HNHacker News
TopNewBestAskShowJobs

jstummbillig

6,610 karma · joined February 14, 2015

submissionscomments
jstummbillig··on GPT-6 Sol and Luna
They hinted this release last week for tuesday already, so if anything it would be Anthropic that tried to make this happen. But I doubt it.
jstummbillig··on GPT-6 Sol
6 Luna and Sol just appeared in my codex. No 6 Terra!
jstummbillig··on I don't want to read what you didn't write
How do you explain professional speech writing?
jstummbillig··on Study: Young users (9 to 18Y) ditch Google for AI, with unknown consequences
> "with unknown consequences"

Beautiful. I shall consider suffixing every random observation I have or action I take with that, for an extra infusion of ominosity and importance.

jstummbillig··on Grok 4.7
How many parameters do Astra or Fable have?
jstummbillig··on I am often wrong
> I can't say I agree with forcing your team to use your own personal "framework" of approaching a problem

Is that not prt of what leadership always is? It might be less explicit. People often don't write their process down. Things get more confusing because people don't understand what is being asked of them.

But when is there ever no process that has to be followed? What leader lets people do whatever however and what's the role of that leader?

jstummbillig··on OpenAI's Sam Altman to Brief UN Security Council Next Week
HN would be better if it did not produce disparaging comments in every thread that has Sam Altman on it.

I get it. We dislike Sam. I got it 3? 4? years ago. Would be nice and more interesting if we could see past that at some point, given that he is a fairly important figure, doing important things, that would be cool to share interesting thoughts about.

jstummbillig··on Don't Be Nice
> 'unaware.' [...] I think that is a little bit of what it is to be nice.

Me, not being nice: Don't use language to be priggish and confuse. We got enough of that going already.

jstummbillig··on AI-generated posters don’t have to be horrible
Something, not anything. 2 years ago AI tools were much narrower and there were more things that you could not have done, but still plenty that was possible.

(Granted, given that you always needed some way of steering the AI and a harness around it, defining "entirely" in this context is a separate, debatable exercise)

jstummbillig··on AI-generated posters don’t have to be horrible
Ach, idk. On average people have horrible design aesthetics, yet are now so arrogant about it, because they imagine their a newly found ability to spot design generated with AI equates to them being able to spot bad design.

No, you silly geese. What you are able to spot is somebody using AI poorly. That's not you developing taste. It's the equivalent of being able to tell something is photoshopped because somebody used a default brush poorly, or masked something poorly and then confounding using Photoshop with using Photoshop poorly.

You have absolutely no way of telling when a good designer created something entirely using AI (unless they are fine with you knowing that) and that has always been true.

jstummbillig··on Xiaomi Mimo 2.6 live post-training dashboard
Wow, spending money on training an almost-frontier-model is much more time intensive than I thought it was.
jstummbillig··on GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code Review?
That very much depends on how code will be written in the future, how much of it and how often it changes. If more of it will be ephemeral (kind of what agents are already doing for all sorts of tasks right now) finding ways to very cheaply check might be of high value.

(I suspect this won't be it, though. Probably something the model providers are going to bake into the models themselves.)

jstummbillig··on Blizzard Revives 'StarCraft' Franchise with Open-World Sci-Fi Shooter
While I can feel the pain, the truth is somewhat less outrageous and more somber: Customers for a Blizzard scale RTS simply do not exist.
jstummbillig··on David Sacks: OpenAI and Anthropic Don't Need Regulations to Pace Frontier Models
Like it! Yeah, they can just take action and pressing them on that is fine. It's not like they asked for policy before they started to mine all the data and sell the resulting product, while, surely, at some point, there must have been the thought that the entire enterprise could be considered vaguely problematic.

To be fair

1) I feel like Sacks is kind of misrepresenting what is going on – at least according to the stories that the two labs tell, they are already taking it upon themselves to act.

2) Understanding if they should do it is more complicated; I don't think anyone can satisfactorily answer that one, so, as a matter of judgement, it seems like a choice to make and they should just consider this option.

