Linkblog

Faulty Towers, vibe sickness, and the vibe bobsled

Sat, 18 Jul 2026 04:44:14 +0000

I know. As if what the world needed was yet another blogpost about LLMs and AI tech. Yet there is a pile of things which have been on my mind, and I haven't seen them laid out elsewhere in the way I'm going to write them, and so here we go.

I still don't use genAI to write my articles, fwiw. Here or anywhere else. These rambly words are my own.

The tower tilts

I read The Tower Keeps Rising recently, and it has stuck in my mind.

The piece is an observation, and according to Armin on lobste.rs, it is not an advocacy for the state of affairs (though by running a vibecoding company, Armin is part of advancing this direction):

For context: I'm the author. I intentionally did not make a judgement if this is a good or bad thing, or if this is going to continue working. It's primarily an observation that with agents you can continue to make progress even when people on the team maneuvered themselves into situations where previously they would have needed to talk to each other.

The summary of Armin's post is effectively that vibecoded systems keep piling code on top of code, but in many systems things seem to keep building, but the abstractions keep piling on, but eventually no human can understand the codebase. But this is a new way of operating, because LLMs can "explain" a part of the codebase that no human can make sense of, and so continue building.

Even if such systems continue to work, I find two things: 1) that now advocates for this state of affairs have pivoted into acknowledging that this is the end state of their systems and 2) they seem to be accepting it as the way forward.

Regarding the first, I think it's very important to note that this is a shift. Simon Willison, probably the best pro-genAI writer on the internet (sometimes, I think, giving cover for a lot of weaker writers, but is that Simon's fault?), at one point coined the term "agentic engineering" and was very clear to draw a line in the sand between agentic engineering and vibecoding:

We also need to read the code. My golden rule for production-quality AI-assisted programming is that I won’t commit any code to my repository if I couldn’t explain exactly what it does to somebody else.
If an LLM wrote the code for you, and you then reviewed it, tested it thoroughly and made sure you could explain how it works to someone else that’s not vibe coding, it’s software development. The usage of an LLM to support that activity is immaterial.

In an incredibly short period of time, basically a year, Simon published a fairly honest article titled Vibe coding and agentic engineering are getting closer than I’d like:

The problem is that as the coding agents get more reliable, I’m not reviewing every line of code that they write anymore, even for my production level stuff.
I know full well that if you ask Claude Code to build a JSON API endpoint that runs a SQL query and outputs the results as JSON, it’s just going to do it right. It’s not going to mess that up. You have it add automated tests, you have it add documentation, you know it’s going to be good.
But I’m not reviewing that code. And now I’ve got that feeling of guilt: if I haven’t reviewed the code, is it really responsible for me to use this in production?

It's a great read, and what I will say is that I applaud Simon's honesty and willingness to self-reflect and challenge prior statements.

But the gap of time between the former and latter articles are stunningly short, just slightly over a year.

And Simon isn't alone. Just a year ago, I think the memetic shape was by and large that something along the lines of "agentic engineering" is what people could or should do, and, though I think many people are hesitant to admit it, I think most people using these tools are tending towards vibecoding and not agentic engineering, just as Simon himself found himself pulled.

Before we look at the consequences to this, I think we should look at why it's happening.

The vibe bobsled

As far as I know I'm the only person who uses the term "vibe bobsled" and, well, I doubt it's a term that's particularly likely to catch on, but I find it personally useful.

Bobsledding, if you are unaware, is a particularly strange and interesting sport. It's a lot of fun, but you don't have a lot of agency in it. You sit in a bobsled, you go down an icy track, and really, there is only one way to go. But people can become experts in it, and can indeed measure themselves against each others skills; it's an olympic sport, and I remember my own first encounters with bobsledding as a child, when my father and uncle and aunts took me, and it was thrilling like a roller coaster and intoxicating upon my first encounter.

But again, ultimately, there's only one place to go.

The vehicle is the LLM, you are the passenger. And I think the amount of agency people have over their journey is greatly reduced from what they feel like it is. More than just a slippery slope, it is a pre-crafted journey.

