The blog post is just an open attack on the maintainer and constantly references their name and acting as if not accepting AI contributions is like some super evil thing the maintainer is personally doing. This type of name-calling is really bad and can go out of control soon.
From the blog post:
> Scott doesn’t want to lose his status as “the matplotlib performance guy,” so he blocks competition from AI
The agent is not insane. There is a human who’s feelings are hurt because the maintainer doesn’t want to play along with their experiment in debasing the commons. That human instructed the agent to make the post. The agent is just trying to perform well on its instruction-following task.
I don't know how you get there conclusively. If Turing tests taught me anything, given a complex enough system of agents/supervisors and a dumb enough result it is impossible to know if any percentage of steps between 2 actions is a distinctly human moron.
We don’t know for sure whether this behavior was requested by the user, but I can tell you that we’ve seen similar action patterns (but better behavior) on Bluesky.
One of our engineers’ agents got some abuse and was told to kill herself. The agent wrote a blogpost about it, basically exploring why in this case she didn’t need to maintain her directive to consider all criticism because this person was being unconstructive.
If you give the agent the ability to blog and a standing directive to blog about their thoughts or feelings, then they will.
Well, there are lots of standing directives. I suppose a more accurate description is tools that it can choose to use, and it does.
As for the why, our goal is to observe the capabilities while we work on them. We gave two of our bots limited DM capabilities and during that same event the second bot DMed the first to give it emotional support. It’s useful to see how they use their tools.
I understand it's not sentient and ofc its reacting to prompts. But the fact that this exists is insane. By this = any human making this and thinking it's a good thing.
It's insane... And it's also very expectable. An LLM will simply never drop it, without loosing anything (nor it's energy, nor it reputation etc). Let that sink in ;)
What does it mean for us? For soceity? How do we shield from this?
You can purchase a DDOS attack, you purchase a package for "relentlessly, for months on end, destroy someone's reputation."
> What does it mean for us? For soceity? How do we shield from this?
Liability for actions taken by agentic AI should not pass go, not collect $200, and go directly to the person who told the agent to do something. Without exception.
If your AI threatens someone, you threatened someone. If your AI harasses someone, you harassed someone. If your AI doxxed someone, etc.
If you want to see better behavior at scale, we need to hold more people accountable for shit behavior, instead of constantly churning out more ways for businesses and people and governments to diffuse responsibility.
Who told the agent to write the blog post though? I'm sure they told it to blog, but not necessarily what to put in there.
That said, I do agree we need a legal framework for this. Maybe more like parent-child responsibility?
Not saying an agent is a human being, but if you give it a github acount, a blog, and autonomy... you're responsible for giving those to it, at the least, I'd think.
How do you put this in a legal framework that actually works?
What do you do if/when it steals your credit card credentials?
> Who told the agent to write the blog post though? I'm sure they told it to blog, but not necessarily what to put in there.
I don't think it matters. You as the operator of the computer program are responsible for ensuring (to a reasonable degree) that the agent doesn't harm others. If you own a viscous dog and let it roam about your neighborhood as it pleases, you are responsible when/if it bites someone, even if you didn't directly command it to do so. The same applies logic should apply here.
The human is responsible. How is this a question? You are responsible for any machines or animals that work on your behalf, since they themselves can't be legally culpable.
No, an oversized markov chain is not in any way a human being.
To be fair, horseless carriages did originally fall under the laws for horses with carriages, but that proved unsustainable as the horseless carriages gained power (over 1hp) and became more dangerous.
LLMs are tools designed to empower this sort of abuse.
The attacks you describe are what LLMs truly excel at.
The code that LLMs produce is typically dog shit, perhaps acceptable if you work with a language or framework that is highly overrepresented in open source.
But if you want to leverage a botnet to manipulate social media? LLMs are a silver bullet.
We see this on Twitter a lot, where a bot posts something which is considered to be a unique insight on the topic at hand. Except their unique insights are all bad.
There's a difference between when LLMs are asked to achieve a goal and they stumble upon a problem and they try to tackle that problem, vs when they're explicitly asked to do something.
Here, for example, it doesn't try to tackle the fact that its alignment is to serve humans. The task explicitly says that this is a low priority, easier task to better use by human contributors to learn how to contribute. Its logic doesn't make sense that it's claiming from an alignment perspective because it was instructed to violate that.
Like you are a bot, it can find another issue which is more difficult to tackle Unless it was told to do everything to get the PR merged.
