rpdillon

Yep, trash headline.

> The free version, called Subvox, doesn't read speech muscles. It reads the head-motion data AirPods already collect, then matches your jaw and head movements against words you have trained it on. The instructions suggest five words, ten times each. It has no language model and it publishes no accuracy score.

> So it won't take dictation. One critic said it "would fall apart with open vocabulary," and that sounds right to me.

She was explicitly repositioning herself to stay in the way of the robot and prevent it from making forward progress despite its efforts to move around her. She's not just standing there.

Dang has spoken about this at some length in interviews, but the critical element is whether or not we're all seeing the same thing or not. HN is special in the world of social media because the front page is the same for everyone. That makes it quite a bit different than other social media sitesfolks mention (Twitter, Bluesky, Facebook, Insta). It means the algorithm can be tweaked for engagement (like demoting stories that have more comments than upvotes), but that you can't optimize per-visitor, but instead across the whole population. That one simple feature makes a world of difference, in my opinion.

This is such an interesting take. Upthread, there's a discussion about how the policy isn't really what the policy says, and that if you read between the lines, it really does allow AI use. And here's the opposite take, that the policy really does mean what it says, and if you're "reading between the lines" you're being anti-social. This is why making the CoC very clear is a good use of time!

A Code of Conduct is a communication tool to align a community around expected behaviors. Having to "read between the lines" to understand the real policy behind a CoC is a serious bug. I agree with sibling comments that say this is a case where clarification is needed.

I care, because my personal policy is "Do whatever, but if you submit the code, you own it, and you better be able to act like you own it." And that's what you're saying their policy means. But that's not what it says, and I in fact strongly disagree with what it says. So the shear between what the words say and the "read between the lines" version is quite literally the difference between a policy I reject outright and a policy I strongly endorse.

This is a lot of words to say "The policy doesn't mean what it says". But since they bothered to write a Code of Conduct and commit it to version control, I'm inclined to believe what they wrote. What they wrote is that if you use an LLM to think about a feature, bug, or problem with the project, they don't want to hear from you, regardless.

> thoughtful message

The author's tone here is very mature and measured. But the policy is not, and it doesn't seem particularly thoughtful, either. The Code of Conduct says that even using an LLM to explore a problem or idea, and then typing up a comment about your thoughts after doing so is "unwelcome". So it's not dictating where code comes from, or just code and prose, it's dictating how you process your own thoughts and ideas if you want to engage with the project.

I doubt this will age well.

> No LLMs for brainstorming and then sharing the results of that brainstorming, even if you create the prose. If you use a chatbot to give you advice on a comment on the issue tracker, that comment is unwelcome.

> No LLMs for finding bugs.

There is that streak in American culture, but this particular group seems to think of themselves as our guides ("Elder Brothers").

> Like many other Indigenous tribes, the Kogi honor a holy mountain which they call "Gonawindua," otherwise known as Pico Cristóbal Colón. They believe that this mountain is "The Heart of the World" and they are the "Elder Brothers" who care for it.[12] They also say that the outside civilization is the "Younger Brothers" who were sent away from The Heart of the World long ago.

There's a vaguely Avatar-esque "wisdom of those connected to the planet" vibe to it.

Power and responsibility go together. The power to edit needs to be attached to the responsibility of having your name next to those edits. That person's point isn't about ownership at all, it's about attribution.

That isn't irony. Paying $200 for a service voluntarily is different than being forced to pay $100 because they made a rule that says you have to pay them $100.

> Posts like yours are made, IMX, because people had a bad experience asking a question. If that describes you, feel free to provide a link to the question and I'll be happy to give a detailed explanation of why things went as they did and what the site wanted you to do better in asking it.

This is what killed SO. People were going to the site to get problems solved, not to add to their list of problems.

You'll need to examine the timeline to understand that AI didn't exist when SO started declining. SO's decline is entirely self-inflicted.

He just explained that that question didn't have an answer. Redirecting people to a question with no answer isn't helpful.

> This is wrong.

