titularcomment

I think an important feature for homelabs is NAT traversal

Also seems to consist of directly AI adjacent topics from a quick skim

You can't decrypt wallets with LLMs friend.

I really disagree with their acceptable usage policy, but you can't say it with certainity that LLM contributions would be a fix for this. They generate so much noise (as the primary incentive behind an independent LLM scan is often cheap contributor brownie points for your CV) that it also could lead to bogus PRs being approved or helpful PRs being lost in the noise.

What is this 'mind' you speak of? As everyone else is intellectually lazy, how do you define the transformer architecture under the hood of LLMs?

I recommend you read the authors' other two latest articles.

Funnily enough you really do need a great prompting and SKILLS setup to use antigravity effectively in contrast to other providers which actually started benefiting from less detailed prompts over time. But I like it this way, its more customizable and much cheaper especially with a sub.

`agy --dangerously-skip-permissions`

FYI, there are ungoogled chromium builds for Android. Firefox Mobile really is a lackluster browser unfortunately both from a usability and security standpoint (e.g. IonStack worked on Fennec)

This is standart, and happens constantly with Invidious (youtube frontend). This happened before on AuroraOSS too. They probably just flagged the accounts and no API change or A/B testing an API change.

We miscommunicated. First, I think LLMs are a problem for newcomers in general, there is a lot of intuition that I saw that you build while tinkering at 3 A.M which makes programming easier in the long run, at least for me. For someone who invested enough time on that, such as you, can easily spot where LLMs fail and already have the knowledge to make the cogs turn. I clarified where I think LLMs are harmful at a sibling comment. This is where I think LLMs obstruct learning. 'Domain knowledge' is an important keystone for agentic coding that bypass this.

As for 'looking cool', I meant the code with its fancy formats and highlighting, no snark intended :)

LLMs are a great tool for building, especially for those who could conceivably be engineer team leads/project managers. For greenfield learning, its sub-par unless its guided well.

I like that they supplied the prompt instead of the working code :)

Reverse proxies work because they're not behind restrictive NATs, and having a few central servers that are not behind Symmetric NAT is a surefire way to establish an ISP-granted monopoly. Even if that somehow worked, this completely breaks P2P for consumers (which I think you intend) as well as torrenting without, again, a central NAT-punching TURN-like server. So yeah, address exhaustion is done but so is the Internet as we know it.

This came out a bit longer than I expected.

Depends on your usage. A rubber duck won't upheave the entire history of written human knowledge to reach for your answer. It doesn't have inherent bias, and its purpose could also be handled by a wall, or the air infront of you if you can stretch your mind a little. What I'm going for here is that as you said, its just a hack to make your brain look for answers in a more efficient manner, by clearly underlining your thought process out loud, tricking yourself into conversation mode, yada yada. Its still you doing all the work, you're just being smarter about it.

A LLM is more akin to a senior pair programmer. It may seem like a rubber duck on the surface, in the sense that you're having a conversation, but the senior programmer may come up with ideas you wouldn't think of. They may have more innate experience in the domain which might pull you away to very different implementations from what you had in mind initially. This is all good, but a common pitfall with LLMs is that they're much more conformative than a senior engineer, much faster, much more responsive. This is where it gets dangerous: they don't tell you you're wrong 80% of the time, like a rubebr duck. They also have innate experience of a domain, like a senior engineer. The result is that your brain never gets a 'hang' signal, and it doesn't feel the need to reevaluate.

The original article talks about 'savvy', which is to me an essence of learning. If 'exploration' is looking at walls of code you don't understand but the linter makes them look cool, you don't need LLMs for that.

Yes, using analogies of comparing LLMs to relatively simple tools and on the same plane declaring them the predecessors to AGI (or for some, we have achieved AGI) is a common hypocritical take.

With all due respect, what you're describing doesn't quite spell learning in my opinion. Sure, its better to read the code of an application you envisioned to learn a programming language rather than an unrelated convulated program; but you're still offloading your thought process to the LLM. You're not quite seeing why it chose A instead of B, and why it uses that struct instead of the other. And honestly, the LLM doesn't quite know why it chose that either.

Funny thing is, I'm glad we can read this output. I'm afraid advances in models will result in more censored reasoning for the models' benefit as well as anti-distillation efforts.

That last sentence is a straight-up threat.

If you think HN is anti-AI, you are in deep. There isn't many places on the internet (sans X/Twit which i don't check) that is as AI-positive as HN. Sure, holdouts remain but the frontpage has a 60% constant vibecoded project submission.

Is ARM truly more user-friendly in this regard? And I dont see no genuine alternative than arm64

Probably to avoid manipulation, there is tons of stuff on HN that hopes to make it to the limelight through this forum channel

This is complete dependency on LLM tools all the way from development to usage, and is risky as well as prone to failure

This does anything but put the user in control. I also think that arbitrary new features on established platforms are not in any close future as it already burns many tokens for you to instruct your agent to fork something and add x, y, z.

lol. Simple case of if you have to ask, you probably shouldn't.

1. It depends on model and tokens spent 2. Models talk the talk but not always walk the walk

Don't play fair in a game others cheat.

I'd respectfully advise you to assume less

GPU driver support is not limited to CUDA. Devices that require binary blobs of firmware to work properly are not always exotic. Some examples off the top of my head that won't work with OpenBSD hassle-free is Vantage (LLL), Solaar ... but I haven't tested. I understand you finding it easier to work with and it's faster install speeds do pique my interest, yet you being on HN already puts your technical literacy on some low % of gen pop. OpenBSD may be comparable to distros like Gentoo, Void or Arch but certainly not out-of-the-box ones like Ubuntu. They serve different auidences.