HNHacker News
TopNewBestAskShowJobs

jstummbillig

6,608 karma · joined February 14, 2015

submissionscomments
jstummbillig··on Mistral's CEO: Europe has 2 years to stop becoming America's AI 'vassal state'
How do you divorce being a European principally from being willing to pay taxes? What part of being uniquely European do you value that does not fundamentally require taxes?
jstummbillig··on Mistral's CEO: Europe has 2 years to stop becoming America's AI 'vassal state'
It seems fairly plausible that they will become useful. I would not consider coding requiring vastly better AI models than the best we have today. The more interesting question is what advantages excess intelligence provides in the future more generally (warfare, system management, etc)
jstummbillig··on We've made the world too complicated
> I've realized this article is slightly naïve.

Well, I would discard substitute "slightly" for a much stronger adverb but alright, directionally: Good!

The uncomplicated past was horrible. The complications exist for a reason. The natural world is mostly fucking mental by any modern human standard.

Yes, we can do without some regulations and maybe whatever social media is. I don't know.

But we are quite smart at identifying things that are bad for us, over some span of time (due to us discussing them, case in point) and also making them better, working against or in spite of millions of years of evolution. Our lives are short and thus we are incredibly impatient with each other – as we should be! It keeps us on our toes and makes the now important. But we do sometimes forget that, if you are close enough to the now, not everything is a smooth curve and it's not always immediately clear which parts in this complexity do what.

But thing's really do get better. Yes, we do not understand all the systems but it's not like the natural world is a system we understood better in the past. We used to think dancing made rain. And yes, things are getting more dangerous as we are getting more powerful through complications. This is a separate, also complicated issue, because it only requires one rogue agent to do incredible harm anyway. Best I can tell: We need to all get better and keep each other in check.

Unless you are a sort of tourist (and that includes all forms of opt-in naturalness today), the natural world is by and large a complicated, hostile and horrible place by all human standards. We do things for good, selfish reasons and somehow, miraculously this works our fairly well.

jstummbillig··on OpenAI and Government of Malta partner to roll out ChatGPT Plus to all citizens
What is it with this new one line paragraph style?

EDIT: Needless to say I loath it and I don't know why.

jstummbillig··on Too dangerous or just too expensive? The real reason Anthropic is hiding Mythos
This makes it sound like some kind of open question or even mystery.

Amodei himself stated quite clearly in recent interviews that they simply can't satisfy all demand, compute wise. Of course, Mythos could get more of the already too small pie, but clearly it's a more resource intensive model and would further increase strain.

jstummbillig··on UK sovereign LLM inference
Hey Ben. I find communication like this fairly off-putting. In so far the 80% cheaper per token (or any part of it) is something of your own making/ingenuity, by all means, do tell, but it requires comparing token cost fairly with comparable models on i.e. OpenRouter and not across different models and pretending it's the same thing.
jstummbillig··on Googlebook
A company that mostly works inside the browser and/or Google workspace. ChromeOS makes sense and just decent devices here are rare.
jstummbillig··on Software engineering may no longer be a lifetime career
> Multiple times per week I have the same conversation.

Really? I mean, good on you if it's true and you like the attention but that's sounds like an implausible amount of interest in someone and their relatively mundane profession.

jstummbillig··on I'm going back to writing code by hand
Not at all. Code is not important, intent is. The leader of a product/company does not have to read code. It doesn't matter if it is generated by humans or non-humans. It simply needs to be correct enough to be usable and then steerable towards better outcomes. Understanding of code never existed from the business perspective.
jstummbillig··on I'm going back to writing code by hand
> The only people I've heard saying that generated code is fine are those who don't read it.

Well, that is problematic. I have to either assume you are disinterested or lying and neither is great for any discourse.

jstummbillig··on Teaching Claude Why
The categories make no sense. Not having to do a job is the entire best case of AI. What we do with that is another thing, but we simply have to accept that any other lense is complete nonsense. The endpoint is obvious and we need to stop being silly about it: We are replacing human labor. Maybe we will find some new jobs to do in the interim. Maybe not. In the end, if everything goes right (in the AI optimist sense), jobs will not be something that humans do.

