Troy Hunt has been a pillar in the cybersecurity community with his Have I Been Pwned project for many, many years, even walked back a purchase of it because he felt they would compromise the principles of the project.
But now he has just gone all-in on AI. On a previous Weeky Update he published a "Robophobia Equality Policy" after receiving pushback to his openclaw usage. Since then he has posted more of his usage of AI, and in the latest he suggest that "people are getting into the program" because he seems to be receiving less pushback and because of Linus' affirmations

No one would have ever called it 'vibe coding' if they wanted it to be taken seriously, and/or respected.

putting SI as a alternate name in apparent support for trump's dumb shit is just the cherry on top of the shitcake

SI as an acronym is already very widely used globally, so it annoys me that it's encroaching on a more useful acronym also.

Sure in principle you could check all the output the AI shits out to see if it's right and then spend time correcting it either by hand or by prompting some more. And in doing so, one might be able to do more in a given amount of time, or less. Recent studies have shown it might have felt like programmers were able to do more, but in reality did less in the same time. But that might vary from person to person and the given tasks.

However that skips over other issues:

Reading and understanding code is so much harder than writing code. This has been known for decades. So instead of doing the easier task of writing the code yourself, you are left with reviewing and understanding the generated code. This add a lot of extra cognitive load, which can lead to fatigue quickly.
This fatigue feeds into another issue, where it's so much easier to just hit approve on everything and skip out on the really in depth stuff. People tend to slack and just glance at the code, which is a huge trap, because LLMs are very good at writing code that looks right, regardless of if it is right.

Then there's the issue where writing the code is a small part of the task usually. There's also understanding the task and figuring out edge cases. It's hard to do that all up-front and at least for me, while writing the code I think of edge cases or get a deeper understanding of the task. By flying through the writing part, the deeper understanding gets skipped as well.

And the issue of bias, where it's often a lot better to think about an open question instead of a multiple choice question. With vibe coding it's very easy to be lead down a path by the LLM, where it suggests a path which seems logical, but might actually be wrong. Then breaking out of that path can be hard. This might be an actual wrong path, or it might just stifle creativity. A common complaint is all AI generated stuff "feels" very samey, which makes sense because LLMs are designed to see common patterns in a lot of the data. The more common a pattern is, the more likely the LLM is to suggest it.

Another very common issue with LLMs is they tend to touch stuff they shouldn't. This is because of how they work is by processing everything as a stream of text. So let's say I want to expand a template with a new part, it has no choice but to read the entire template and then output the entire template as well. It's supposed to just add the new part, but often it makes other changes as well. And this happens on a multiple-file base as well.

One big issue I've noticed is the regression of skills for people who use AI for their work (coding or other). It's so tempting to just press the big green EASY button on just about any task. But then when the AI isn't able to do the task right, people are stumped and are no longer able to do the things they used to be able to do for decades before. Especially with coding, it isn't like riding a bicycle where once you know how, you are able to do it forever. It's a lot more like a muscle you need to keep excising and if you don't it's going to lose power and capabilities.

And we're not even touching other issues with LLMs in general, like concentration of power, vendor lock-in, obfuscated pricing, resource usage, (un)fair use of source data, hallucinations, manipulation, destruction of the internet etc.

So for me the people who say: "It's just a new tool, you wouldn't go back to hand punching paper tape right!??!!?" are being very disingenuous about the actual issues. Sure it's just a new tool, but one with real consequences to using it. As with every tool we need to weigh the pros and cons and just because it's available, doesn't mean we have to automatically use it. We still see people enjoying a command line or text-based interface, even though most people use graphical interfaces. And I personally code without any form of autocomplete, because I find it annoying and distracting. Just because tools exist, doesn't mean we should all use them and it certainly doesn't mean we should be blind to the downsides attached to certain tools.

Linux is not vibe coded but for some reason claim it is but macOS isn't. I see no basis for either assertion.

Offtopic, but how do you get the content from a blocked embed?

No shade to Troy, but he’s in the business of promotion and tech evangelism. He’s going to be on this train.

a feeling that those weaponising the term have now accepted

That's an interesting approach, to declare that everyone who says it "doesn't really mean it"

Linus Torvalds

Linus' take was more nuanced, I'm not sure I agree, but at the same time he was talking up AI, he also expressed that he saw it as useful for onboarding, review, and throwaway efforts in fields you don't get. Also he made the caveat that he doesn't really code the kernel so much anymore, so he explicitly stated he had lost some perspective. Also that it wasn't suitable for full-on "vibing" maintainable code. However AI bros have gone all in on feeling vindicated and feeling like Linus said the kernel should be vibe coded now.

Vibe coding is, by definition, not coding. It is the exclusive authorship of code by AI, specifically by a lazy developer. Yolo coding would be a better term, in hindsight

I knew this would happen the moment Linus made a statement. Id be very surprised if there wasnt a notable uptick in ai usage after he posted.

It's a Firefox feature. It blocks external calls to social media that might track you, then adds a summary below. I'm not sure how it works. I have the "AI kill switch" thing enabled and still get the summaries, so I assume it's not using AI. At least not locally. My guess is some proxy server requests the content, summarizes it (probably with AI), then adds it to the page.

Yeah, he's been like this for a while, here's a choice quote of his from back in April.

Demanding to “speak to a human” solely because the bot is an AI may constitute robophobic exclusionary behaviour and will not be tolerated.

However AI bros have gone all in on feeling vindicated and feeling like Linus said the kernel should be vibe coded now.

