And your starter “production” deployment of the Nomad/Consul/Vault stack is literally 12 VMs, comprising three independent Raft clusters. There is no decent way to do zero-downtime instance replacement without building your own orchestration layer, but also they’ve had a years-long track record of shipping bad upgrades and following up with only manual remediations or workarounds instead of a fix.
As someone who has productionized and maintained truly hundreds of those clusters across several jobs, it is hard at this point for me to recommend Consul, Nomad, or Vault to anyone serious about building reliable applications. Too many broken upgrades and manual click-ops tasks just to keep them online. (…and I’ve said nothing of the actual product!)
Personally I find LLMs absolutely terrible at writing Terraform and full of hallucinations. But also Terraform is my bread and butter and our use/workflow is fairly advanced. And we're multi-cloud + baremetal. That wasn't an area where I was going to get a ton of value out of LLMs anyway.
I've definitely found what you're describing at bigger companies, but I also previously had experience writing software at smaller non-technology companies.
Legal marketing specifically. Weirdly, my work had more impact, respect and longevity there than the place where I'm a much more senior engineer supposedly directing the work of a whole organization of engineers. I had it better where I was a 1 of 2 than a leader among hundreds.
Small successful companies are great, but the hardest thing for me psychologically has been when I'm at a small company that is struggling to convince anyone to use its product. Being a small cog in a giant machine serving lots of users is more satisfying (to me) than building things that nobody is using.
Not just isolating them from customers but also from engineering-adjacent work that isn't code.
I've been at a place that is basically microservices slop (several dozen services per engineer). They're all poorly maintained and at least a solid 40% of all this code that they've written could have been just a traefik or nginx configuration/container.
When you have a lot of inexperienced (relative to industry) and overworked software engineers, the solution to every problem becomes to write code and writing new code should be a last resort.
Worse still, there's just a poor general understanding of the internet protocols they're working with and of how to do distributed systems right. Unfortunately with LLMs I've been seeing this get worse, not better.
They use the LLMs for code generation but not architecture review. Bad ideas are getting fully-baked quickly before anyone with good sense can intervene.
I find the stream rips to be really shitty quality.. The original source is very low bitrate, compressed tae fuck. I find for stream rips from netflix for example I need to download a 4k rip in order to watch in 1080, and that's acceptable.
I don't recall exactly when it went onto streaming, but I'm pretty sure I got a good quality version before that. It may have been released for streaming in other regions earlier than I thought though, I don't keep super up to date with that sort of thing, as I generally don't watch movies super soon after they're released
It's not super hard. You just have to listen to when people are asking for things, try to help and read an org chart.
90% of the engineers I've worked with in bigger companies wouldn't know how to find someone in the company outside of their direct reporting structure.
Honestly it's pathetic. The rest of the organization can't work like that and these are table stakes social skills IMO.
I seriously think the "headphones on, get into flow" trope is the most damaging meme in our industry. Management also takes huge advantage of the low-information environment that engineers seem comfortable in. Most of them don't even (really) know what our product is or how it's sold and marketed.
For most people it's hard, especially for the stereotypical "IT nerds".
I think the best tip for people who have a hard time is: Watch who of your colleagues know "everyone" and spend as much time as possible with them. If they ask you to go for lunch together, always join. If you can work on a project with them, do it. They will casually introduce you to all the people over time, and might just tell you the newest company gossip.
I'm closing out 3 decades+ around the "IT nerds" and if anything they spend _more_ time socializing than "normies". The difference is typically that their socializing tends to specialize around fandoms/activities instead of generalist/community things like civics, sports and family.
Not to be overly harsh but the problem is with seeing those who don't share their interests as people equally worthy of attention. IT nerds typically have trouble meeting people where they're at instead of the other way around. And most of the time it's because they've never made any effort to do so.
That's why I say it's not super hard.
It's also becoming more of a societal problem in general as younger people spend more and more time isolated and socializing in bubbles. I think it's a serious and growing problem that people don't have friendships outside of their immediate peer group age-wise.
My brother's kid is in her early 20s and her and most of her friends don't see people at 30+ as people. They don't value their opinions and it has all sorts of negative effects on their lives like they struggle to obtain/keep jobs, etc. That's not a blanket generalization though -- we have some team members in that age-range and they're great.
We ended our intern program and so did a ton of other companies.
IMO one of the worst decisions we've ever made because 80% of the time the interns we take on and then hire are absolute superstars.
And even before we ended it we had a couple of years where we stopped competing for talent from Waterloo. I guess Trump made that harder but yeah bad decisions all around.
I worked at a top 5 hedge fund in the early 2000s. They had a large team of E&Y auditors onsite at all times that I worked somewhat closely with.
Some things stuck out at me:
- They were all in their early 20s.
- They were all incredibly checked out. Honestly they still seem like an outlier to me decades later.
- They partied hard. Yes, with drugs.
- Most of them were in rotating intimate relationships with each other and unusually open about it. Office scuttlebutt was literally "who is fucking who this week".
- They seemed busy for maybe two or three weeks out of the entire year and then it was long stretches of Minesweeper/Solitaire.
I filed this away in my head as "provides no value" and that was decades ago. If the industry itself is worse off today I can't imagine how much worse it actually is from my experience.
It's insurance. The same way you don't want to go too far with germline editing. There are genetic variations that you might need to save your species some day. The cost of that is people suffer.
We may need a close connection to that way of life again and not have to relearn it from scratch.
This strikes me more as one of those things that is shocking to hear but not incorrect. People get more upset that someone said it without actually having a counterargument.
He's not saying that we _should_ live that way but that we might need to.
I think you're underestimating how much we simply don't know about how to live that primitively. Maybe you could do it in small groups of people. Maybe up to a dozen. Most experiments at this fail hilariously early by the way.
Can you do it and sustain hundreds of people? I doubt it.
At least they're here to be potentially observed. You don't have to _totally_ wing it. People living like that through history had bigger day to day survival concerns than documenting the finer details of sustaining their continued existence to us.
The last closest analog we have to them would be the Hadza people and they've had agriculture since 500 CE...
The problem is that when you run this long enough you want K8s features anyway.
reply