Top
Best
New

Posted by tosh 8/30/2025

You Have to Feel It(mitchellh.com)
381 points | 156 commentspage 4
sublinear 8/31/2025|
> When you feel it, you know. The feature makes you smile when you use it. It fits right in, like it was always meant to be there. You want to use it again. You want to tell people about it. This is the difference. This is what metrics, specifications, and demos miss.

Totally absurd! The metrics and specifications are what make all of that possible.

This feels like it was written for execs and managers who bury their heads in the sand when they're overwhelmed.

bongodongobob 8/30/2025||
As sure as fuck don't. Just keep the paychecks coming.
bubblebeard 8/31/2025||
Whenever my logic said one thing and my gut feeling told me different, I used to ignore my gut. This would more often than not cause some sort of backlash I could not possibly have foreseen. Sometimes immediately, sometimes several days later.

A year or so ago I got tired of this, started doing the opposite, and I will never go back.

Feelings overall are underrated.

hendry 8/31/2025||
Vibe coding
bonoboTP 8/31/2025||
Yes, absolutely true. And this fact will be used against you.

Two sides of the same coin. Humans have an internal need to see that their work output is valuable and they are useful to society/community. This is not some unrealistic expectation. For most of history, it was fairly easy to see how one's work directly results in value. Farming, butchering animals, making clothes and shoes, paving roads or building houses or whatever it may be. People could feel satisfaction of jobs well done, applying their experience and being valued for it. It is normal. Pointless drudgery eats away at one's soul and leads to burnout, much more than simply being "overworked". If much of your work-related mental energy goes towards timing when to jump ship before it sinks, that won't be sustainable. Working on some pointless project that you know will get axed, but for some internal politics reason it has to remain up and running for a few more months to get someone promoted before killing it, etc.

When people say that a job is a job, and you wouldn't get paid for it if it was enjoyable, it's just not right. If you produce value, you do deserve the pay, it's not supposed to be compensation for being miserable.

On the flipside, when MBAs internalize this message, what they take away is that "passion" and "motivation" makes employees more productive and prone to put in extra effort. So they come up with the genius idea to test whether you are "passionate" before hiring you. Resulting in a dreaded and ugly song and dance ritual where you have to exaggerate your passion and talk in unrealistic language, and grifters and smoothtalkers win in that.

That's not the kind of passion that is needed, the outwardly, enthusiastic-glowing passion. Just something where you feel that it's neat, it's a net positive for the world even if small, and you left the world a bit better when you go to sleep at night compared to when you woke up in the morning. Even extra money can't compensate for that in the long term (short term yes).

fuckaj 8/30/2025||
But I feel depressed every time I use Terraform. I see why it is useful, good etc. But I just enjoy it least of all the programming activities.

Pulumi isn't much better. I feel IaC done that way isn't the way we will settle on long term.

losvedir 8/30/2025||
To connect the dots, Mitchell is the creator of terraform. And honestly, I kind of feel the same. I think it points out that even if you "feel" it as the creator, that doesn't mean others will feel the same way, necessarily.
azriel91 8/30/2025|||
Does my side project provide any hope?

https://peace.mk/blog/checkpoint/

(old blog post, but I'm slow in making progress)

chanux 8/31/2025||||
FWIW: https://x.com/mitchellh/status/1938066378185183279

With Terraform I have felt I'm fighting it at times but I also understand the reasons.

ricardobeat 8/30/2025|||
They left the role of CTO many years ago, then became an IC again, and then left, safe to assume he wasn’t feeling it either.
invertedohm 8/30/2025|||
I get the feeing you're looking at this through the specific lens of a programmer. Terraform isn't made for programmers - you'll miss all the flexibility a real language gives you. It's made for ops people who deal with wrangling a whole bunch of different types of systems with different API's and languages and just need some way to standardize the management of disparate systems, whatever counts as "infrastructure".

The state file thing gets a relatively large part of the hate but it's that and the limitations of the DSL that make the DAG possible and useful. Pulumi and all the other wrappers don't solve this, though they can plausibly solve the "closer to programming" problem and I'm sure that has a valid audience.

I guess what I'm saying is, I think it'll stick around and we will in fact settle on it for a large part of operational work. I'll add that I also think k8s should die a quiet death and _that_ will be seen in retrospect as a necessary step to something better.

elcapitan 8/31/2025|||
Is there anybody who enjoys any form of configuration management? It feels like the most joyless of all technical activities, like filling a tax form.
theteapot 8/30/2025|||
Please try some Cloudformation then reflect and reevaluate.
danw1979 8/30/2025|||
Not a popular opinion I know, but I really like writing Terraform in HCL2. Like, if I’d invented this, it’d give me the feels and I’d want to share it.
chanux 8/31/2025||
When you eventually understand the constraints and learn to work with it. A zen moment :D
sethammons 8/31/2025||
I quit a $400k+ job to get away from IaC. I loath yaml. More and more, my day was filled with "you have an error on line 1, good luck." It was more k8s than tf, but I get the same sneer on my face when messing with tf.

I refuse to let such a shitty experience be what defines my day.

I was hoping pulumi would help. Haven't used it yet, but it is sad to hear it doesn't live up to the hype.

fuckaj 8/31/2025||
[dead]
judithh 9/1/2025||
[dead]
tomdickandharry 8/31/2025|
[dead]