IT/Software career thread: Invert binary trees for dollars.

TJT

Mr. Poopybutthole
<Gold Donor>
46,819
128,188
A guy that works for me, if he finishes at 4:59 with a task he'll ask me what else he can do. He closes at least 50% of our tickets. Nothing wrong with it but sometimes I want to tell him to touch some grass
I am firmly capable of talking to people. I also spent years making my own projects at my old job because I found deficiencies and corrected them. Which is how I ended up being the de facto head of product telemetry despite not even working on the product itself. As well as engineering the entire billing system end to end. Nobody tasked me with that, I just saw something that nobody was identifying as an issue and did something about it. Nonetheless every dollar of revenue we ever received since like 2020 comes solely through the billing system I created.

If you give me two tickets that I quickly identify as being simple for me to do. I will do one, then do the other one. Then I will usually chill and my natural curiosity fills in the blanks. Which is why I have a proposal to make their RBAC more coherent and automated already, and another process change to bring uniformity to some of their data management. Thankfully my new manager started the same day as me and we get along great and agree with the issues this company clearly has.

This company has a much bigger data org than my last one but they are way behind when it comes to the platform itself. So it's really my time to shine.
 
  • 1Like
Reactions: 1 user

Noodleface

A Mod Real Quick
39,549
18,130
Just saying as a guy who gives out the tickets, I appreciate the guy that works like that but he's also a mental burden pretty much all day. Not saying that's you, just my experience.
 

Phazael

Confirmed Beta Shitlord, Fat Bastard
<Gold Donor>
16,173
36,383
Good work output never comes without a toll. Some people are social cancer, others just straight up assholes (me), and others sketchy/untrustworthy. Learning to cope with that to get shit done is pretty much the way you get to be successful as a leader, especially in IT. Well adjusted people are all path of least resistance lazy types who lack the OCD or mental damage to power through work. I think we can credibly say just about everyone in this thread has at least a little mental damage, for an anecdotal example. It also makes work more interesting, in the Chinese sense of the word.
 
  • 1Solidarity
  • 1Cringe
Reactions: 1 users

Phazael

Confirmed Beta Shitlord, Fat Bastard
<Gold Donor>
16,173
36,383
Guess it's not just us coding nerds


But hey, jobs reports keep coming back aces.

I dub this phenomena Schrodinger's Poo, which basically means when it makes things look good Poo numbers are included and when it does not, the reverse happens. Its massaging statistics but done in a way to make things look more rosey and obfuscate how much horse shit with H1B Poo infestation is going on. The gov and corporate america do this shit a lot.
 

M Power

Silver Squire
225
202
Well my interview today with a tech adjacent company seemed to have went well. It does seem like a lot of work to build out and develop the project with a tight timeline but if they are willing to pay me what I will ask then I think it's worth it. The one huge positive is the emphasis on growth opportunities for the company and within the company.
 
  • 3Like
Reactions: 2 users

TJT

Mr. Poopybutthole
<Gold Donor>
46,819
128,188
Wrapping up week two. I'm very glad that my manager started the same day I did. The overall org has heavily favored internal promotion. Which I am all about. This has resulted in many of their developers being fully self taught within the org and formerly salesman, analysts, client success, and others starting as devs and working their way up. Things that make this apparent in their architecture are just clearly growing pains and years of solve the problem right now. They went from $0 to $2B+ revenue in like 9 years.

My manager has also been around a long time and seen a lot of this so we are really in a good spot to start setting standards and best practices. Which is a new challenge for me and kind of exciting. The org is all in on this so its interesting how little internal friction there is for it. Not sure which one they want to tackle first but I am implementing the first one right now. These are just small changes but the org does not realize the impact of just standardizing certain things. The director does but its interesting its just never something they considered.
 
  • 3Like
Reactions: 2 users

Noodleface

A Mod Real Quick
39,549
18,130
Got a really short timeline to spin up a new server line with our product. I estimated just porting our code to be 3 weeks plus a week of validation. Now given I baked in a buffer because we all know things never go smooth.

One of my goals for the year was to utilize AI to increase platform porting, because I knew this wouldn't be the only one. I simply asked Claude to port our code and gave it two hardware schematics to compare the done platform and the new one.

