I know some people who do this kind of work in various organizations.
As far as I (and they) can tell it all should have been automated down to 1/10 or less the staff it takes now decades ago with computers, but their orgs keep getting basically scammed by vendors promising the moon and delivering a mimeograph of a smeared Xerox of a bad photo of the moon, so it never happens—there's constant tool churn, but nothing ever gets faster or better. Anyone internal smart enough to automate any of it themselves either quietly does it some for them and their buddies so they can slack more, or does none at all and tells no-one what they're capable of because it'll just mean more work and management'll probably fuck it up anyway, if not be angry about it.
It's a leadership failure and it seems to be more common than not. Maybe we'll reach a tipping point where such orgs simply die from that kind of thing, but it really seems like we ought to have by now considering how much "low hanging fruit" that couldn't been tackled in the 80s is still around.
Unfortunately the software is so bad and so complex that upon deployment, they realised they couldn't just roll this out for direct access to staff. It is full of travel-industry terminology nobody understands combined with corporate org policy terminology few understand.
So they designated specific staff as "travel managers" who would be the ones to book travel for their group. These people then get special training etc. In practices however none of the managers have time for this so we are all delegating it to admin staff that already do other admin type work for our teams.
And so the whole exercise has brought us full circle to where dedicated staff are effectively manually booking travel for us. And of course then COVID hit so nobody traveled for 2 years after that and we all prefer remote / zoom as much as possible anyway.
10,000 person tier 1 tech company, 80% of the staff are software engineers, so let’s be conservative and say the average cost of an hour of each employee’s time is $100.
I’ve got a great idea, lets implement a tool that means it takes a minimum of 4 hours to complete the process of booking flights, so that we can better enforce budget policies and make sure people don’t spend an extra $50 on flights.
Let’s also ignore the fact that the “cheaper” flight the tool makes you book is several hours different timing to the one you wanted, meaning you’re now losing nearly 8 fucking hours of productivity…
Yeah… but you saved $50 on the flight…
Yeah… but you saved $50 on the flight…"
How does a cheaper flight lose productivity, I have never been flying or in a plane before.
Then also red-eyes, it's hard to sleep on a plane, so you may get to somewhere in the morning but if you didn't have a full nights rest, what's the point? You're exhausted.
https://www.smithsonianmag.com/air-space-magazine/how-much-w...
It's really not a worthwhile effort getting everyone to learn the domain specifics, and a few domain experts to translate works pretty well, despite looking silly.
The idea "this could all be easily automated so those who don't automate it will go out of business" presumes efficient market theory is true. Markets are not actually efficient. Markets are based on relationships between human beings. And a lot of human beings have a vested interest in things staying exactly the way they are.
that is why i automate some things but in general "keep my head down and my mouth shut". the people who get all angsty and revolutionary about this stuff tend to get fired / laid off / asked to leave / burned out. i know because i used to be one of them.
the people who survive are quiet and don't complain about "leadership failure" as you put it. we do not get paid to point out failures of our leadership. or to do things which would by their very existence imply a failure of leadership. we get paid to do what leadership tells us. our continued supply of health care, shelter, and food depend on it.
if i want to develop my talent / skill, well that is in my free time. i can build robots in my basement and play around with new algorithms and nobody will get mad at me or tell me to stop. now you'll have to excuse me, there is an STM32 board calling my name.
My employers and I were much happier when I switched to 30 hour work weeks and went to grad school part time :P
Imagine an economy where there were no ATMs. Any time you wanted cash, you would have to walk to the bank and wait in line to talk to a bank teller and cash a check. And every week you'd be paid by a check. Paper records were collected for everything and an army of employees had to reconcile the records. No spreadsheet software was available.
In fact, the number of bank tellers (not even bank employees) actually doubled since the advent of ATMS:
> The number of human bank tellers in the United States increased from approximately 300,000 in 1970 to approximately 600,000 in 2010. Counter-intuitively, a contributing factor may be the introduction of automated teller machines. ATMs let a branch operate with fewer tellers, making it cheaper for banks to open more branches. This likely resulted in more tellers being hired to handle non-automated tasks, but further automation and online banking may reverse this increase.
But somehow today more people work at banks than in the past, even accounting for population growth. Why would CRUD apps that make some stuff more efficient change this?
The larger a company gets, the more careerists it attracts. For careerists, “number of people managed” is a key performance indicator (and a boardroom/resume bragging right).
The human motivations of decision makers - to advance their own careers by being the manager that manages more people - is completely misaligned with these future efficiency hypotheses
As long as that remains true, we will continue to see bloat and companies hiring way more people than they need.
How do you go from Sr. Manager to Manager, you manage a large team, and create a need to have your org have multiple managers.
So yes, lots of perverse incentives to grow your org so-as to grow your career.
