Work Somewhere Dysfunctional

People who want to study earthquakes don't get paid by the earthquakes.

Author works for a defense contractor.

I'd need a ton of evidence that this particular defense contractor has the goal of doing something inherently safe with AI before I'd read the author's analogy as anything other than the written equivalent of getting drunk at the local bar to cope with a dysfunctional marriage.

That said, I’m grateful for people who are really into the org stuff. I’ve worked for a savant at that stuff and it really did improve things. Thank you all who suffer the politics on behalf of the rest of us.

Part of what makes a place dysfunctional is that the rewards have no correlation with performance or impact.

Then there is stuff like this:

> People who want to study earthquakes need to go where there are fault lines. People who want to solve hard problems need to go where those problems live.

Sure, if you want to learn about dysfunctional places, go into them. Learning about them isn't a priority for most people, and if you have a technical profession (not administrative), it would be a hobby at most. So, do you enjoy it there?

In a truly dysfunctional company, you'll burn out and/or become a pariah if you try to tackle foundational, intractable problems. I've worked places with 'emporor's new clothes' level disconnections with reality. You seeing the problems others don't see is irrelevant.

Focus on things you can improve and small changes you can effect, and build up a library of answers to behavioral interview questions. Accept that your efforts will be undone and be focused on your long term goals, which do not include dying on this hill.

> On the other hand, people who have pragmatic and slightly selfish goals in addition to wanting the do good are more resilient. If the system change they envision doesn’t work out, they still have something to show for all their efforts. That keeps them grounded and calm for much longer in the same environment.

I disagree wholeheartedly. Many companies (from big tech to fast growing startups) have hard problems that need solving, and they needn't be dysfunctional. In fact some high performing organisations have plenty of hard problems, as solving hard problems usually yield even harder problems to solve.

Sadly disagreeing with the premise makes the entire article a weak claim.

I'd definitely agree that you'll build the most skills at a dysfunctional place. The type of skills you learn are typically the opposite of what makes an established organization work though. There's a common saying of "work the people, then the problem". The challenge here is that when majority of your work is to improve the culture, the actual skills you were hired for start to slowly take the back seat.

I don't agree that being selfish is going to help you at all. In fact, being selfish is the exact reason why most organizations are dysfunctional. Here's one that comes off the top of my head. Say you're a line engineering manager and you're incentivized to deliver a large feature that year by your manager who is largely clueless of what your group is doing. But, the whole team realizes that the feature you're incentivized to deliver is not in the best interests of the team right now. What do you do with your role power? Most people will selfishly get that feature implemented to get their max bonus & rewards that year. That type of behavior breeds more dysfunction.

Here's a whole list of dysfunction: output over outcomes, obsession with internal metrics, lack of customer research, optimizing everything for little gain, shipping features like an assembly line, over-dependence on data/spreadsheets, fast paced twist and turns of work, over-engineering, trying to keep everyone happy, flip-flop decision making from people with role power.

This seems to conflate dysfunctional markets with dysfunctional workplaces. It seems like the author's thesis rests exactly on this logical error.

You may have started at a wonderful company that is acquired or for some other reason becomes dysfunctional. The place isn't dangerous, and the pay covers the bills, but you are losing market value (skills are dated, less relevant). Changing jobs is a pain, it could be bad timing for some reason, and perhaps you lack the experience or training (certs) you want at the next place. You may be jumping out of one pot into another, or worse, into the fire itself. In this case, this approach makes sense to me.

However, I have to agree with the majority of the comments, avoid getting into one in the first place. Don't seek them out, and if you arrive and it's bad, get out quick.

Software Safety and especially AI Safety are hard technical problems. You're not going to figure out how to add a few 9s to the reliability of a computer vision system by studying org theory.

Making progress on The Hardest Problems requires hiring the excellent ML engineers and scientists, and then having even half-decent management. Management does matter, but the strong IC talent is a precondition to progress.

If Defense had armies of competent ICs but was still failing, I guess focusing on non-compensation-related organization issues might be reasonable. That's not even remotely the case.

