets are the most difficult ones to fill. Software engineering skill sets are not fungible. The highest paid ones are often those with the least elastic supply.

What are these software engineering skillsets that are paying over $1,000,000? There are people who are technically ICs making that, but at that salary its mostly down to politics not technical skill, and they are ICs in name only.

>That's not a "wage shortage" unless you expect me to believe that software engineers are choosing to grind away on ERP software for $150k instead.

There are a ton of software engineers who are grinding away on ERP software are capable of handling the vast majority of work at most SV companies. Many of them can't pass leetcode style interviews despite being capable of doing the actual work. Some of them could but don't think they could. But that leaves plenty of engineers who just don't think the prep time is worth it. Those are the people who you could convince by increasing wages.z

I've worked with many smart software engineers in specialties outside of backend web development/distributed systems who are capable of retraining in a matter of months and would retrain if the pay differential were high enough.

Of course the other option is for companies to come up with more realistic hiring tests, lower the bar and fire quickly, or train employees and set a realistic retention budget.

A skillset I am familiar with that can pay extremely well is the design of complex systems software from first principles. Most software companies don't have this skillset on their payroll, and honestly most companies don't need it. Companies that do need it really need it though, and they can often directly attribute many millions of dollars to the work, so the exceptional compensation seems reasonable, demand greatly outstripping supply notwithstanding.

These days, people with this skillset are commonly employed to design data infrastructure software for cases where open source software has material deficiencies in terms of scalability, performance, or efficiency. Optimization of complex AI/ML processes is another emerging area where I see friends being hired. It typically isn't a leetcode style hiring process.

A software engineer cannot be retrained to do this role in months, the domain is too deep. It requires multiple years of diligent self-study to achieve functional competency, plus many more years for mastery, and I've never seen an exception to this. Much of the essential theory is diverse and never covered in undergrad e.g. greedy routing theory, universal sequence prediction problems, agreement-free consistency, etc. Efficient reduction to practice on real hardware is esoteric, non-obvious, and poorly documented. You need similar levels of low-level knowledge about how silicon behaves as software optimization specialists. In most cases, all existing software is demonstrably unfit for purpose which often implies a likely requirement for inventing a novel and/or unorthodox software design that is superior to what has been done before (the alternative is that all popular implementations that might be fit for purpose are just really poor, in which case you do a blank-sheet implementation of a conventional design).

Consequently, the new talent pipeline is almost entirely sourced from people who devote years of their life to mastery solely because they love the problem space. We simply have no practical way to produce this skillset at a rate sufficient to meet demand. In principle you could build a full-time educational program that focused on training people up effectively but it would still require years to produce more talent like this.

That speciality actually sounds extremely interesting. Is this the kind of thing that companies bring in a consultant for, or are these full time positions?

I’d be very interested to hear more. Do you have a blog, or can you point me to any relevantly communities?

At the margins.

But let’s say there are 100 heart transplant specialists in the US getting paid $1M per year, but you need 200.

You can increase their wages as high as you want and you can’t create 100 of them out of thin air.

I’d suggest for highly technical roles, shortages can exist.

Unskilled or semi-skilled labor is a better fit for your statement.

In the short term, paying more will pull doctors out of retirement/semi-retirement, and delay retirement for existing doctors.

In the medium term, you'll start attracting doctors from related fields who are willing to retrain.

And in the long term, paying more will attract more candidates graduating from med school who will be performing useful work in 5 years or so.

In software engineering, attracting engineers from other less profitable industries wouldn't take nearly as long.

Reply on news.ycombinator.com