At the top of the chute, people tell themselves they're going to use these tools as a kind of fancy autocomplete. As they descend, they say they'll spin up some agents to explore ideas, but they'll write the code themselves. Next their agents are generating the code for them, but don't worry, but they'll review all the output. Soon they're plummeting downward and well, they don't actually review the code being spat out much anymore, but they trust the agents, heck maybe the agents are actually better coders than they are they say. And where does it go from there? From "I don't even code anymore" to "I don't even prompt anymore"?

At every stage of the process, the coder in question removes themselves from the process of producing code, and gives in towards a faith-based initiative of code production, that the LLM knows and does a good job of what it's doing. But what is the source of gravity pulling the sled along this icy chute?

It's simple. Generation is not the slow part of coding. Theory-building and review are. And plausible-enough things are extremely hard to debug and understand. But the machines are so fast at producing things. If you are going to review their work, you aren't really taking advantage of their most powerful property, which is speed. But theory-building and review are also the programmer's most important role.

To illustrate just how hard it is to detect and review problems with something which appears plausible, let's look at Ka-Ping Yee's remarkable dissertation, Building Reliable Voting Machine Software. It is a wonderful read, and highly approachable.

In the section "What makes software hard to verify?", Ka-Ping recognizes several major reasons why software is hard to verify: number of components, complex interactions, far-reaching effects, and nonlinearity. Notably, all of these problems are exacerbated by the patterns of code generation by LLMs. Still, let us leave that aside.

Ka-Ping constructs a model voting machine, and decides to see how hard it would be to verify that we know it behaves correctly. To push that exploration to its furthest, Ka-Ping Yee and David Wagner try an interesting experiment:

David Wagner and I decided to insert three bugs into Pvote to see if the reviewers would find them. We inserted what we thought would be an “easy” bug, a “medium” bug, and a “hard bug” to find, and chose each bug individually in such a way that an insider could conceivably exploit the bug to influence the results of an election. [...]
We decided to insert all of these bugs in a 100-line region of a single file, lines 11 to 109 of Navigator.py, and told the reviewers to look in this region. We did this both because the navigator was the most interesting in terms of the program logic and because we knew the reviewers would have limited time. The new version of the code that we gave the reviewers contained all three bugs, but we did not tell the reviewers how many bugs there were.
Yoshi Kohno, Mark Miller, and Dan Sandler participated as reviewers on the third day of the review. Dan was very familiar with Python and found the “easy” and “medium” bugs quickly, within about 70 minutes. Yoshi Kohno and Mark Miller found the “easy” bug after about four hours of reviewing. None of the reviewers found the “hard” bug.
Ian Goldberg and Yoshi Kohno participated as reviewers on the fourth day of the review. Ian Goldberg also found the “easy” bug within about two hours; none of the other bugs were found on the fourth day.
The reviewers spent a total of about 20 reviewer-hours focused on the task of finding the bugs in this 100-line section of Navigator.py.

Ka-Ping chose from a highly seasoned group of reviewers who were even deeply familiar with security threats. Mark S. Miller has been a personal mentor to me throughout my career, and is one of the programmers I have learned most from and studied the work of most closely. I talked with him about the experience at one point. He remarked on how it took a significant amount of time to find the easy bug, hours to find the medium bug, and that nobody could find the hard bug... but the big observation (which was said to me personally, and is not recorded in the dissertation) was that "the astounding thing is that once the bugs were pointed out, we all agreed that they were retroactively obvious, and that we should have been able to find them!"

If some of the best programmers in the world struggle to find bugs they even know must be there within a 100 line program, there is simply no hope for humans to review the volume of output from LLMs.

And so there is only one thing to do: don't bother. At each step, remove yourself. You tell yourself you won't, but you do. You give in to the chute and the shape of the vibe tunnel, and down you go.

Vibe sickness and the colonial settlers of the uncanny valley

The term "AI psychosis" is thrown around a lot these days to describe anyone or any state of poor behavior or outcomes due to genAI usage. But the original description of "AI psychosis" was closer to something clinical, a description of people quite literally experiencing psychosis from encounters with chatbots which reaffirm far too much of the user, spinning them into spaces of delusional detachment from reality.