Labor = capital/energy in an AI complete world. We have to start from that basis when we talk about alignment or anything else. The social issues that arise from the extinction of human labor are something we have to solve politically, that's not something any model company can do (or should be allowed to do).

jstummbillig··on Discord Incident – Resolved
I don't exactly know what you are comparing. No popular IRC network came anywhere near what we would find acceptable in terms of reliability today. It was an absolutely (wonderful) wildfire.
jstummbillig··on Child marriages plunged when girls stayed in school in Nigeria
Or, potentially, you have less time to marry (among other things) when you go to school?
jstummbillig··on Vibe coding and agentic engineering are getting closer than I'd like
Organizations usually are not looking for employees who change things that work fine, just because it disagrees with the "vision" of one employee.
jstummbillig··on Vibe coding and agentic engineering are getting closer than I'd like
That's not how most organizations work, AI or not.
jstummbillig··on Flue is a TypeScript framework for building the next generation of agents
That would be really interesting. I doubt it's the case, actually probably the opposite? The harnesses seem very happy to write extensive test suits, without me having to ask much.
jstummbillig··on Open Design: Use Your Coding Agent as a Design Engine
It's midness. People don't find it good, it's just less effort to meh it.
jstummbillig··on Your website is not for you
In a mom-and-pop shop you still want good electrical wiring.
jstummbillig··on ASML became the chokepoint for cutting-edge chips
> By betting on extreme ultraviolet lithography long before it worked, ASML became the chokepoint for cutting-edge chips.

Makes one wonder: Would we be much better off of worse off if we reshaped society to do more of things, where a new technology is unlikely to work but highly beneficial in the limits? Would we sooner have 10 additional ASMLs or waste a lot of resources?

jstummbillig··on Show HN: A Karpathy-style LLM wiki your agents maintain (Markdown and Git)
> My hypothesis is that “do things that don’t scale”[0] will still apply well into the future, but the “things that don’t scale” will change.

Say more?

jstummbillig··on Tell HN: Claude 4.7 is ignoring stop hooks
Simple: You can ask a LLM and can get a good explanation for why it did something, that will help you avoid bad behavior next time.

Is that reasoning? Does it know? I might care about those questions in another context but here I don't have to. It simply works (not all the time, but increasingly so with better models in my experience.)

jstummbillig··on GPT-5.5
> This quote is more sinister than I think was intended; it likely applies to all frontier coding models. As they get better, we quickly come to rely on them for coding. It's like playing a game on God Mode. Engineers become dependent; it's truly addictive.

What's the worst potential outcome, assuming that all models get better, more efficient and more abundant (which seems to be the current trend)? The goal of engineering has always been to build better things, not to make it harder.

jstummbillig··on GPT-5.5
The only difference being that we still do care, very much. The models can still get a lot better before we stop caring.
jstummbillig··on GPT-5.5
Tbf, I have not super kept track of what is actually happening inside the "thinking" portion of recent releases. But last time I checked there still was a lot of verbosity and mistakes, that beat the actual amount of required, usable code generation by a wide margin.
jstummbillig··on GPT-5.5
We'll find out!
jstummbillig··on GPT-5.5
You are paying per token, but what you care about is token efficiency. If token efficiency has improved by as much as they claim it did (i.e. you need less tokens to complete a task successfully) all seems well.
jstummbillig··on Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
> yet below than OpenAI/Anthropic

This is the competitive advantage. Being better.

jstummbillig··on ChatGPT Images 2.0
1) Is there a moat? Is there no moat? Are open models as good as the closed ones? I keep getting confused.

2) As one of these artists, I am entirely fine with my entire body of work being used for the purposes of model building. The tech is astonishing and fantastic, and I sincerely hope we will be better through it. As the parent suggested: The idea that people in general previously gave a fuck about compensating artists is hilarious. MS builds models with my work, random people bought, idk, another vacation in Thailand or a fourth pair of shoes with the money that they never spent on art. I know which one I would prefer.

But I do find it particularly juicy that people, who, on the whole, never thought too much about paying artists (which I am also fine with btw!), all of a sudden can't stop wringing their hands about the injustice of it all.

jstummbillig··on Anthropic says OpenClaw-style Claude CLI usage is allowed again
I have no trouble believing that all labs are trying really hard to come up with an enticing bundle of something works for a wide variety of users, but it's hard to anticipate the popularity of something like OpenClaw, which completely blows through all previous usage patterns at population level.

It seems like a tall order to set lasting rules in this space at this point, where nobody really understands what is going to happen in a few weeks.

jstummbillig··on MNT Reform is an open hardware laptop, designed and assembled in Germany
> You can make things at home.

Price aside, most just won't be very nice.

← PreviousPage 13 of 34Next →