Programming thread

  • 🔎 If you see an alert for analytics.usips.org, this is a self-hosted system I am running as a test. It is GDPR compliant and collects no invasive information.
  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account
UERISIMILITUDO said:
I've read an account of a teenager waking up with police guns in his face for the crime of bypassing security mechanisms protecting a video game
Click to expand...
More specifically, he hacked into Valve's servers and leaked just about everything they were working on, including Half-Life 2. One, he was fucking stupid and fell into an obvious trap, two, he was German, and he got caught in Germany, and Krauts are one of the most despicable subhuman nations on this planet, and they in particular have a gigantic hard-on for enforcing copyright law to the point where having a gun pointed at your face for pirating a movie would be seen as perfectly normal by these untermensch.

Here, no one would care, piracy is a non-crime. Unless you pirated a dogshit comedy drama no one in their right mind would watch. But anyway, that story, besides reminding us that copyright law should be abolished, also reminds us that Germany should've been carpet bombed every 50 years as a preventive measure.
 
It's bullshit to compare Claude to the average person.

Name a single person in whom the entirety of human intellectual output and billions of dollars was invested?

The average person is given a bullshit public education filled with garbage ideas like whole language reading and no child left behind, all designed to stifle students for the sake of keeping retards in the running with normal people. Gifted people are treated with contempt and given no support in favor of equity in education. Higher education is treated like a scam where everything from tuition and textbooks are up-charged to ridiculous prices because the government is subsidizing it, and the government doesn't care about this exploitation because the debt is given to the student.

And yet the average person is held to the standard of the biggest and most expensive machine ever invented, one which would be nothing without the consolidation of mankind's total accumulated knowledge?

Even then, it still doesn't equate to a real human brain, the way they work is fundamentally different. There is no such thing as a "context window" in human neurology, there is no central processing unit, the brain is a series of branching neurons that interact and disengage dynamically, computers are made up of metallic, static lanes that operate linearly and determine value by the number of times an electron is fired and received, any imperfection or sudden change to this system will destroy it. The human brain doesn't operate solely on electricity, computers don't possess hormones like serotonin or oxytocin. The human brain isn't restricted to one build, it's constantly culling and replacing neurons and undergoing restructuring and information consolidation.

And on the notion that humans can't create original ideas, mankind invented math and writing. There was once no concept of a zero in mathematics, then someone invented it. There was once no concept of squaring a circle, then someone invented that. There was once no concept that gravity and time were interrelated, then someone came up with it. There was no concept of a Turing machine, then Alan made one.
 
New AI productivity study dropped: Artificial Intelligence in the Firm: Bottlenecks in Software Production.
It is still data from pre fable era(january 2021 - march 2026), so take it however you want.
My main takeaway from it is that Jira tickets/epics are not completed faster in statistically significant way. 1791508333047.png
Yet lines of codes increase:
1791508382450.png
They also state main bottleneck being code review. But we already have been given solution to this problem: just stop reviewing it!
 
Least Concern said:
Hopefully I don't need to tell anyone here how useful LoC and commit counts are for determining software quality. But they sure are easy numbers to graph, aren't they?
Click to expand...
I'd add "tickets closed" is also often a near-useless measurement of actual productive activity (one "hard" ticket solved that cures a problem thousands of users experience is seen as "less valuable" to this kind of bean-counter than a dozen "password reset" tickets closed by a tier-1 support technician).
 
moocow said:
I'd add "tickets closed" is also often a near-useless measurement of actual productive activity (one "hard" ticket solved that cures a problem thousands of users experience is seen as "less valuable" to this kind of bean-counter than a dozen "password reset" tickets closed by a tier-1 support technician).
Click to expand...
Epics completed should be a better measure. The sample size is in this study is pretty big as it comes from 718 firms with 725,938 workers, of which 55% are devs. So unless usage of AI itself affects epic sizes, that's probably the best measurement we can have.
 
ADHD Mate said:
New AI productivity study dropped: Artificial Intelligence in the Firm: Bottlenecks in Software Production.
It is still data from pre fable era(january 2021 - march 2026), so take it however you want.
My main takeaway from it is that Jira tickets/epics are not completed faster in statistically significant way. View attachment 9682134
Yet lines of codes increase:
View attachment 9682139
They also state main bottleneck being code review. But we already have been given solution to this problem: just stop reviewing it!
Click to expand...

Anecdotes don't change signals.
Organizational things are organizational.

