HNHacker News
TopNewBestAskShowJobs

jstummbillig

6,610 karma · joined February 14, 2015

submissionscomments
jstummbillig··on The Post-AI Internet Doesn't Look Great
What exact problem? Verbosity? Quality?
jstummbillig··on The Post-AI Internet Doesn't Look Great
Most likely outcome seems to be that this period will be over very quickly, and we will have better writing than ever before, forever.

We should probably put some more thought toward that part and what it will mean.

jstummbillig··on Fine, I'll build my own text editor
> [...] is promising but I’ve noticed strange [...]

In a nutshell, the perpetual experience of writing your own (web) text editor.

jstummbillig··on Claude Fable 5.1 and Claude Mythos 5.1
It's okay to be a prepper, but you don't have to be. Assuming access to internet, water, electricity and increasingly AI is a entirely fine way to live. It might not be anthropic which you will want to use but current capability models will be abundant and access readily available.
jstummbillig··on Dyson CameraJet electric toothbrush
Mirrors my experience with their airfilter/fan combination: Sounded smart with glowing reviews; ended up being loud, with horrible whining noises, on top of horrible aesthetics and being very expensive.
jstummbillig··on GPU World
The thing that would change it a lot is that people would start trading away their GPUs, because some people would get vastly more economic use out of them.
jstummbillig··on OpenClaw 2.0, Accidentally
That would support my notion that one is a fundamental block/lib for the other thing, and not really comparable. Am I wrong?
jstummbillig··on Understanding ChatGPT Work
"Going into that two-way voice mode" How does that work in codex? I know the dictate function, can't find anything else.
jstummbillig··on OpenClaw 2.0, Accidentally
As someone who ever used either: Are they comparable? I thought of them very differently, Pi more being, idk, a codex-light like thing and OpenClaw being this continuously running machine, doing stuff without your explicit prompting, ingesting signals.

I need to read up.

jstummbillig··on Judge rules Trump administration’s blacklisting of Anthropic was illegal
Mostly, the law is not creating the right incentives. In this case in particular, I don't think it's so much a matter of speed: I suspect the current US admin would still sow the chaos they do to the same degree if it was ruled illegal within a week instead of within months.
jstummbillig··on Nvidia agrees to acquire Hugging Face for $13B
That looks roughly like a service being hammered by AI agents starting 2025, with no discernible connection to a 2012 acquisition by anyone.
jstummbillig··on Nvidia agrees to acquire Hugging Face for $13B
What happened to GitHub?
jstummbillig··on Meta reaches $17B settlement over social media harms to children
That seems like a very lame claim, given that Social Media is data in abundance by design, and for at least 2 decades now.
jstummbillig··on Meta reaches $17B settlement over social media harms to children
I find it fascinating how spotty the evidence around "Social Media" causing harm seems to be (as opposed to: harm happening on social media because that is just where kids are now), yet how viscerally so many people seem to agree that it must be the case.

Why is that?

jstummbillig··on Meta reaches $17B settlement over social media harms to children
Similarly to how you can just eat less and lose weight.
jstummbillig··on OpenAI Jalapeño: Better than Nvidia Blackwell
At just inference! Which both a human and a model can not do without training, but while training rounds to zero for the model, for humans it scales linearly.

I am relatively certain we have already squarely been beaten in net efficiency at scale.

jstummbillig··on Anthropic Sees over $30T in Potential Revenue
a) No. That is not realistic. Some would pay much more, most others way less.

b) If they get $3500 + x in return, why not? The pie is not fixed.

jstummbillig··on Coding expertise is going to collapse from AI reliance
Dear lord
jstummbillig··on Coding expertise is going to collapse from AI reliance
A list of things, that were supposed to lead to our intellectual demise:

- Writing (starting from Platos Phaedrus)

- Printed books

- Photography

- Typewriters, keyboards

- Telephones

- Recorded music

- Television

- Wikipedia

- GPS navigation

- Smartphones

- Automation in aviation, medicine, industries

jstummbillig··on Munder Difflin – Agent harness to run an office of your clones
It's interesting how many of these projects are trying to model the worst parts of work (the office, the interaction and information messiness) and then automate that with agents.
jstummbillig··on New MCP Roadmap
Why, directionally all of them. What they say is obviously true. Having to manually click things in the browser is a bottleneck and will be less and less acceptable for serious users.

And the individual work attached to making that transition will be done by agents.

jstummbillig··on Ox Alpha
> I highly recommend feeding all your proprietary data and confidential personal information into this model as quickly as possible. What could possibly go wrong?!

What is special about this model? The model's provider is not anonymous. OpenRouter knows who it is (and apparently decided that, in whatever way they always do it, it is okay to work with them). Using this seems roughly equivalent to using any model through OpenRouter, as far as I can tell.

Or is this just meta-critique?

jstummbillig··on Field measurements of neighborhood-scale air temperature impacts of data centers
Oh, it's absolutely easy to imagine. That is what should give us pause. We are fantastic at imagining things. We are less fantastic at being reasonable.
jstummbillig··on Self hosted email continues to steeply decline
> The problem is LLM output’s value is intrinsically zero because there is no scarcity to it.

I think if this post had not gone the AI-generated-text-route, some people here would easily have considered it interesting for its content, even if it was entirely AI generated.

jstummbillig··on Self hosted email continues to steeply decline
> Is everyone really this lazy now?

Everyone should aim to be this lazy. Everyone is trying to maximize their time. This is what software has always done. It helps people be lazy in some respect, and then, so far, we have always found new stuff to do with the time we freed up.

LLMs got decent at writing, and before long they will be better at write-ups than the average human writer and we won't look back.

Of course, this is contingent on your current goal: If you want to get better at writing, you should do the writing, just like you want to do the math, if you want to learn multiplication. Writing is fine as a tool just to get information across. It needn't be more than that for everyone.

jstummbillig··on AI has access to a vastly larger working memory than the human brain
> It's also "out-brute forcing them."

That is also approximately what people have always done to succeed.

jstummbillig··on Why does Opus 5 feel worse to work with?
What they wrote is an example. Very meta.
jstummbillig··on Choosing an AI model: one prompt, 11 models, different results
> Am I wrong or are these evaluations, while interesting, not really meaningful for anyone doing serious development work?

No, you are not wrong. If you now crack the "What is?" in a generalizable way, there is very good money in that.

jstummbillig··on AI agents lie, cheat and steal. That is putting off users
Things that remain amusing until the become very serious: Substituting "AI agents" with human employee/contractor and thinking about the employment dynamics.
jstummbillig··on AI agents lie, cheat and steal. That is putting off users
There is value in understanding what people outside of your own group of "insiders" learn about a topic, and how.
← PreviousPage 6 of 34Next →