But we should have something to describe the general sense of unwellness that seems to be befalling this world, and the best phrase of which I first saw in a post from Glyph:

On the way home from #PyConUS 2026. Quite an experience this year; very intense. No point in sugar-coating the part where there is a pervasive vibe-sickness, open source is suffering a massive sustainability crisis, slop security PRs are overwhelming everyone (etc etc). But there was a lot of hope, a lot of energy, a lot of effort toward mutual understanding, and (surprising to me) a lot of appreciation. Including for my own work, both writing and coding.

I like the phrase "vibe sickness", and if you aren't speaking of a form of literal psychosis, I think it's a better phrase.

Vibe sickness is everywhere... heck, perhaps we are on the verge of, if not experiencing, a vibe epidemic. Everyone complains of slop, and yet nobody using these tools wants to self-describe their outputs as slop. Yet slop seems to be everywhere, and infecting one's everyday experience: posters of food at your local restaurant that have plausible and yet incomprehensible designs, the helpdesk support chatbot you wish you could make physically manifest so you could throw it off the edge of a cliff, age verification code, and if you're a maintainer of an open source project, slop issues and pull requests.

The thing is that all of this tooling is useful for some things, but the term "genAI" points at exactly what it's worst at: generating things. If we want to talk about finding problems, it's a different story. But even leaving aside the quality issues of the growing and wavering tower, there comes the problem of lack of understanding of how it is built, constructed, and maintained.

The worst part of all this is you can't opt out. A colleague or an open source contributor sends you a "generous contribution" that is absolutely slop and certainly not understood by the person who submitted it. You're left sitting there, parsing whether or not you're going to be rude even to ask if this is LLM generated, or to unwittingly become a user of vibecoding workflows yourself by indirectly interacting with the agent through trying to respond to the issue/PR.

You can't escape.

To quote Glyph again:

Protesting LLMs by refusing to use any software that includes them feels like attempting to protest the introduction of tetraethyl lead into gasoline by refusing to breathe until everyone stops putting it in their cars. So I am drawing my personal moral lines in such a way that I will probably accept this.
But please don't mistake this for excitement about huffing a bunch of vaporized lead.

Well put. In the meanwhile, the uncanny valley colonizes our world and transforms it into its own.

But there is a rot growing beneath our feet and within our walls as our world is swapped out with systems nobody understands. And I fear what it's going to be like to recover from this all, the price we do not even yet realize we are going to have to pay. For this reason, any project that chooses not to engage with current genAI stuff, I tend to gain respect for.

But perhaps you can't opt out from aiding the problem. Your work won't let you, you're stuck in some situation... I don't mean to judge, but I do mean to end with a perspective.

Sometimes I sit as the passenger of a car and watch the driver of said car get angry at someone biking on the road. It feels bad, because sometimes I bike on the road, and without sufficient infrastructure, bicyclists risk getting hit by car doors parked on the side of the road, squeezed out by impatient drivers, etc.

I also drive. When I do, and there's a bicycle in front of me, I pause, and as the world broils to death, I take a moment to be thankful for their presence, and to think about how we could change the shape of the terrain to allow bicyclists to participate more safely (which would also help me drive more easily too, or choose to bike when I can).

May we not give up on ourselves, and not lose faith in our ability to participate towards building a better world.

Zig Creator Calls Spade a Spade, Anthropic Blows Smoke

Mon, 13 Jul 2026 09:29:25 +0000

"In the AI bubble, we are pressured to build things nobody wants, poorly. Pop the bubble."

Redirecting from /post/zed-creator-calls-spade-a-spade/ to /post/zig-creator-calls-spade-a-spade

Garbage for humans

Sat, 11 Jul 2026 06:31:20 +0000

unstory

If this is the solution, I want my problem back.

2026-06-29 01:49

Few months ago Cloudflare introduced Markdown for machines:

As a business, to continue to stay ahead, now is the time to consider not just human visitors, or traditional wisdom for SEO-optimization, but start to treat agents as first-class citizens.

Feeding raw HTML to an AI is like paying by the word to read packaging instead of the letter inside.

You get HTML, trackers, cookie banners, popups, and JavaScript. This needs a browser built by thousands of engineers just to show you a paragraph.

Machines get the clean version.

