aspe:keyoxide.org:JKVZEDBRMLKHYWMJWR6L6UDDY4
- theshrike79AFAIK the enemies are also trained with machine learning.
- Some people’s brains are just wired for that.
There are people who are just fine being a cog in a machine doing the same thing all day for years on out.
My family is from a factory town and many of them were literally standing next to a conveyor doing a monotonic task that could’ve done by a robot.
I tried it for one summer job and my brain almost melted from the boredom and monotony.
- Kagi or DDG
Google is useless
- I've completed actual features by saying "look up issue ABBA-1234 and create a plan to implement it" to Claude.
Then I wait, look through the plan and tell it to implement and go do something else.
After a while I check the diffs and go "huh, yea, that's how I would've done it too", commit and push.
- If we extrapolate the Dr example:
There is the one doctor who learned one way to do the operation at school, with specific instruments, sutures etc. and uses that for 1000 surgeries.
And then there's the curious one who actively goes to conferences, reads publications and learns new better ways to do the same operation with invisible sutures that don't leave a scar or tools that are allow for more efficient operations, cutting down the time required for the patient to be under anaesthesia.
Which one would you hire for your hospital for the next 25 years?
- Expert Generalists are a thing: https://martinfowler.com/articles/expert-generalist.html
BUT they're completely wasted if you just use them to turn JIRA tickets into end to end features =)
- There are a definitely few ulcer-inducing events in my past that would've taken me an afternoon to fix with a current SOTA LLM vs 2+ weeks of swearing, crying and stressing out.
- Depends a lot on the type of software you're doing. Startups will have hungry people willing to learn, more traditional companies won't in the same percentages.
Not all people are curious, they go to school, learn to code and work their job like a normal 9-5 blue collar worker. They go to company trainings, but they don't read Hacker News, don't follow the latest language fads or do personal software projects during nights and weekends. It's just a day job for them that pays for their non-programming hobbies.
I've had colleagues who managed the same ASP+Access DB system for almost a decade, with zero curiosity or interest to learn anything that wasn't absolutely necessary.
We had to drag them to the ASP.NET age, one just wouldn't and stayed back managing the legacy version until all clients had moved to the new stack.
...and I just checked LinkedIn, the non-curious ones are still in the same company, managing the same piece of SaaS as a Software Developer. 20-26 years in the same company, straight from school.
- It's the old saying: "$10 for the part, $990 for knowing where to put it"
You get a feel for what works and what doesn't, provided you know the relevant facts. Doing a 10RPS system is completely different than 300RPS. And if the payload is 1kB the problems aren't the same as with the one with a 10MB payload.
And if (when) you're using a cloud environment, which one is cheaper, large data or RPS? It's not always intuitive. We just had our AWS reps do a Tim "The Toolman" Taylor "HUUH?!" when we explained that the way our software works is 95% cheaper to run using S3 as the storage rather than DynamoDB :D
- If you really distill it, I've been doing API Glue for about a quarter century.
I connect to a 3rd party API with shitty specs and inconsistent output that doesn't follow even their spec, swear a bit and adjust my estimates[0]. Do some business stuff with it and shove it to another API.
But I've done that now in ... six maybe seven different languages and a few different frameworks on top of that. And because both sides of the API tend to be a bit shit, there's a lot of experience in defensive coding and verification - as well as writing really polite but pointed Corporate Emails that boil down to "it's your shit that's broken, not ours, you fix it".
At this point I really don't care what language I have to use, as long as it isn't Java (which I've heard has come far in the last decade, but old traumas and all that =).
[0] best one yet is the Swedish "standard" for electricity consumption reports, pretty much every field is optional because they couldn't decide and wanted to please every company in on the project. Now write a parser for that please.
- Yea, I don't understand why people use LLMs for "facts". You can get them from Wikipedia or a book.
Use them for something creative, write a short story on spec, generate images.
Or the best option: give it tools and let it actually DO something like "read my message history with my wife, find top 5 gift ideas she might have hinted at and search for options to purchase them" - perfect for a local model, there's no way in hell I'd feed my messages to a public LLM, but the one sitting next to me that I can turn off the second it twitches the wrong way? - sure.
- IMO LLMs are forcing us in the other way.
To get the maximum ROI from LLM-assisted programming it needs proper unit tests, integration tests, correctly configured linters, accessible documentation and well-managed git history (Claude actually checks git history nowadays to see when a feature was added if it has a bug)
Worst case we'll still have proper tests and documentation if the AI bubble suddenly bursts. Best case we can skip the boring bits because the LLM is "smart" enough to handle the low hanging fruit reliably because of the robust test suite.
- Me: "Boss, this takes at least 4 weeks to complete properly including QA time."
Boss: sucks in air through his teeth "Best I can do is one week. Get to it."
Me, with a massive mortgage and the job market is shit: "Rogerroger, bossman"
- Not all problems are "work" vs "doesn't work".
We're not talking about making a calculator that can't calculate 1+1. This might be a website that's a bit slow and janky to use.
25% of users go away because it's shit, but 75% stay. And it would've too much effort to push the jank to zero and retain a 100%.
A website that takes juuuust too long to load still "satisfies requirements" in most cases, especially when making loading instant carries a significant extra cost the customer isn't willing to pay for.
- "You can't polish a turd" =)
- We have an AI doing the first pass PR review using company standards as a prompt.
It catches the worst slop in the first pass easily, as well as typos etc.
- I've had calls with Principal Architects who couldn't code themselves out of a wet paper bag.
And according to the company experience chart, they should've been a "thought leader" and "able to instruct senior engineers"
My title? Backend Programmer (20 years of experience). Our unit didn't care about titles because there was a "budget" for title upgrades per business unit and guess which team grabbed all of them =)
- A coworker had this anecdote decades ago.
There's a difference between 10 years of experience and 1 year of experience 10 times.
YOE isn't always a measurement of quality, you can work the same dead-end coding job for 10 years and never get more than "1 year" of actual experience.
- It’s also possible to implement an MCP as a skill
- They’re not a perfect solution, but they are a good one.
The best one we have thought of so far.