There are some other tides working in the favor of bloated salaries and payrolls. One is the focus of "stakeholder capitalism". Most of the large companies are held by ETFs. The large banks that run the ETFs are less concerned about individual efficiencies of any one company (they own them all anyway), but are more interested in other things (green initiatives, social causes, etc). So the decisions being made at these companies are less about profit and more about servicing "stakeholders", be it their employees, the community, the environment. That's all fine, but it's difficult to track and hold executives responsible. A bloated staff is fine because you don't want to lay off employees. Or various money losing departments are okay because they're servicing the community.
No org is perfect, but I keep seeing this take again and again and again, yet rarely any form of acknowledgment that it could be good for the company to incentivize proactively growing a team
Perhaps that's not quite the right comparison - even though the population hasnt doubled since 1970, the number of people employed has (80M->160M, approx [0]). Adjusting for that, Bank Teller employment level is basically unchanged.
I think it is hubris to say that anything a human can do cannot eventually be replaced by a machine. When we will actually reach that point is the real question.
Fednow should alleviate some of this.
The number of tellers is expected to drop by 12% over the next decade while general jobs are expected to grow by 5% according to government forecasts btw.
Being a cashier used to be a tougher job because you needed to know all the inventory. Now it’s ok for a grocery store cashier to not know what parsley is.
If automation makes it practical to do an embarrassingly small task in order to provide a value to someone richer, yeah, you can be hired into it. Maybe a lot of people will. But that’s not prosperity that’s capitalism holding you by the balls.
Not everyone is equally capable, which implies that some people are on the left side of the distribution. Some jobs that simplify the essential actions bring the job to meet the capabilities of the worker.
Lowering the bar is not needed or helpful. People who are so incapable that these jobs are their peak potential should probably just be on social programs.
(I had a family member, since passed, who was significantly mentally handicapped by a difficult birth sequence. She loved nothing more than to feel like she was included and being helpful to the limits of her ability.)
Jordan Peterson - What Kind Of Job Fits Your IQ
Population growth doesn't account for regulatory changes and it takes a while to automate anything that interfaces with the law because of the inherent risk in getting it wrong. Banks are acutely exposed to this because they're often the ones who are ultimately responsible - they're the default insurance policy against counterparty risk.
Edit: Tellers also tend to be general purpose account managers at smaller banks, so if you called up the bank you'd eventually get forwarded to someone on the floor at your local branch instead of a call center.
Additionally, people are much more reliant on financial services and payment rails. Credit cards weren't as common (and required for almost everything) 20 years ago. People used cash for probably 10X as many transactions as they do now, so you probably had more 'cash economies'
All those branches closed during covid. I have 6 within walking distance and there isn’t a human anywhere.
I count half a dozen Chase branches that have gone up near me in the past few months. I don't get it.
I’m not sure I buy into the bank analogy (probably due more to technology making bank products cheaper/more accessible and thus more customers and thus more bankers) but I definitely buy this.
If you believe fundamentally in basic market dynamics and that humans are generally productive creatures, then any automation that comes about will simply shift where the human labor is most useful, and we will generally always have enough human powered work to go around. The tech and automation is just making it more efficient for everyone.
If I could get quarters form an ATM, there would be no reason ever for me to.
What are people using tellers for?
They need to look you in the eyes to see if you are pure of heart and deserve a mortage.
Why am I stuck with WF? A longer, boring-er story.
It hasn't in 250 years of automation.
Regardless time will tell
You can see this principle with ERP implementations - they can literally bankrupt an entire company if you get too ambitious with what can be automated.
Edge cases & changing business context aren't going away.
Psycologically for some it is a massive change - having had numerous colleagues in the past who had core values like "if you are not physically sat at your desk you are not working".
On every metric, excluding commuter congestion and possibly office rental rates - no change.
In my experience, it's the exact opposite. WFH creates more of a fog of war, so companies end up hiring more managers to stay on top of it.
Some employees are great at self-directed WFH, but many others struggle without the physical, social context of the office to get them focused. Again, companies rely on more managers to compensate for those employees who struggle more with WFH.
I know people will say "Just hire people who are good at WFH", but if we transition large amounts of the workforce to WFH then you can no longer pick-and-choose just the good WFH people. You have to deal with the realities of managing WFH at scale.
Senior ICs are split on WFH. Managers who can code are split.
Managers who can’t code are unanimous.
It is a real issue though and the answer isn’t obvious. I tend to think that something like A16Z’s model will wind up as the end state: modest amounts of flex office space, geographically distributed, with frequent, family-friendly off sites.
Pair-programming software infrastructure will have to improve, and maintaining a hub of urban office space where Senior ICs get a COL bump to be near the young folks will probably be part of it.
It’s a new world, and there will be an adjustment, but there’s enough critical people who simply aren’t going back that it’ll be management’s job to figure it out. I imagine figuring it out well will pay commensurately well.
> but there’s enough critical people who simply aren’t going back that it’ll be management’s job to figure it out
Pendulum is swinging, my friend. With a big recession looming, companies will have all the leverage. Elon has already paved the way and other big companies will soon follow. On the other end of the spectrum, a lot of startups will be bootstrapped in this recession and a majority of startup founders want to have the initial bootstrapping team to be together[0].
