Kaligalis

It is - and tons of bugs are just about fucking it up or just not doing it at all. Incomplete or faulty parsing and validation of untrusted input is absurdly common. It got better in the past decades, but most software developers write worse code than AI by now (not me though; obviously, I am still better than the best AI).

We know that some humans can be trained to do that just fine. Humans are natural neuronal networks. That implies, neuronal networks can in principle do it. We just don’t have any human-capability artificial neuronal networks yet.
LLMs might never get there. But humans aren’t LLMs. If we ever manage to properly model a human brain, that probably will be able to do that task with human-level accuracy (which actually is pretty good if you only look at professionals of the filed).
Hopefully, it doesn’t actually need a human brain for the task - because modeling that might still be a century off.

Sales and marketing don’t count. Critical thinking doesn’t sell. So you won’t find critical thinkers in those departments.
From a security standpoint, those departments are to be considered hostile. But you can lock down the PCs there as much as possible to reduce the offline time because computer-illiterate employees don’t care about being able to install stuff or change settings.

They have a domestic market to serve first. And they are fast at scaling up - but not that fast.
Give them half a decade.

I wouldn’t enter a cartel-protected market either if I had to finance that myself. If I had a captive bloc-sized market at home though…
China is big. There are a lot of customers there. And the government tends to plan long-term and is heavily subsidizing chip making tech for geopolitical reasons. Entering the western market is high-risk. Entering the Chinese market as a Chinese company and then also selling outside China after scaling up is low-risk.

For western-oriented companies, it’s a trap. All RAM makers outside China are in a single cartel. When anyone tries to undercut prices, they all lower their prices too and starve out the new competitor. Then, prices go up again.
This is why only China can solve the RAM crisis. They need their own full-scale domestic RAM production for geopolitical reasons and are therefore not disincentivized by the cartel existing.

Bragging about how product is at doing something is the easiest way to advertise product. No conspiracy theories needed.

They want you to use their LLM to screen your code for bugs. When you do that, you see how surprisingly well it works now and that leads to you wanting to also use it for generating code. That is when you realize that the 20 bucks starter subscription isn’t enough to actually use it professionally… which means, you upgrade and they get moar money.

Microsoft being slow to fix bugs isn’t news.

He can do what he wants - except closing Pandora’s box. It stays open. The race is on, whether he participates or not.

Is there no criminal investigation about this? Whoever actually did it should go to jail to heavily disincentivize planned crimes like that.

Because it’s literally impossible to solve. The party which would do that would literally be guaranteed to lose the next election and the other party would then revert the decision. A massive part of the US population gets really pissed when you try to take their guns (or even just the possibility of them having guns in the future). Trying to actually restrict gun ownership is political suicide in the US.

What could be done is solving the actual problem: Kids turning into psychopaths by being bullied over years until their mind breaks.

Actually restricting guns for the plebs is impossible in the US. The right to bear arms is literally in the constitution. And when looking at the (pretty damn short, but it does exist) history, it makes perfect sense (even though it doesn’t actually work to keep the government honest and the feared British revenge war never happened).

What would be possible even in the US would be making it less likely that the abundant guns (which just ain’t going away anytime soon) are used by people with serious mental problems - by reducing the likelihood of serious mental problems to happen.
You need a village to raise a kid.

Damn. I wished we had those in our school back then. What mischief I could have done with a borrowed drone like that…

I didn’t see any CSAM on Reddit. But I can confirm that there indeed is plenty of adult porn there.

They should have learnt to do proper QA before they get to college. That is the actual problem to be solved.

The problem isn’t that they cheat. It is that they are not doing proper QA when using AI.

Everyone will use AI (or AI-powered robots) in some way or another for basically everything in a few years. Companies banning AI will just disappear. People who see using AI as cheating will become unemployable.
But no one knows when the massive AI reliability issue will be fixed. It can be tomorrow, but it can also take years or maybe even forever. Until then, AI requires a human doing proper QA to make the results reliable in almost all cases.

Doing QA mostly requires a completely different skill set and attitude than what the educational system teaches. That needs to change. Good QA is also important for doing anything without AI. But with AI, it’s almost the only thing left for a human to do.

Btw, the article is bullshit. When the students select the text to copy it, the white on white text becomes visible. That is when people intelligent enough to be able to use AI (and yeah, that bar is really low; AI is pretty easy to use) are usually detecting that there is something fishy going on.

The idea is probably to have an easy joker reason to randomly ban any repo which is for some otherwise hard to justify reason not wanted on the site. It’s a common corpo practice to have extra vague and impossible-to-disprove rules like that.

Just one problem with that: Every piece of work - code or otherwise, made by a human or not - may violate copyright law. There really is no way to know without actually checking against all other works (which you can’t because you don’t have access to all of them).

What GitHub does is what a sane person would do: React to claims. They just assume that what they get uploaded isn’t violating copyright until someone claims otherwise. And as it is literally impossible to actually know, that is the only sane way to approach this problem apart from just not hosting other people’s stuff at all.

Just use that cat girl captcha thing, every other FOSS site uses, and the problem is solved.

Let’s be real here: Most future projects will be mostly written by AI.
Sure, the good ones will have proper human QA. But Claude Code is so much faster at writing and rewriting surprisingly good code, that ignoring it is just like ignoring IDEs was a decade ago.
You can still write code without AI, syntax highlighting, code-suggestions, documentation tooltips, inspections, and fully integrated version control. But all these tools legitimately make software development easier.