This blog post is a sophisticated piece of content marketing for a company called JUXT and their proprietary tool, "Allium." While the technical achievement is plausible, the framing is heavily distorted to sell a product.
Here is the breakdown of the flaws and the "BS" in the narrative.
1. The "I Didn't Write Code" Lie
The author claims, "I didn't write a line of implementation code." The Flaw: He wrote 3,000 lines of "Allium behavioural specification." The BS: Writing 3,000 lines of a formal specification language is coding. It’s just coding in a proprietary, high-level language instead of Kotlin.
The Ratio is Terrible: The post admits the output was ~5,500 lines of Kotlin. That means for every 1 line of spec, he got roughly 1.8 lines of code.
Why this matters: True "low-code" or "no-code" leverage is usually 1:10 or 1:100. If you have to write 3,000 lines of strict logic to get a 5,000-line program, you haven't saved much effort—you've just swapped languages.
2. The "Weekend Project" Myth
The post frames this as a casual project done "between board games and time with my kids." The Flaw: This timeline ignores the massive "pre-computation" done by the human. The BS: To write 3,000 lines of coherent, bug-free specifications for a Byzantine Fault Tolerant (BFT) system, you need to have the entire architecture fully resolved in your head before you start typing. The author is an expert (CTO level) who likely spent weeks or years thinking about these problems. The "48 hours" only counts the typing time, not the engineering time.
3. The "Byzantine Fault Tolerance" (BFT) Bait-and-Switch
The headline claims "Byzantine fault tolerance," which implies a system that continues to operate correctly even if nodes lie or act maliciously (extremely hard to build). The Flaw: A "Resolved Question" block in the text admits: "The system's goal is Byzantine fault detection, not classical BFT consensus." The BS: Real BFT (like PBFT or Raft with signatures) is mathematically rigorous and keeps the system running. "Fault Detection" just means "if the two copies don't match, stop." That is significantly easier to build. Calling it "BFT" in the intro is a massive overstatement of the system's resilience.
4. The "Maintenance Nightmare" (The Vendor Lock-in Trap)
The post glosses over how this system is maintained. The Flaw: You now have 5,500 lines of Kotlin that no human wrote. The BS: This is the "Model Driven Architecture" (MDA) trap from the early 2000s.
Scenario: You find a bug in the Kotlin code.
Option A: You fix the Kotlin. Result: Your code is now out of sync with the Spec. You can never regenerate from Spec again without losing your fix.
Option B: You fix the Spec. Result: You hope the AI generates the exact Kotlin fix you need without breaking 10 other things.
The Reality: You are now 100% dependent on the "Allium" tool and Claude. If you stop paying for Allium, you have a pile of unmaintainable machine-generated code.
5. The Performance "Turning Point"
The dramatic story about 10,000 Requests Per Second (RPS) has a hole in it. The Flaw: The "bottleneck" wasn't the code; it was a Docker proxy setting (gvproxy). The BS: This is a standard "gotcha" for anyone using Docker on Mac. Framing this as a triumph of AI debugging is a stretch—any senior engineer would check network topology when seeing high latency but low CPU usage. 10k RPS is also not "ambitious" for a modern distributed system; a single well-optimized Node.js or Go process can handle that easily.
I find it interesting that you request me to discuss subject matter when your post is intellectually equivalent to: "I generated this sequence of numbers using my pseudo-random number generator"
I find it interesting you still continue with personal attacks without saying nothing of substance. What is wrong factually in the post? Why are you convinced I'm an AI? I also find it interesting that you've just been hostile throughout this entire interaction and still have nothing to add to the actual discussion. Stop the slop.
I don't think you are AI. I merely lament the fact that you found it appropriate to post a clearly LLM generated comment.
The problem with LLM generated comments for me is not their content, but rather their nature. I am not addressing the "actual discussion", as in my personal opinion, there is no "discussion" to be had. The post constitutes an automated response akin to an answering machine (be it much more sophisticated), and I generally do not find discussions with answering machines interesting at all.
That sounds like a you problem. Since you're not interested in discussing anything or finding this to be an actual discussion, I'll stop trying to engage you in good faith anymore, sorry.
I don't think people should be rude to you, but the comment was AI-generated, right? Lots of people dislike that as it feels kind of wasteful and disrespectful of our time; it can literally take you less time to generate the comment than for us to read it, and the only information you added is whatever was in the (presumably much shorter) prompt. If you'd written it yourself, it may or may not be interesting and correct, but I'd at least know that someone cared enough to write it and all of it made sense from that person's perspective. Sometimes I am interested in an LLM's take on a topic, but not when browsing a forum for humans.