Just shy of 70% of CS PhDs from US institutions are foreign nationals, which means Defense is already talent-constrained. The 50% or so who are qualified to work on AI safety and qualify for relevant clearances would probably want a close to a 0 added to what Defense offers (and not even for moral reasons... security clearances and "must work at the office" are real drags on quality of life that require significant additional compensation).

I've seen first-hand that defense systems are insecure and unsafe because defense chooses not to purchase excellent IC talent. The biggest organizational challenge in Defense is the lack of adequate compensation for engineers and scientists.

The largest management problem in the Defense industry is figuring out how to get the org to pay for excellent engineering and scientific talent. Your competitors in the labor market are paying high six or low seven figures, don't require security clearances, don't require drug tests, and have much more hybrid/WFH flexibility. For me to take a job in defense to work on ML Safety, you'd probably have to pay north of $1M.

The risk of a dysfunctional workplace is the negative feelings that start consuming you, crawling into your personal life, impacting you emotionally, and causing burn out.

Don’t work somewhere dysfunctional.

One of my jobs had the dysfunction cranked to 11 with politics (government and inter-office). I can honestly say that screwed me up for a good long while. It took two job switches to be normal again. You don't realize how far down you are until an outside observer points out you haven't smiled in months.

The more obvious each casting choice is, the more dysfunctional the workplace.

I might have a different definition of a dysfunctional place though - for me it's a place that works against me and my productivity, a place where I have to go through a number of hoops in order to do the job I was hired to do. And let me tell you, by the time I made it through all these hoops I was so drained of energy I lost all the passion and will to continue

As a side note, the Chinese words for "crisis" (危机) and "opportunity" (机会) share a common character (机). A useful mnemonic (crisis can end in an opportunity, opportunity can begin in a crisis).

I worked at such place that boasted security on the outside. From within, it was spit and duct tape. I couldn't believe it was a fortune 10. After couple of weeks working on one of the smaller projects, I created a new branch and called it rewrite.

When I had free time, I would rewrite the entire project from scratch. I really hated all the in line mysql connections in side_bar_bottom_new_v2.php. It took me a month to realize that I couldn't rewrite an application that was built over the course of 10 years, in my spare time. Only when I embraced it and took the time to understand how it was built, was I able to turn it into a secure app (messy, but still secure).

Now when I meet dysfunctional systems, I don't get mad. I challenge myself to figuring out what the heck it is doing.

I mean seriously - why take shit when there's a 90% chance you're working on another elaborate CMS/CRUD-looking system? There's millions of them out there, you don't need to subject yourself to terrible experiences. Just go someplace else.

Don't waste your life on shitty organizations. They're not worth participating in. You're not growing useful skills.

source: currently wasting my time in a shitty organization.

Many years ago, I found the character of the "The Wolf" in pulp fiction fascinating: a person who basically does normal, regular things in dysfunctional environments and charges a premium for that. I love when people call me, when the project is overdue, when nothing seems to work, when they ran out of options. Strangely, I seem to thrive in these settings.

Spent a decade in big tech surrounded by dysfunction, pushing truly innovative tech through the muck to release and expending significant effort to clean the muck as I went for myself and others. Now I'm just burnt out, jaded, and trying to exit tech altogether.

At the end of the day if a direct line cannot be drawn between a problem (or its solution) and revenue, then management doesn't care and are not capable of filling in that line in their minds. And if you fill it in for them, you'll get at most a pat on the back.

There is no need for those places to be dysfunctional though. You can go study earthquakes in California or Afghanistan. You aren't any better for it dodging the Taliban though.

>bellmar.medium.com

I did a double-take to ensure it wasn't "ballmer.medium.com" :D

That

The secret to maintaining a vital mental landscape is to resist the temptation to think the whole world conforms to what you've seen so far in your short-ass life. Be content to "know" a lot less, and probably say less as a result. Ice Cube once rapped something about "shut my M.F. mouth if I don't know" and that seems like a pretty good policy, though if everyone did it, there would be a lot fewer blog posts in the world, unfortunately (or fortunately?)