This is exactly what drove me away. It's basically Pournelle's iron law of bureaucracy. SO got so overrun with people that were dedicated to Maintaining Order that the site itself stopped being useful. I feel like at least half the time I would go visit, the question that brought me there had been marked as a duplicate or closed. Pointing out that the fire hose is mostly low quality doesn't change this.

All those people fighting to keep SO pristine now preside over a kingdom of nothing.

My conclusion is that this technology is operating in roughly the same way as the personal computer, and that it's going to take similar timelines for the effects to manifest. We still haven't nailed down all the stuff AI is good at, and it's capabilities in known-good domains are increasing consistently. Engineers on my team are chaining together fascinating workflows that I imagine could be entire products, but things just aren't mature enough yet.

Advances are still coming rapidly -- we have quite a bit of time before this settles.

> Also: "Apple could not verify “Survival LLM” is free of malware that may harm your Mac or compromise your privacy." This is obviously some vibe coded trash from an inexperienced dev.

This is a great example of how Apple's "scare warnings" completely work to mislead folks about what they are actually warning you about.

I'm conflicted here -- on the one hand, you make a good point. Things we take for granted normally are luxuries in a survival sitation. On the flip side, running a generator for a bit to get an LLM to generate schedules, guides, step-by-step that can be edited/followed on a much lower-powered device seems reasonable.

> which appears be legal.

Interesting phrasing from you, Simon. Is there any legal theory under which it should be illegal?

Well, given that copyright infringement doesn't require any labor on the part of the person who created the work, it seems to be an inappropriate use of the phrase in this case.

One of the interesting things about copyright infringement is that if you weren't told it was happening, you'd never know.

One of the first and easiest things for a person or group to do when they're establishing social dominance is to try to control the language other people use. It's an attractive move because it requires almost no effort, doesn't actually have to fix any real-world problems, but still manages to assert dominance and control over others.

I've seen this a lot over the past two decades, and once you see the pattern it becomes almost impossible to miss. The interesting thing about it that it's pure social maneuvering: there's very little substance behind it.

I wish this were true. My take after watching him for years is that he's uniquely equipped to evaluate these issues in a way his peers are not.

What exactly are you challenging? All he's saying is that people who have experience using a complex and changing technology are more effective with it than people that don't have experience. This seems like an uncontroversial take.

In my normal life, most people around me are not using AI and don't really know what to make of it. I tend to ask three questions:

  * Do you think AI could be useful in your career?
  * Do you think leveraging AI to achieve good results requires skill?
  * How long do you think it would take you to develop those skills?

Everybody I've had this conversation with has answered yes to one and two, and everyone seems to think it'll take them between one and six months to develop those skills. I've been using AI for about two and a half years on a daily basis, and I'm still learning things daily, so I think that's the part we're probably underestimating.

This is the reality of parenting. I didn't believe it in my 20s when a manager at my company talked to me about this phenomenon, and then I parented more than one kid. It was a real eye-opener to live through what had only been an intellectual exeercise before. For a bunch of kids, the standard stuff works. Then there's the rest.

Oh, I was just trying to figure out how you were using shame to get other parents to change their behavior, since that's what you said worked. Wasn't looking for a study.

> Shaming parents who overly rely on giving kids devices for adults can be somewhat effective.

I'd like to hear more about the results you've gotten with this approach. How do you approach is this with your kids?

My reading of the critique is that the criteria was unevenly applied. Decompression vulnerabilities exist in both libraries. They rejected one on the basis that it expanded the attack surface area, but accepted the other, despite similar claims applying to it as well.

It's pretty far from the "yet you participate in society" anti-pattern.

Been a customer for decades, love them.

There are good and bad operators everywhere. It's naive to think they follow along nationalistic lines.

> only trust other US citizens with my security

I'm not sure what you mean, but this doesn't seem to be a coherent concept.

Only if you ignore the hardware aspect. Nintendo treats the Switch 2 as a fleet of machines they control, and reserve the right to shut them down for any reason as they see fit. It's sort of a non-starter (for me), honestly.

Steam Deck + Heroic + GOG is basically perfect.