20 minutes. It ported all our code in 20 minutes. I spent the next 2 or 3 hours massaging things, fixing very minor changes and at the end I had a full build. So let's say 4 hours total on the high end.

I definitely want to be the guy promoting and guiding AI rather than the guy opposing AI and manually coding. This stuff is breaking so much ground.
 
  • 1Like
  • 1Truth!
Reactions: 1 users

TJT

Mr. Poopybutthole
<Gold Donor>
46,819
128,188
Totally. I get to set AI workflow standards for this org so I am just now getting them in tune with it. We will be using Cursor as they already have a small footprint with it and I am the most familiar with it anyway. This manager prefers Claude Code but they don't even have it so we probably won't be going down that road. I want to try full agentic automation as one of their use cases is quite standard. They have a ton of Kafka stuff and they are setting it up in a really manual way. This is something you just have an agent do all of it and don't even think about it anymore outside of approving a PR on github. That's the direction this is going.

They do have Gemini but I haven't really bothered spending any time trying Antigravity. Which is the AI IDE Google spun up.
 
  • 1Like
Reactions: 1 user

dragonbr

Trakanon Raider
108
121
Got a really short timeline to spin up a new server line with our product. I estimated just porting our code to be 3 weeks plus a week of validation. Now given I baked in a buffer because we all know things never go smooth.

One of my goals for the year was to utilize AI to increase platform porting, because I knew this wouldn't be the only one. I simply asked Claude to port our code and gave it two hardware schematics to compare the done platform and the new one.

20 minutes. It ported all our code in 20 minutes. I spent the next 2 or 3 hours massaging things, fixing very minor changes and at the end I had a full build. So let's say 4 hours total on the high end.

I definitely want to be the guy promoting and guiding AI rather than the guy opposing AI and manually coding. This stuff is breaking so much ground.
I really feel like the ones opposing it fall into 2 categories. The first being people that used early iterations of chatgpt, saw the glaring slop issues with the lack of integration, and then formed their opinion and never looked back.

The second set are people that have seen what the latest models are capable of, and their survival mode activates. They convince themselves that its still bullshit due to the fear of change and the threat of becoming obsolete.

It's still by no means perfect but in comparison to what agents were capable of even just a year ago it's made massssssive jumps.

It definitely still requires guidance and hand holding. If you are giving it garbage prompts and or dont have a solid understanding of what the finished product should look like, then you are probably going to get garbage out. But even that issue is somewhat being solved with planning modes. After using that mode for a bit it still surprises me with how well it knows what questions to ask me regardless of the subject matter.
 
  • 1Truth!
Reactions: 1 user

Noodleface

A Mod Real Quick
39,549
18,130
I really feel like the ones opposing it fall into 2 categories. The first being people that used early iterations of chatgpt, saw the glaring slop issues with the lack of integration, and then formed their opinion and never looked back.

The second set are people that have seen what the latest models are capable of, and their survival mode activates. They convince themselves that its still bullshit due to the fear of change and the threat of becoming obsolete.

It's still by no means perfect but in comparison to what agents were capable of even just a year ago it's made massssssive jumps.

It definitely still requires guidance and hand holding. If you are giving it garbage prompts and or dont have a solid understanding of what the finished product should look like, then you are probably going to get garbage out. But even that issue is somewhat being solved with planning modes. After using that mode for a bit it still surprises me with how well it knows what questions to ask me regardless of the subject matter.
I almost fell into the first category. I wrote AI off when we first got access to the gpt models and llama. It just didn't understand our code. It was useful for making debug prints and stuff so I assumed when we moved to this it would be the same. But anthropic is so far beyond that it's actually overwhelming at times. It is writing UEFI code so quickly that if you know exactly what you need there's no reason for you to manually write anything. I've even used it on new features.

My survival mode is telling me to be the guy that knows how to manage AI.
 
  • 1Like
  • 1Solidarity
Reactions: 1 users

Deathwing

<Bronze Donator>
17,723
8,698
How about people that are opposed to it because the ethical, moral, security and environmental issues and economic costs haven't been sufficiently addressed?

Can you tell me how many man-hours it would have taken to do what Noodle did? Do you know if you ended up paying that cost in electricity and data center hardware depreciation anyway?

I'm more than willing to admit AI is cheaper, but I want to be able answer that question. The remote compute makes cost a giant black box.