AspireOne

I'll be honest, I have no idea what gopher or ring NIC is, lol.

Thank you!

Do you remember details about the rm -rf ~/ folder thing? Model, harness (omp?), context window size, reasoning effort (none/low VS anything higher is enough), when it happened?

Without those details, it's an interesting cautionary tale and a valid anecdote, but as actual practical datapoint, it's useless.

> What's preventing me from feeding this into Claude and reconstructing the company for $200?

Like, everything. For starters.

You deserve a medaile for this.

Man... This is so sad, disappointing and worrying. For some reason, your explanation in particular struck a nerve with me. Who could have predicted all the nuanced social consequences that AI would bring upon us. I don't yet know what position to take in my head on all of this. It's too much change, despite believing that the world will change and adapt. In particular, that human work will become more valuable - but as a craft. As an endeavor you do because it's artful, not because it's mandatory, or the most optimal way to quality or speed. And those two categories - human and ai work - will become very separate. While the economical value or even expectations of absolute quality in the human work will plummet.

Similarly, I expect "raw"-ness to gain on value. Photos untouched by AI or filters. Physical objects. The non-digital. Who knows, maybe physical media will make a comeback.

Either way, it's not good times we're living through, and I don't particularly appreciate all of this happening after *just* becoming an adult in the world.

As somebody who attempted that recently, I seriously doubt that percentage :D

I disagree with this completely.

I am glad about it as well. Don't get me wrong, I love OSS, I love personally being able to get my hands on the frontier stuff and frontier cybersec capabilities, and I dislike a few entities having exclusivity on it. Everybody does. But that absolutely doesn't mean it's the correct approach. I think people are blinded a little bit by general principles and by personal feelings, rather than operating from a point of pragmatism, consequences, and the big picture.

Also, to note: I don't claim to have the answers either. I don't even hold a strong position on this issue yet. I'm just sharing some of my intuitions, and reacting to people who, as far as I'm concerned, are making this issue all too simple.

The Pandora box can expand even further, and it matters. If it didn't matter, like you seem to be suggesting, then by the same logic there isn't a reason to push for further open-source development.

> Kneecapping cyber vulns only serve governments.

Uhh, no? Kneecapping cyber vulns server most of us. There are negatives to kneecapping it, but let's not pretend like it's perfectly okay and safe for everybody in the world to have a magic box that can hack anything and cause havoc all over the world. The more powerful, cheap, and accessible this magic box is, the worse the consequences.

This is a hard problem and there's real tension - like you said, the playing field not being leveled is a problem (major companies / entities / govts having exclusivity on the best cyber abilities etc), but everybody having access to it is a problem as well.

So let's not pretend like this is a simple problem, and don't say that it only serves the govt. That's very silly.

> One thing about these founders is that they also have a hard time understanding context and follow-up questions

I feel like you're purposefully being sarcastic.

I like it. If the increase in unrestricted cyber capabilities can be slowed down, please be it. The world needs to get ready, and that takes time. Way more time than even a slow progress would allow.

Unless you want to see more of what happened the past few days.

I think the argument that "cyber capabilities should be democratized, thus keep producing them now, fast, and for everybody" is a very naive one.

Furthermore, "UI design" is not a narrow enough category, I think. Sure, statistically significant differences in performance between models do manifest when benchmarking against "UI design", but there's still just so much variance within the category. Opus 5.5 can have excellent one-shot results, but totally shit the bed when it's placed in an existing codebase and should uphold the UI/UX conventions, design language, and existing component primitives already established in an app.

But that's just the obvious. There are task by task differences (component by component, page by page, context by context) that the model is worse and better at, even though they are all within the general category of UI design.

I honestly don't have the clarity of mind to explain this coherently, but the crux of my point is that there's a huge variation in it's performance in different contexts, often ones that seem near similar, and in result, even Sol can beat Opus 5.5 in some variations of UI design.

And this is my problem in benchmarking and comparing model performance in general. Presently, we try to create broad categories to benchmark against, create a set of tasks, and then average out the results. That works generally, but it's kind of misleading and it's partly what drives people to feel a disconnect between benchmarks numbers and real world usage on their specific tasks.

It's such a messy field that it's kind of maddening.

What "types" of companies do you mean exactly?

They pretty obviously aren't. Why do you even feel the need to say this? What's the point? I'm not speaking on the behalf of the original commenter obviously, but as a passer-by, I find this comment to be very rude and tone deaf considering how much time likely went into writing such a thoughtful and comprehensive comment.

OP says that a change in leadership for THE one existing linux kernel project is warranted. Your reply is "just fork it"? The Linux kernel? And what exactly is that going to be good for?

No, it's not a fair suggestion, because it's not a real suggestion at all. Over 2000 skilled maintainers work on the kernel, there's hierarchy, processes, culture, community, the linux kernel project both adapts to and shapes computing globally, companies are in talks with the linux foundation in regards to drivers etc, distros are tightly coupled and tested on it, there are north-star people like Linus... It's very much a big ship.

So firstly, yes, there IS something "standing in the way of anyone forking a kernel and building their own version" - for example that it's absolutely useless, unless you can do the job of 2000 maintainers, your "fork" will be the upstream base + 3 little patches on top and you'll spend a month getting it stable on your system.