I still work so much fucking faster with AI.
I get log jammed by QA.
 
moocow said:
Oooooh don't let the anti-AI fags in the "AI solves maths faster than mathematicians" thread hear you say this, it'll boil their blood and they'll remind you how everyone died when Ford invented the assembly line :story:
Click to expand...
AI isn't going to solve mathematics.
AI is going to clear one kind of log jam, build tools to go after things we haven't done yet, and be a partner along the way.

Absolutism is retarded.

If you want to be humbled, but also see the fractal "haha fuck you" jack-in-the-box from hell, look at P!=NP. It's obvious. But try proving it.

Wait, can't use that, because it's vacuous.
Can't do that because you're restating the problem!

Wait, proofs are programs?! Well wait is the proof program polynom- fuck.
 
Sneed's Feed n Seed said:
AI isn't going to solve mathematics.
AI is going to clear one kind of log jam, build tools to go after things we haven't done yet, and be a partner along the way.

Absolutism is retarded.

If you want to be humbled, but also see the fractal "haha fuck you" jack-in-the-box from hell, look at P!=NP. It's obvious. But try proving it.

Wait, can't use that, because it's vacuous.
Can't do that because you're restating the problem!

Wait, proofs are programs?! Well wait is the proof program polynom- fuck.
Click to expand...
Yeah, their outrage utterly blows my mind. "OMIGOD this random tool is solving our cute math problems!" So fucking what? Come up with harder math problems and use those tools to solve them too! Advance our understanding of the universe with whatever tools you can. That's the big win. Not the sinecure of comfortably butting your head gently against "hard to solve" problems. But oh wait, they want the sinecure, not to actually solve the problem. lol my mistake :story:
 
CS glazing the proof barriers makes me want to smack people with my cock. Yes, duck. I'm about to turn left.

TALK TO OTHER DEPARTMENTS! CHANGE PERSPECTIVES! Academia is full of faggot silos. Actually shaking my damn head.

AI will likely help clear out stuff like "did anyone fucking check!?" (Jacobian, etc) and "we didn't do a literature review or invite other silos in". It will speed up things where it can automate out the cruft and boilerplate. Great.

Tao's a faggot atm. I'd expect more from him.

Nigger, actual fucking retard, do you think we grasp how drugs work half the time!?
 
Goose the Redeemer said:
Are there any good resources for "modern" webdev/hosting/general sysadmin stuff? Security in particular.
Click to expand...
Now there is obviously more to it and I'm not claiming to be an expert, but classical web exploits come from not sanitizing the shit out of your input enough.
Perl has this cool feature called tainted mode which tracks what values came from the outside and will scream at you if you try to use one unsafely.

You may also want to look up `Bubblewrap`.

As for ports, setting it to something unusual helps against those who are batch scanning ips. The extra compute adds up, especially when they would have to verify what kind of service it is.
 
I'm tracking statistics about requests for this personal project.

Compiling these stats is expensive and affects performance, so I can't just analyze every request. So I have sampling where every X (say 20) request gets analyzed.

The service consuming this data can just assume all the numbers are actually 20x higher, with the caveat that it's super imprecise and there's gaps. There's enough churn with the requests and the type of data I'm collecting that you'll probably see evidence of whatever you're looking for. The numbers will just be fairly useless.

So in addition to the sampling, I have a selective "exact" data collection scheme where the consuming service can set specific criteria for requests where they get full, exact data collection. "get me exact data on requests from this IP address range"

So the consuming service can essentially mark some types of requests as "suspicious" and it'll return exact stats for them.

There's no technical problem here. This code is a comfy, lazy Sunday afternoon project.

But the naming is driving me up the wall. Like conveniently, "sample" is a noun and a verb.

So my sampling function (it gets called on batches of requests) is RequestFrequencyTable.sample_requests. It gets called with every batch of requests and internally keeps the request modulo counter. I don't know what to name the exact version though. I'm just saying fuck it and calling it RequestFrequencyTable.exact_requests. I know exact isn't a verb. Anything else is going to irk me more though. I need my metaphors and naming to line up or else I'll be thinking about this tonight when I'm trying to fall asleep.

Idk, I guess I could rename the older sample_requests function.

Technical stuff in programming is fine and dandy. But the real challenge is the autism in naming shit.
Lard Shart said:
I could not continue watching after she praised a paper about feminism in glacier research, putting forward the idea that male researchers climb mountains for bragging rights.
Click to expand...
Lol when I was a kid, my dad used to tell me "y'know what people say when they ask why they climb mountains? they say 'because it's there' " and then he'd cackle.

"Because it's there" the most airheaded hippie, "live laugh love" type thing to say. As a joke, I didn't get why it was so funny until I was an adult.
 
You must log in or register to reply here.
Back
Top Bottom