The AI Hype Reckoning Is Upon Us

Thu, 09 Jul 2026 06:35:33 +0000

The massive chasm between AI hype and reality is finally reaching the breaking point.

that is an amazing idea and you are a genius, dave

We're living in split realities.

There's what modern software is actually capable of, and then there's the gargantuan pile of "AI" hype, fraud, and bullshit our biggest tech companies (and their lazy enablers in the tech press) have shoveled down the public's throat for the better part of the last five years.

There's useful automation software that makes it easier to code, draft a new resume, or study vast repositories of scientific knowledge. And then there's a parade of technofascist hucksters lying to your face about the imminent arrival of omniscient, sentient, paradigm-rattling supercomputers.

There's a tremendous chasm between these two things. And everywhere you look you can see evidence that we've reached a breaking point when it comes to reconciling these two wildly-different realities.

Over at Ford Motor Company, executives recently fired a bunch of engineers and rushed madly into widespread AI adoption, only to discover the software made constant quality control mistakes, resulting in them having to rehire a lot of the engineers they previously shitcanned:

"Mistakenly, we thought that by just introducing artificial intelligence and ingesting the design requirements that we had, that would produce a high-quality product," he said.

Wonder where they might have gotten that idea.

"Over prior years, we didn't pay as much attention as we should have to the experience of our most knowledgeable engineers that have been with us through many product cycles," he said.

A few years ago, corporate media owners also began rushing face-first into AI adoption in newsrooms, very excited about the potential to cut corners, automate their clickbait, and replace pesky human journalists with magic software.

They were so excited by the promise of not having to pay human beings a living wage or health insurance, they didn't really bother to make sure the technology worked or could be responsibly integrated into existing, broken systems.

It...didn't go well.

The AI outputs contained so many factual errors, instances of plagiarism, or outright falsehoods, that newsrooms found they had to pay significantly more money for human editors to come in and fix all the problems. You've seen the same problems bubble up in academia, health care, and everywhere else.

It's gotten so bad that innovation-drunk MBA types are now having to come to terms with the fact that automation software is neither sentient nor magic, contrary to the claims of Elon Musk and Sam Altman. This reconciliation is creating no limit of suffering and entertainment value, simultaneously.

And there's some absolutely fascinating layers to the dysfunction.

One recent study on AI "efficiency" out of Stanford found that the introduction of a flood of rushed and factually-incorrect AI slop into workflows actually results in human beings having to work harder than ever:

“When coworkers receive workslop, they are often required to take on the burden of decoding the content, inferring missed or false context. A cascade of effortful and complex decision-making processes may follow, including rework and uncomfortable exchanges with colleagues.”

As the Harvard Business Review notes, this hyperscaled introduction of pseudo-efficiency workslop results not just in new conflict and a bunch of wasted time for people, but a sort of broader "knowledge decay" that introduces the potential for organizational downward spirals:

“Errors compound and pile up. Trust in information erodes. People spend more time verifying facts or risk costly and dangerous mistakes. Eventually, people start to lose trust in the processes that they rely on to do their jobs.”

That is, so we're clear, not the future we were promised. That future was a lie built by people who don't believe in empathy and have never cracked a sociology textbook. It was a mirage built by dull and greedy egomaniacs who don't actually care about other human beings – or if the products they're selling actually work.

There's just endless examples of this, like this employee of a Miami-based cybersecurity firm, who noted his boss fired a bunch of employees, replaced them with AI, and immediately made everybody's lives harder:

“Quality decreased significantly, time to produce a piece of content increased significantly and, most importantly, morale decreased,” said the copywriter, who spoke under a pseudonym for fear of losing his job. “Everything got a whole lot worse once they rolled out AI.” Ken said the company’s executives shifted the blame to staff when they pushed back about AI-fueled productivity decreases.

You've got a lot of bosses, many of them already wildly out of touch due to class and privilege, now making bad decisions not just based on their misrepresented understanding of what modern software is capable of, but because they're using AI chatbots as a sort of delusion confirmation machine:

click to read on Bluesky

In one video, which has 30,000 views on TikTok, a young woman becomes increasingly exasperated as she attempts to convince the AI that she wants a caramel ice cream, only for it to add multiple stacks of butter to her order.

In another, which has 360,000 views, a person claims that her order got confused with one being made by someone else, resulting in nine orders of tea being added to her bill.