Secondly, this isn't something you solve by saying "screw it, I'll do it myself" and abandoning the ship. Even if you did build your own linux kernel to your liking, this doesn't address the concern about the global linux kernel project that other 99.9% of the planet and distros and hardware uses.

This "don't like it? Fork it" argumentation fallacy has been present ever since open source became a thing, and it's quite frankly insufferable, but applying it to the goddamn Linux kernel, that's a new level.

> LLMs cannot reason, period.

You cannot be serious.

... Can you quote any existing definition of "reasoning" that LLMs don't fit?

There's no "every time" for a LLM. It's weights are frozen. You're asking it for the first time, again and again.

A human will also say 7 disproportionately often. Clearly, our output is statistically constrained and we don't actually think, right?

... Besides, how do you even make the jump from "LLM picks 7 when asked for a number between 1-10" to "llms don't think"?

Just... Nothing actually makes sense in what you're saying. You didn't define "think", you didn't actually make an argument, you didn't present evidence, there's just nothing but empty words.

> People at the big labs have been terrified of losing the chain of thought because they currently monitor it

Yes

> to prevent users from using their models in certain ways they consider unsafe or bad for their business.

No.

That's not the only reason why they're terrified of losing CoT, which makes your statement a mischaracterization - a lie.

They don't want to lose also because CoT is one of the key tools in alignment training.

> Personally I think chain of thought monitoring is a bad way to do this and obviously needs to be replaced by some way of analyzing the model's activations directly.

"Personally"? That's literally the consensus amongst, like, everybody. Nobody in the field thinks CoT is an optimal way of monitoring. But it's the best way we have now.

Come on.

For time reasons.

> Amodeo could drop it all right now, today and spend the rest of his life with his family, or nose deep in the research project he's the most passionate about. He's not doing that, instead he's spending the remaining irreplaceable moments of his life pursuing the accumulation of power.

Again you go with this simplistic and reductionist view. This is an either-or fallacy.

I don't want to pursue this discussion any further, sorry. No offense to you. It just makes me too emotionally invested for some reason.

What is even your point man?

You know what?

Why does it always have to be some ulterior motive? Every. single. time. Anything a high-position person does - make a statement, post a blog, start a company, make a business decision... -, the reaction of the keyboard warriors is that it must be motivated by money, or power, or some need for destruction, and that anything that leaves their mouth is a poisonous lie.

I think this is a terrible property of humans, this hyper-skepticism, and that it has nothing to do with logic, nor deduction, nor reality.

Dario Amodei is a real person who has a family, a life, a profession. He worked hard to graduate from biophysics, he worked on ai for decades. He's very much a nerd. In interviews, he's very transparent. he left openai because of safety concerns, and everything he does and all his messaging EVER since is motivated by AI safety. His ultimate motivations are probably very messy and broad, some morally worse and some morally better, just like in all other humans.

You are going to stand here and, without flinching, without a single bit of shame, reduce him and others to wanting to "order legions of people around, buy and sell politicians, flood ad markets during elections" and "that's what motivates these people"?

Sorry but where did you leave your sense of nuance? What allows you to say this so confidently? Where is any proof?

This is simply absurd. Absolutely absurd, reductionist, and childish.

Yes, that's the reason. Anthropic said as much in their statement (they said that there's a tension between writing human-understandable prose and prose optimal for other agents to read).

As a side note, I am kind of sad that basically all models nowadays are RLd to hell and back specifically for coding and deep reasoning, and as a result, they lack in creative writing, human-readable prose, and even general knowledge and abilities (for example, qwen 3.8 27b is a beast for coding, but reportedly falls off the cliff in anything else). Gemini 4 Argon seems to be an example of what you get when you DON'T do that - Google seems to be one of the only labs optimizing for general performance.

I also fear that models cannot be excellent in everything at the same time. I fear that there will always be a heavy tradeoff between what specific field the labs optimize for VS everything else. I don't know.

To be honest, Gemini CLI was disastrous. The Antigravity CLI is much better.

"Disastrous" is obviously a loaded term, and it can be debated, but from what I collect, the reasons are

- Dots. People don't understand it. Their presentation didn't explain it properly at all. Additionally, many people thought it's stupid, and that the cutesy mascots are dystopian.

- Voice mode didn't work in their presentation. The demo failed.

- 200$ plan goes 20x -> 10x usage. 500$ plan introduced, 25x usage. The contrast is stark.

- "Ultrafast": 8x pricing on subscription plan (6x API). That's 80$ input, 400$ output, 100$ cache write. Hello?! Additionally, since it generates 8x faster, it consumes your quota 8x faster. Per wall-clock minute, it's 64x more usage spend per minute of generating. People report that on the 500$ plan, their weekly usage is gone in 40-100 minutes.

- People believe gpt-6.1-sol was planned to be released as astra-minor, but OpenAI renamed it at the last possible time, because gpt-6-sol is a disaster (this I can get behind)

Very easy to do. It's a nice silly thing to try, so I'll conduct this very important experiment and update you on the results if I remember to do so

One or the other. Anything in between is where economic suffering lies.

Then do. Enlighten us and save the ad hominems.