fleebee

I'm running Iocaine[1] which is essentially the same thing on my tiny $3/mo VPS and it's handling crawlers bombarding the honeypot with ~12 requests per second just fine. It's using about 30 MB of RAM.

[1]: https://iocaine.madhouse-project.org/

Nice stats!

I've only had GPTBot reach depth 92 on my honeypot. I guess it's not as interesting.

I came across a blog post[1] (HN thread[2]) recently that sheds some light on the issue. The tl;dr is that the PR view can render over 100 000 DOM nodes, many of which are invisible inline SVG nodes, and SPA routing makes navigation a lot slower.

[1]: https://yoyo-code.com/why-is-github-ui-getting-so-much-slowe...

[2]: https://news.ycombinator.com/item?id=44799861

4 KB + whatever curl takes (540 KB on my machine).

> This is similar to good audio engineers back in the day wiring up a dirt cheap car speaker to mix albums.

Isn't that the opposite of what's happening?

I have decent audio equipment at home. I'd rather listen to releases that were mixed and mastered with professional grade gear.

Similarly, I'd like to get the most out of my high-end Apple display.

Optimizing your product for the lowest common denominator in music/image quality sounds like a terrible idea. The people with crappy gear probably don't care that much either way.

Do you have tracking protection on "Strict"? The player only started working for me after changing it to "Standard".

This works. Thanks!

Site seems to be down? I can't access it.

For me it was OpenAI. GTPBot hammered my honeypot with 0.87 requests per second for about 5 weeks. Other crawlers only made up 2% of the traffic. 1.8 million requests, 4 GiB of traffic. Then it just abruptly stopped for whatever reason.

I love this. It's pretty and really easy to digest.

It's configuration for software. If I want to be in control of and understand how my tools work I need to read (and write) those files.

Now that's a blast from the past. Before we had good one-in-all plugin solutions for solo play, I'm pretty sure I ripped the hook code from ProKreedz for my listen server. Then I got a checkpoint plugin from another guy, LJ stats from somewhere else and so on. I could tinker with my server freely and make it work just like I wanted.

That's what I really loved about CS 1.6. It allowed so much freedom in terms of what kind of maps and plugins you could create. We got amazing community-cultivated game modes such as KZ, HNS, surf and so many more out of it. And what's more, it was relatively easy to whip up your own map in Hammer and get it out there for everyone to play.

Community servers were first class citizens back then, prominently displayed as soon as you launched the game. These days someone getting into the game might not ever find out about the rich variety of experiences provided by community servers because they get funneled right into the default 5v5 matchmaking experience.

I tried TF2 recently and it took me a minute to figure out how to play a game without queuing into matchmaking. It's a bit sad.

I honestly think developers undervalue the power of moddability in adding value and especially longevity for their games. Fortunately, and as you pointed out, CS 1.6 is still there, and there's still a lot of active communities around that game. I believe that's because the game allowed the community to carve out a space for the themselves and build whatever they wanted.

Well, not just FPS games. It got ported to Assetto Corsa[1], which is a driving simulator.

[1]: https://www.youtube.com/watch?v=Yeh5vFG1GK0

What's worth noting is that the companies providing LLMs are also strongly pushing people into using their LLMs in unhealthy ways. Facebook has started shoving their conversational chatbots into people's faces.[1] That none of the big companies are condemning or blocking this kind of LLM usage -- but are in fact advocating for it -- is telling of their priorities. Evil is not a word I use lightly but I think we've reached that point.

[1]: https://www.reuters.com/investigates/special-report/meta-ai-...

I found this "advancement" creepy. It seems like they deliberately made GPT-5 more laid back, conversational and human-like. I don't think LLMs should mimic humans and I think this is a dangerous development.

As a note to future historians, this sounded just as crazy back in March.

The reading of the quote that the tweeter provided is even worse.

I'd rather not be subjected to fake news on HN.

The intention was to prevent tooth decay by regulating candy intake. But it turns out that mandating a special weekly(!) day for candy consumption had the opposite effect. Children are naturally interested in things that are forbidden from them, and so when the restrictions are lifted they tend to eat large amounts of candy. Furthermore, restrictions and rules are not a good way of teaching moderation. Practices like Saturday candy facilitate an unhealthy relationship with food, which is why it's not longer a recommendation in Finland -- I'm not sure about the other Nordics.