[0] Based on my first hand observation working as an advisor for a VC firm.
With that said, we've been through this before: companies always have leverage with many/most of their employees, and that is going even further up for awhile, but Google Search doesn't stay up without a meaningful number of people who have made enough money to not be pushed around.
This is the flip side of being one of the "CEO/VC class" and having a boom decade or two: most of the real pros made some serious money on stock and can't be herded like cattle the way new grads can be.
The American economy has become an experiment in how close you can get to indentured servitude for your working people without calling it that by twisting the macroeconomic knobs until the kid doesn't get medicine unless you hump that bullshit job. I'm sure the money guy who figures out how to do that to hardass infrastructure hackers or founder-caliber kids will be hailed as a hero (in private), but it hasn't happened yet alhamdulillah.
If you try doing that remotely you need permanent surveilance with a webcam, and people get all uppity about it when that happens in their own home.
Jim sits and watches cat videos on youtube all day - is 0x plan, and will be promoted to being your boss next year.
The tactic never seems to be to hire more people to do the actual work though, curious!
It's the most ridiculous thing I've ever seen, so it must either be a legal requirement or maybe the organization had a really bad experience with expense fraud and is now paranoid.
“You pay me enough that I can afford not to deal with Concur or our expensing process.”
This is all baloney. There is noone organising the information and documents in "modern" companies. There is just a rolling reset of knowledge with an average lifetime of about three years.
> Having physical pieces of paper as a fungible business tool is still something used by a lot of firms but they are simply behind the times,
They are not behind their times. They are using a system that actually works.
Nah, I’ve seen legacy code and how much maintenance it requires. Regular devs are safe.
This has been said over and over. In the 1980's people worried automation would leave lots of people unemployed. And long before in 1931 Keynes [1] predicted we should now be living in a world of abundance. He said our the biggest concern would be how to spend all this free time.
Yet I see people working harder than ever, and burnout is the number one reason for absence of work in many countries.
Every step you automate opens up new possibilities and more employees are needed. I work as a software engineer now for around 15 years. Never have my software reduced costs, because as soon as my software automates something people want more.
Although I'm skeptical, we could really facing a turning point, Keynes made his predictions for 2030, so we'll see :)
[1] Economic Possibilities for Our Grandchildren (1931)
I do have a pet-theory that humans aren't well equipped for long-term 9-5 computer thought work. Anecdotally, I see people burning out in < 5 years with thought work careers. People under 30 packing up their careers to go live in a van or whatever it may be.
There are very few industries that haven't already digitized, the ones you mentioned that have human CRUD apps are relatively rare. Accounting, law, healthcare...who else? Basically nobody.
How many new companies exist now because of automation?
Sure, there are some prestige managers out there, but I think their prevalence is overstated. Most managers I know in tech companies are overworked and have no time on their schedule.
Increasing automation...the increasingly widespread adoption of enterprise software, data analytics, and machine learnings are increasingly rendering office jobs moot as much of the traditional number crunching, document processing, and repetitive analysis performed by office professionals is now being performed by software.
This is not even wrong, in the W. Pauli sense. Just one of the many reasons is enterprise software works less and less for the enterprise that 'rents' it and more and more for the enterprise that controls it.
Everyone hates enterprise software who is on the receiving end of it.
As for the accounting example DM5 uses, ask his partner about when three people have to co-ordinate a fixing a single ppv in a single po. It can take days to fix a single mistake. Meahwhile dozens or hundreds more po's have piled up.
po = purchase order, ppv = purchase price variance 3 people = seperation of duties, because people try to steal.
At this stage I'd welcome any automation in the accounting space but efforts I've seen thus far have not been promising. Domain specific software is tricky - either you get SWE in then they don't know the domain, or the opposite then the code is garbage. That gap means endless meetings and committees about requirements and end user testing in practice even for the most trivial of automations.
Outsourcing is way more likely to kill accounting jobs.
It also removes geographic limiters that drive up the cost of certain pools of labour.
Until some genius writes delete from invoices ...
I don't think that we are nearly close enough to have digital recordkeeping done right.
WFH will reduce middle management and "prestige" employees as more efficient workflows evolve from the old "filing cabinets and typewriters" methods that some companies are still using to this day.
Automation is coming, and it is definitely coming for the white collar workers who act as human CRUD apps.
The ability to collate, search, and analyze data used to require labor power, it does not anymore.
I watch my partner's experience in accounting and the difference are stark.
There are many many bs steps in accounting with paper that disappear when digital tools are used.
Having physical pieces of paper as a fungible business tool is still something used by a lot of firms but they are simply behind the times, there are relatively simple software solutions for most paperwork needs and this reduces labor power required.
"We've always done it this way" is an excuse for firms that will fall behind other digitally empowered firms.
What does this mean for investors? Don't invest in companies that insist on using filing cabinets.