What I saw was mainly AI haters that couldn't form a well thought out take, saying that Linus has gone full slop and the kernel is now slop-first.
The debate was killed before even starting because one side can't express anything other than "slop", and refuse to hear the full arguments presented.

usually you would click on the little shield to the left of the url bar and it would say something like "allow content from twitter".

Yes I use Floorp but have never seen the summarization.

Yes I use Floorp but have never seen the summarization.

So his brain melted?

He thinks the LLM has personhood? Or feelings?

Fucking moron

Not sure if there's any difference in Floorp. I think you need to have enhanced tracking set to strict or custom though. I have my ETP set to custom with the settings shown here, and the summary is shown automatically. (I don't need to click anything, it's just there when the page loads.)

Nah there's opinionated idiots on both sides, one is organically ignorant and the other is parroting corporate published ignorance

Thanks for proving my point

LOL, no.

Vibe coding is still 100% a derogatory term. Go use MS Teams for a day if you think vibe coding produces anything useful. There are new bugs every single day that we have to work around, as MS apparently vibe codes this shit AND no longer properly tests their work product before releasing it.

you can use AI to write code, but you still have to check behind it.

And this is, in large part, why AI is an unprofitable, useless product. So much of the promised productivity is crap, because if you have to validate and test anyway, it takes less time in the end to do the work yourself. If you're building a tech product, you also need to understand how it works because you need to be able to service it properly, and frequent AI use dulls that ability too.

Every single person who uses AI, by now, has seen it hallucinate bug out often enough that it's demonstrably not something which produces trustworthy output.

handing things over to a computer without knowing what it’s doing

So, it's very easy to understand different levels of abstractions and how the lower level abstractions fit into higher level ones.

Like the simplest adder circuit can be expressed in a series of nand gates, a nand gate can be implemented in various ways on the silicon. In the other direction, you have a group of circuits than can be driven by a clock to demarc discrete states, and the collective you have becomes memory, CPU, FPGA etc.

The "software" is similar, truth tables, machine state etc whatever it is on to the highest level software.

The act of vibe coding abdicates understanding the abstraction by the programmer, and if they take up that abdication they cease to be a programmer, or a professional in general.

Edit fixed grammer but this looks worse

Ouch you should read over the posts, I definitely did not.

While it is a fair observation that critics also failed to recognize nuance in Torvald's take, it is a bit wild to assert that it was 'mainly AI haters' that didn't get the nuance in a post specifically about someone so pro-AI they declare that everyone agrees with their take, and even people who claim otherwise don't mean it and secretly agree with him, and that Torvalds gave an unequivocal thumbs up to full on vibe coding.

Its been shitty before vibecoding

You're not wrong.

SI

😭

Anyone calling it Super Intelligence is a moron anyway.

Heck, go use GitHub for a day

I would love to live in a world where robophibia is a legitimate issue we need to overcome, but that's not anytime soon.

Great way to have people tell on themselves, however.

A red flag doesn't even describe it. It's more like a jolly roger or a hazmat placard. They are announcing their tacit support for an idiot fascist.

It's just the whole dam breaking, really. First the support then everything else, until there is nothing anymore.

I feel like you intentionally left out some context there:

This is so "peak 2026" - writing an equality policy to ensure people treat our AI bot with the same respect as they do their human counterparts. It's intentionally a bit tongue-in-cheek, but it's there for a purpose: we simply don't have the capacity to deal with every request we get, and we need Bruce to be the coalface of support.

The genie is out of the bottle unfortunately. Ain't no putting it back. It's here to stay one way or another.

eyes closed coding

brain nope coding

crap coding

slop coding

douche coding

hands free butt plug coding

stupor intelligence coding (sic for short)

derp coding

not coding

bro coding

kardashian koding

aihole coding...

preferably as open source models no one cares about that are seen as the useless slop makers they are

robophobic exclusionary behaviour

good punk band name

Is there an argument to be made other than slop? All i have seen as far as arguments for is: "this other thing is slop, so why can't I slop here?", "what if it's only a little bit of slop" or "what if I serve you slop anyway and you don't notice".
I don't care if you only put a thimbleful of dogshit in your soup, it's still shit soup.

Fuck HBIP, fuck Troy Hunt.

Yeah, he should be in an asylum...

We saw this happen with Kent Overstreet too. I think the takeaway from these cases is that developers need mental health care. AI is just the canary in the DSM coal mine.

I've been waiting my whole life to stand in the living room drunkenly yelling "No daughter of mine is going to marry this... fucking roomba!"

And she'd be all crying like 'That's why I couldn't you tell you I was robosexual, dad. You're a hypocritical bigot! Mom would roll over in her grave if she knew what you and that Bimbo4000 subscription on your desktop were doing!"

Then I'd punch a hole in the wall, which would be easy because all public housing would just be refrigerator cardboard boxes and yell "Why can't you be normal like your transmasc ace furry brother??"

Then she'd run out of the house crying and jump into her lover/Electric VTO transport that also plays in a band and burn thruster donuts in the front yard.

But nooo, we get this stupid shit going on here.

Sans Intelligence.

It's a hype cycle like any other. Some changes will last, others won't.

Wrf is a coalface is that an aussyism?

Linus also said competent, which means it’s being tested before it’s committed, etc. This person is just fulfilling his bias. The AI cuck coders won’t be testing because they’re tired of testing AI’s fuck ups.

^ Non-believer in regulatory power.

Spreadsheet Impersonator?