I feel a bit warm and fuzzy inside when the forum (and the guestbook) seem reasonably active even if they're hidden under a rock. It's a treehouse kind of vibe.

There is an option in GitHub Copilot settings to enable GPT-5 already.

That cracked me up. "The wall's design is remarkably well-done"?

I could definitely see this working quite well with more similar topics, like steering users towards the cloud instead of self-hosting etc.

Well written. I've been thinking about this topic lately and it's hard to put it into words, but you've captured into words a lot of the feelings I've had about it.

Judging by discussions I've had around this topic, a lot of people don't seem to mind reading LLM edited or straight-up generated content as long as they feel like they gain value out of it. To me, it feels like a violation of a social contract. Communication isn't just words; the words are there to help you interpret the implicit meaning of the author. When there is no author, that meaning doesn't exist and I think that's part of why it's so repulsive to decipher LLM output.

Nice post! There's a lot of stuff here that I had no idea was in built-in already.

I tried making a standalone executable with the command provided, but it produced a .blob which I believe still requires the Node runtime to run. I was able to make a true executable with postject per the Node docs[1], but a simple Hello World resulted in a 110 MB binary. This is probably a drawback worth mentioning.

Also, seeing those arbitrary timeout limits I can't help but think of the guy in Antarctica who had major headaches about hardcoded timeouts.[2]

[1]: https://nodejs.org/api/single-executable-applications.html

[2]: https://brr.fyi/posts/engineering-for-slow-internet

Not sure I like the fact that Copilot inferred that Dylan likes football so it offered to buy tickets when Dylan launched his browser.

In general, this seems to be less about empowering users and more about shoving AI in their faces. You don't need to prompt anything, it just is there.

I don't even want to imagine how bad it will get if a legitimate moat does surface and users get entrenched ever deeper into the ecosystem of a single provider. A lot of the companies in this space have a track record of squeezing as much value out of their customers as they can get away with.

You can still write simple, understandable React. All the weird, obscure hooks they have added in the recent years are opt-in and are often better left for library maintainers. You probably don't even need useMemo or useCallback. That's not to say I don't see developers (mis)using them all the time, but that's self-imposed convolution and you can only blame React as a footgun supplier.

I think it's cool that producers are experimenting with co-fermented coffee in search for new experiences, but truth be told, the co-ferments I've had have always left me disappointed. I'm guessing the process is delicate and it's easy to end up with over-fermented, overripe fruity notes that overshadow the subtler coffee notes. I've tried these coffees from several award-winning roasters and the experience has always been similar. The first sip catches you off-guard like "wow, I didn't know coffee could taste like that", but once your palate catches up, it can be a chore to finish the cup. The flavor is intense, but I think it comes at the expense of clarity and depth. It works great for competition where you need to stand out, but not so much for home consumption.

I feel the same way about beer with fruit adjuncts. Yeast and hops -- like coffee -- can express an unbelievable variety of aromas and flavors on their own. For example, a Flanders red ale can taste so much like sour cherries that you couldn't believe none were added to the ferment, mostly thanks to the magic of yeast.

When you have this kind of expressiveness in your ingredients to begin with, co-fermenting with fruit is a bit of a shame in my opinion.

I on the other hand find agentic LLMs mentally draining.

I can't enter a flow state since the workflow boils down to waiting and then getting interrupted, and then waiting again. Often the LLM does the wrong thing and then instead of moving to implement another feature, I'm stuck in a loop where I'm trying to get it to fix poor decisions or errors.

It's possible I get a feature implemented faster thanks to agentic LLM, but the experience of overseeing and directing it is dreadful and pretty much invariably I end up with some sort of tech debt slop.

I much prefer the chat interfaces for incorporating LLMs into my workflow.

moment.js has been deprecated for, what, close to 5 years now? I wouldn't recommend choosing it over any of the modern alternatives.

Depends on what you're using for routing.

In Tanstack Router it's a boolean you set when creating the router. The documentation nicely lays out what's being done under the hood, and it's quite a bit.[1] I wouldn't try that at home.

In React Router you just chuck a <ScrollRestoration /> somewhere in your DOM.[2]

[1]: https://tanstack.com/router/v1/docs/framework/react/guide/sc...

[2]: https://reactrouter.com/6.30.1/components/scroll-restoration