At the same time, I don't think that requires the labs to not press for policy changes. Again: Claiming there is an obvious way to do that, that is both realistic and correct seems a little far fetched, but arguably one of the more important and pressing questions of our times to get a hold of.

jstummbillig··on Nvidia dismisses "circular financing", says every $1 it invests brings back $100
> Where are those 99$ coming from? Who is actually paying this money?

Companies who pay 99$ to make >99$ in return. I am not saying it works in all cases but that's the idea when a company spends money.

> Because it seems like so far everybody is losing money with no reversal of this trend in sight

I am not sure what you are seeing: Anthropic (as one of only two major companies that do just AI) is starting to return profits, while demand for AI is accelerating and, clearly, compute is maxed out. And I mean: On the entire planet. They are turning profits despite everything being in full buildout mode.

jstummbillig··on Real-SWE: Benchmarking AI models on private, real-world, enterprise codebases
I am trying to estimate if my reaction to seeing GPT-5.6 Sol last on that list is reasonable or or mostly emotional and find that I have no way of telling.
jstummbillig··on Shopify is moving from React Native back to Swift and Kotlin
In a sane society we would be less willing to pass heavy judgement on so many things at passing glance.
jstummbillig··on Ask HN: Can we please limit the AI news flood?
It's unlikely that right now is peak anything AI.
jstummbillig··on DeepSeek v4.1 Flash
Well yes, it is propaganda. They really think that.

I think it would be foolish not to debate it. I remember a time in my life where the majority of people around me found the idea of farm animals being capable of fear or pain laughable, while having no trouble thinking of dogs that way. Humans are dangerously incompetent beings. Being more careful is fine.

jstummbillig··on DeepSeek v4.1 Flash
If you are distilling from other models (according to Anthropic reports they are [1]), there are probably a bunch of things that you can just do away with.

[1] https://www.anthropic.com/news/detecting-and-preventing-dist...

jstummbillig··on DaVinci Resolve 21.1
For anything that can create better margins, and that is not outlawed, there will be demand.
jstummbillig··on Tell HN: OpenAI brings back 5 hour limit for plus and business standard users
The rules are not changing by chance. People want to use the new model. Rules are adjusted to better meet demand.

People want things to be better more than they want things to not change. That's a good thing!

jstummbillig··on Tell HN: OpenAI brings back 5 hour limit for plus and business standard users
The equivalent alternative is not "no hourly limits", it's simply less overall weekly allowance for Plus users, or higher prices overall. It's a way to balance a scarce resource, and maybe the least frustrating one?

People have a fair chance to learn resource management in 5 hour chunks, while not being limited to silly models, and burning through all tokens in the first hour of the week. Seems mostly good.

jstummbillig··on It's time for Mark Zuckerberg to resign from Meta
WELL THEN, if Joan Donovan thinks so.
jstummbillig··on The Real Luxuries In Life
Statistically, yes. And statical health things are the only health things to really worry about (because you can't ever do things to avoid getting cancer, you can only ever do things to decrease your chance of getting cancer).

But then again, the list of things that statistically keeps you healthy while being expensive is actually very short (GLP-1s currently being the most notable item on it).

Supposedly this is going to change in the near to mid future with more wonder drugs incoming. But, for the time being, that's nice!

jstummbillig··on Ask HN: Why were OpenAI, Claude, and Grok simultaneously down?
Only if you ignore baseline rates. Which brings us back to my original point.
jstummbillig··on GPT-6 Astra in code review: Gains, privacy, and cost
Roughly how we price (high skilled) human labor.
jstummbillig··on Ask HN: Why were OpenAI, Claude, and Grok simultaneously down?
Why would that be probable? People on average are fantastically bad at getting probabilities right.
jstummbillig··on Invisible Companies
As always the writing could be denser, but that is very lossy compression. I enjoyed the full read!
jstummbillig··on The Post-AI Internet Doesn't Look Great
I have no idea how one would have to slice their data to conclude that GPT 5.6 is not a significantly better writer than GPT 3.
← PreviousPage 5 of 34Next →