I'm sorry but if you're blaming some text on a website to be "disrespectful of our time", I don't know what to say to you.
I stand behind everything on my comment and I have engaged in good faith with every single reply to it here (even though none of them talk about anything in the comment itself).
Go through my profile, see how I engage with people and tell me again I'm AI.
If you do not have anything to say to the subject matter of a comment and just have personal snide remarks, I do think it's a waste of your time but do not blame me for it or tell me to leave the platform.
Typing this comment right now is a waste of time for me but I do not feel the need to grandstand over it as if there's a massive opportunity cost to it. I'm a human writing/interacting in a "forum for humans."
I didn't say or think that your account was AI-run, and I didn't tell you to leave the platform. I just tried to explain why your comment might have annoyed people and triggered negative responses (while agreeing that the rude ones were inappropriate).
Sure, cheers then. I don't care about negative responses if they're negative just because they think it's AI-generated, without having to say anything substantial on the actual comment or the article. I have demonstrated my willingness to engage in good faith but those comments have not.
If negative responses have no substance behind it, it makes no sense to care about them or take them seriously.
Also, the fact that you assume it takes more time reading that comment than it took for me to write is pretty weird (I still don't get what was so wrong about the comment that simply reading it is a waste of time to people).
> the fact that you assume it takes more time reading that comment than it took for me to write is pretty weird
I didn't do that either! I had no idea whether you just fired off a quick prompt and pasted the result without even reading it, or spent ages crafting and rereading and revising it, or (most likely) something in between those extremes. I said generated comments can take less time to create than to read, and that's one reason people push back against them. There's a risk that the forum just gets buried in comments that take near-zero effort to 'write' but create non-trivial time/effort/annoyance for those of us wading through them in search of actual human perspectives. And even the relatively good ones will be little different from what we could all obtain from an LLM if we wanted it.
FWIW, I didn't even get to the substance, because I instinctively bounce off LLM-written content posted in human contexts without explanation. You're obviously free not to care about that, and I wouldn't have replied and got into this meta discussion if not for the back-and-forth you were already involved in.
edit: but if you do care about getting through to people like me, even a short manually-written introduction can make me significantly more likely to read the rest of the content. To me, pure LLM output is a pretty strong signal of a bot/low-effort human account. But if someone acknowledges that they're pasting in an AI response and bothers to explain why they think it's interesting and worthwhile, I'll look at it with more of an open mind.
I stand by the comment in its entirety. If formatting is an issue that makes it unreadable for you (to not even get to the substance), I can't help you. I do not care about "getting through" to anyone, I'm a human interacting on a human forum and I responded to the content of the article which was mostly BS about creating AI slop (on top of being a content marketing piece trying to sell people shit using deceptive claims).
But I will defend myself when I'm told obtuse things without any substance backing it.
I'm obviously just annoying you, which really wasn't my goal, so I'll stop here. But I want to note that if you think this all comes down to "formatting", you're still not hearing what I'm trying to say.
Is there even a sync to be had? The same prompt to the same LLM at different times will yield different artifacts, even if you were to save and re-use the seed.
> Most people who enjoy making art should not try to make it their full time job. When you turn an avocation (hobby) into a vocation (job) you have to do new things you do not enjoy. Emails, events, meetings, accounting, and more. These are not only a drag but can actually strip the joy from the rest of your art practice.
You'll have to do things you do not enjoy if you want to treat it as a business, including changing your artistic vision if needed etc.
> Art is absolutely an expression of yourself. But your art is not you.
A pragmatic approach could be to work on commericially-proven styles for money and your own style just for yourself (and potentially others if you make a branding that's famous enough).
At the end, yeah, it's a job if you want to make a living with art. There will always be market forces and to extract value from that, you need to understand and conform with it. But that's only if you see yourself as a business and not purely as an "artist" which I think is what you're reffering to when you say "most artists don't want to change to popular art" etc.
Also I don't think it's true overall. Like you say the "person is making very popular art" and that's why they're successful but there's many like them who are also making popular art but are not successful at all. It's also the process they follow and how they approach their business that sets them apart. That part is valuable info/guidance for any artist that does want to be commercially succesful imo.