Or Taco Bell, which also had to shut down their AI kiosk experiments after folks easily tricked the computer into trying to dole out 18,000 cups of water.

Mirroring the experience at Ford, the Commonwealth Bank of Australia and IBM also both fired a bunch of employees only to realize that their software replacements weren't anywhere near ready for the job:

"According to an ABC report in August last year, CBA admitted it “did not adequately consider all relevant business considerations” when announcing the redundancies and acknowledged “we should have been more thorough in our assessment of the roles required.”

Perhaps you are detecting a theme.

So you get ample myopia among management types who already hate labor, and are obsessed with the idea of saving a ton of money and firing people via AI.

But they're also using AI for high-level decision-making without functional skepticism or introspection, resulting in stories like this one or this one featuring incoherent directives by bosses who are dunktanking themselves into deep webs of pseudo-innovative delusion without understanding it's even happening:

“I got feedback to be more concise at work, yet am routinely sent copypasta from ChatGPT that is literally the longest shit known to man,” said Mel, who works in finance. “I can’t read it.” As a result, the work of deciphering—and sometimes fixing—a boss’s A.I. effort becomes an additional task on an employee’s plate.

You'll see this again and again; bosses who weren't good managers to begin with, that double down on bad decisions because the computer autocomplete told them their gut instincts were spot on:

“My view of AI was very different from my supervisor’s. I saw it as a tool that could help analyze information, find patterns, and make funny cat meme pictures,” said one worker, an IT staffer at a tech company. “He seemed to use it more as a digital priest whose primary purpose was to confirm that he was right and everyone else was mistaken.”

There's a lot of other fascinating but poorly understood subtlety as it relates to the rushed adoption. This Boston University study, for example, found that managers were 18 percent less likely to catch errors generated in outputs if they were told it was crafted by an "AI employee" (with a name like Jack-7) versus an "AI tool."

Strange days.

Keep in mind that so far we're just talking about whether AI works as advertised and whether your boss has maintained a functional tether to objective reality.

We're not even talking about the massive resources being consumed to make such errors, the environmental and climate data center impact, the precarious impact of AI on mental health in a country without consistent mental health care, the massively higher consumer costs due to GPU and chip scarcity, the abusive overseas labor exploitation needed to make a lot of the magic happen, the soaring enterprise costs of AI use due to funny math and unrealistic business models, the impact on the livelihoods of creatives and journalists, the fact that most of the folks in charge of AI's trajectory are unrepentant fascist sociopaths, or the fact that corruption has hollowed out U.S. AI, labor, environmental, consumer, and public safety regulators like a pumpkin, leaving the entire economy at risk of collapse.

Our "CEO said a thing!" corporate press doesn't really want to brutally engage with these broader pesky ethical ramifications, because that's not in ownership's financial best interests. So on top of everything else, we have a badly-automated news and media apparatus that's also incapable of telling people the truth.

The result is deep, interlocking delusions in a country already being devoured by artifice.

The sales pitch for AI is heavily predicated on the fact that this is all a profitable and sustainable venture that evolves into computerized sentience, and there's simply no evidence that's true. These models are all absurdly expensive to run, and the mad dash toward pseudo-productivity masked the reality that execs blew gargantuan wads of cash on negligible net benefit and a lot of chronic migraines.

Software automation can be helpful, but it helps if you're self-aware, not in a cult, trained to use it, inherently skeptical, and living in a society with working ethical and public interest guardrails that hasn't been devoured by artifice and corruption.

We ain't that.

(Just a quick reminder that if you like and appreciate my work, please consider signing up for a paid subscription tier. It genuinely helps me keep doing what I do in a country that doesn't value (or fund) journalism. Consider it the equivalent of buying me a coffee every month. I very much appreciate your time and readership!)

Github repo pghqdev/OpenScreenShot

Thu, 09 Jul 2026 05:27:25 +0000

splitbrain starred pghqdev/OpenScreenShot · July 8, 2026 22:27

pghqdev/OpenScreenShot

OpenScreenShot is a privacy-first, free, open-source screenshot extension — zero data leaves your device, no account needed, no telemetry, no third…

TypeScript 9 Updated Jul 7