Hacker News Reader

Comment by woodruffw | Hacker News Reader

> The Pareto principle is “80% of consequences come from 20% of causes”. And as an extension of this, 80% of results are delivered by 20% of engineers.

As an extension of this, I propose we remove 4/5ths of each airliner. The 20% that remains, so long as we pick the right 20%, will surely fly correctly.

Seriously: who writes this dreck? You don't have to believe that every single employee at a company is productive or necessary to understand that you can't saddle your remaining employees, even if you picked the "right" ones, with the remaining workload.

View on HN

>You don't have to believe that every single employee at a company is productive or necessary to understand that you can't saddle your remaining employees, even if you picked the "right" ones, with the remaining workload.

This is exactly the problem with this thinking. A lot of people don't contribute as much as the top contributors, but if you remove them, suddenly the most productive people have to do their normal jobs plus all these other peoples' too. Then they're going to feel overworked and look for a better job elsewhere.

Sometimes the most productive people are productive precisely because they can pick tasks that play their strengths and manage to ignore the rest. If they're suddenly forced to take on whatever shit work pops up, some of them aren't going to be as effective anymore.

This is a very good point. An overworked, under resourced team is running just to stand still. There's no time to take a step back, consider what might be done differently. It just has to be done right now. It's a recipe for stagnation.

Another problem is that the better employees can move to better jobs more easily. So when management becomes a problem, they'll move elsewhere, and that will inspire more employees to leave.

You end up with the less productive employees remaining.

Once you've fired the first 80%, 80% of the work will get done by 20% of the remaining 20%, so it's safe to fire another 80%, and so forth.

Ultimately, each company should employ exactly one engineer.

And they should only work 20% of the week.

Scott Adams knew.

> Prediction 36: In the future, all work will be outsourced, until all the work on the planet is being done by one guy.

Indeed. And let's say you made the cut, you're staying, phew. What about next year, when there's the next cut? And the year after that? You've every incentive to hire _bad_ people, to protect your good ones. It's in your interests that other teams fail so that yours can survive.

This right here is why I always half-assed when my company asked me to participate in interviewing devs. Given the comapanie's track record of laying off quality devs for cheaper options, they were all potential replacements for ME. The expectations around loyalty in this industry bewilder me.

The management class does not want this Twitter ordeal (likely a huge public failure) making engineers think their skills give them power over companies.

I agree completely.

On top of that, it feels like classic "magical" thinking by a manager who doesn't understand technical scope or split responsibilities. Why bother to understand why you've hired so many people before blithely opining that 4/5ths of them are fireable? Who the hell wants to work for someone who says things like that? Certainly not the "high-quality" engineers who are supposedly being separated from the chaff.

How do you define "public failure"? My definition would be entirely contingent upon the intentions of the new owner since it's his to do with what he pleases. If said owner is well into middle age and bored, but has more money than literally any other person...? Who am I to judge what he decides to do with it?

"It's just money" says Elon and if his goal were to be the world's most massive troll (which I suspect is a large part of it), then it's achievement accomplished, full stop.

At least to my mind, "public failure" includes but isn't limited to embarrassing outages and service degradation on one of the world's most popular social media sites, one that was more or less running smoothly immediately prior.

I notice every VC cheering this absurd management style on. If Twitter's new management can be somehow spun as a success, it will normalise treating tech employees in general and engineering in particular very poorly - longer hours, less compensation. For VCs and senior management, that's a fantastic outcome.

In a month they're going to say "look! we didn't have a single outage during the World Cup" and act like it vindicates their approach.

The irony here is that you could easily fire 80% of the C-suite and VPs and a company should survive for some time. This is kind of a tautology because leadership at bigger companies should be mostly operating at the months-to-years timeframe, so it should take a long time (or a unique emergency) for their absence to be felt. We could run both experiments (firing engineers vs firing execs) against each other across a sample of companies and I know which I'd bet on to still be mostly functional 3-4 weeks later.

I think the assumption is that 80% of the workload can safely be dropped on the floor, that the resulting 80% of results is sufficient to survive. However goofy it is to apply that to Elon's haphazard cuts to SRE and administration etc, the premise is not crazy. There's certainly no suggestion of multiplying the workload of the remaining employees by 5.

Losing 80% of employees is not the same as losing 80% of the workload.

Much of that workload isn't going away, even if they are ignoring it.

I'm inclined to agree for the non-eng parts of "the workload". But swe workload in a company that already pretty much does its job is a very flexible concept, and I'm not sure it has a firm existence apart from the workforce, firm enough to say there's something there that "isn't going away". And remember the standard for "the workload" that must be done is "survival", which is not a high bar.

I guess where I'm going is that the article states a roughly true thesis, but applies it in a situation where it's not terribly relevant.

>>As an extension of this, I propose we remove 4/5ths of each airliner. The 20% that remains, so long as we pick the right 20%, will surely fly correctly.

Not sure if they will fly correctly. But most airlines I'm sure have wildly profitable routes, some moderately profitable routes, some mild. And some I'm sure are loss making routes.

You start plotting a graph and a bell curve emerges. The key is this- Do the profitable routes have more demand you can't meet now, but can meet if you divert resources from the loss making routes. If the answer is yes, you can always eliminate the loss making routes, and use those resources to supply to profit making ones. This makes perfect sense, because it not only saves money, this makes money.

There's a bakery near my place which used to sell its things in a shop. Soon enough they realise finding talent to bake is a big problem and most people were buying bulk quantities to sell it to others for a small profit. Eventually he realised he could sell to people who could run shops but had no talent to bake. That way we was able to divert his time and resources to make things at scale, so obviously his profits and by definition market capture increased too.

The real question in case of companies like Twitter is- What is Twitter's core competence? Should Twitter run a in house payroll team? Should they be running a PR team? What sort and size of Legal teams should they be running? Should they be running a in house content moderation team? If their business benefits more making newer products and features. It does make total sense to cut down resources from places are a cost center -ve to hiring more programmers. And it does make total sense for Musk to say things like they will be fairly intense on the engineering and programming side of things.

"Airliner" here means the airplane, not the airline.

(And there's a reason airlines fly routes at a loss: an unprofitable route brings travelers to hubs where profitable flights can be allocated cheaply. It's never, ever as simple as "just trim the fat," because the things you think are "fat" almost never are.)

Isn't that exactly what most people did with the economy class seating. They modified their plane to transit more people. So this logic still holds. The remainder is as simple us understanding what basic structure of a product holds good as a product to sell.

>>It's never, ever as simple as "just trim the fat," because the things you think are "fat" almost never are.

Fair enough. But do trim the real fat.

I can't tell if you're intentionally being obtuse or not. The point of the original comment was that "Pareto-style" reasoning here is ridiculous: the engines and wings do 80% of the "work" in an airplane, but you can't remove the rest and expect it to still fly. You can't really quantize work inside of complex systems (like airplanes, or corporations) in this way; adjacent sub-threads in this thread explain why.

We’re not talking about airlines. We’re talking about a place you go to see the Iron Sheik f-bomb someone.

The point is that it's bunk reasoning, the same kind that results in managers proudly proclaiming to have used two women to make a baby in only 4.5 months.

The Pareto Principle in no way implies that you can just fire 80% of your employees and retain 80% of their "effect"; that's just not how tasking and scoping work.

Reply on news.ycombinator.com