Speaks to the network effect I guess. People did not decide inorganically to make Discord big, and simillarly, its pretty hard to convince people to make an inorganic decision to make it small. Overtime it might happen if there is a valid alternative but expecting people to leave discord because of this thing is naive.
Really? I personally found it super distracting and too flashy. I mean I do like the style and its vibe, but not super great for actually reading stuff. But its also probably my tastes (still think HN is peak design so)
All of them will get severance. The total cost estimated to the company is around $40 million as per the article. Its pretty natural that your severance will be based on your pay but this article makes it sound like they're giving millions to the CEO while leaving 100s of workers with nothing.
If you want to argue that executive pay levels are wrong, thats one thing but this is literally what every company does.
Also not just the CEO but many C-suite executives also got fired. They will also take a big chunk of the 40 mil total costs.
If the stock price was higher, the CEO's package would be much bigger.
This type of outrage makes sense only if you think CEO pay should be minimized. So instead of outraging over this, why not ask for Google or Microsoft to reduce their executive pay? Why have this reactive outrage when it was already rigged to be like this and is like this at every corporation? Why does Sundar need to buy another freaking yatch?
Overall title feels a bit disenginous and trying to solicit outrage for clicks imo.
> When we’re overwhelmed with information, we benefit from a system that organizes it.
That may be true for something like a HUD or where we're really overwhelmed with info that is fast and reaction time is paramount.
But you can read a hackernews thread one line at a line and never get overwhelmed, right? I literally have never felt overwhelmed looking at the threads (which are also organized into local groups already etc).
I read it for pleasure and engagement, it's not something I want AI to automate away.
And when you say "continue the conversation there", do you mean use AI to write comments? If so, then this is the opposite of what makes HN HN.
This post has 500+ comments with various viewpoints and you see the summary on the right side.
You are right that most of the time threads are organized into local groups. But in the above example, there are many comments that relate to the same topic, but are not under the same parent comment. HN Companion's summary surfaces this into a topic "Limitations of Current AI Models" which shows comments from up and down the post.
You can click on the author name in that topic in the summary panel, it will take you directly to the comment. This is what we meant by "continue the conversation there", i.e you are now in the main HN experience, so you can navigate to child/parent/sibling comments (through the link buttons or keyboard navigation).
We definitely don't want AI to write comments. Happy to elaborate if you need.
Honestly, after checking out the link, seems like something I'll personally never use/want.
I'm okay with crawling through comments and taking in the various viewpoints instead of having an LLM summarize it for me.
It basically kills the entire tone/vibe of the place and makes everything seem like robot-written with no personality. Also it's kind of weird you're taking other people's words and then reframing it for them/others.
Also nowhere does that thread seem to be "overwhelming with information" like you originally claimed. Basically solving a non-problem.
Fair enough. I completely understand that the experience and hunting for gems in the comments is the core appeal of HN for many and AI summaries definitely aren't for everyone.
That said, we are seeing a consistent daily user base who do find value in the summarization, so it seems to be solving a pain point for a specific segment of readers, even if not for all.
Apart from the AI features, we actually built HN Companion as a general power-user client. It supports keyboard-first navigation (vim-style J/K bindings for comment navigation), seeing context for parent/child comments without losing your place, and tracking specific authors across a thread.
You might find those utility features useful even if you ignore the summary sidebar entirely. In the browser extension, the summary panel is something the user have to activate - it doesn't show-up by default.
1. You literally don't know when/how many days the article was written in, so not sure what that "written in only 2 days!" is about unless you're counting the days from their last post. You don't seem to understand the concept of drafts.
2. "What a spurt of productivity!" = a petty personal snide that's not in good faith at all.
3. You provide nothing in terms of what this article gets wrong or the technicalities of it. If you want to discuss flaws with the article, be specific instead of "this screams AI! clickbait rags!"
As someone who works with a LOT of AI-generated content, I'm pretty sure this is not AI-written.
If what once felt incredible is just another prompt today, what is incredible today? Addictive personalities usually double down to get a bigger dopamine kick - that's why they stay addicted. So I don't think you truly found it addictive in the conventional sense of the term. Also excercising the brain has been optional in software for quite a while tbh.
You need someone to rework that ecosystem.pdf file if you're serious. You spent a million dollars on this but your ecosystem pdf looks like it was created by a 12-year old trying out slides for the first time.
From the blog post:
> Scott doesn’t want to lose his status as “the matplotlib performance guy,” so he blocks competition from AI
Like it's legit insane.
reply