I find it interesting that this moved from the top position within minutes down to the second page. Below a post with 23 points in 11 hours, while this post has 102 points in 48 minutes. It has obviously been moderated down.
I suppose this won't work. The mechanism behind it is probably something like an elevated sink rate.
But it would be very interesting to scan for posts with a higher than usual sink rate. Obvious violations against the rules would just be flagged or simply downvoted. If it sinks like a rock despite having good numbers, however, then it is probably extraordinary interesting.
I had it open, then decided to read it later because my kid needed attention. Ten minutes later I was looking for it on the frontpage and couldn’t find it anymore. I spend lots of time (too much) on HN and it definitely felt like some coordinated effort, very non-organic.
Yes but by whom do you suspect? Anyone can create an account and just start flagging articles. In this case it's a bit of a screed against Silicon Valley, justified or not, so it hits certain people a certain way. Who do you think is doing the moderating down?
It has not been flagged, at least not visibly. That's the point.
Regarding your question, HN is YCombinator and the article is very much in opposition to YC's business model. Paul Graham is being criticized, too.
Not my hill to die on, but this is very much at odds with the common perception of HN as a platform promoting and enabling the free exchange of ideas and highlights its utility as a VC propaganda machine.
This is not even remotely related in any way to Gaia-X. The only connection is your marketing blitz label, which can be anything that involves public communications, really.
The code is only part of the specification. It does rarely document the actual requirements to a degree you can rely on for decision making. Sure, the code should speak for itself, but it mostly speaks about the WHYs of the implementation, not the reasoning behind the actual requirement. I found that OpenSpec actually helps a lot in this regard.
While the post is specific to Claude, interestingly I found the same terms and language patterns appear noticeably more often in recent Open AI conversations as well. Also in the same timeframe as for the Claude models. And I can see two reasons for that. One being that one company is training on the models of the other company, but I highly doubt that because of the apparent similar appearance without noticeable delay. And the other being that there has been some fundamental realization on how to improve and shape conversations.
I see people around here giving advice on how to reduce this vocabulary in the model output. I’m not certain this is a clever thing to do, as it appears to me, the emphasis on certain words is deliberate, and an important step towards overall quality as the words are being played back to the model in the next turn and thus supporting the model shape its own thoughts and stay on course.
It’s obviously more complex than this, but my intuition is that the previous messages of the conversation define the playing field for the next response in the same way as thought blocks do. consequently, when the model arrives in round n and “notices” a lot of words like “plain”, “genuine” and so on in its own previous responses, then, in its pursuit of continuity, it takes this as a strong trigger that it should not be talking fiction here. If that picture is right, then trying to remove those terms would be a bad idea. At least if you intend to continue the conversation.
Don’t you feel your conclusion is preceding the reasoning, forming a confirmation bias?
Have quick test. In a codebase or PR (assuming you’re a dev) with claude’s signature wordings unmodified, command claude (/code-review) with same “thinking” effort and model. You may find (as I have many times) that claude points out the comments it made in earlier sessions, as factually wrong or misleading. Repeat this in a fresh session a few times and on different code.
I recently set this up on my homelab, not really expecting to switch to it, more out of interest. But wow, after I saw how unbelievably fast it was, I’ve moved everything over. I’ve now set it up to mirror things automatically to Github if I want to make them public.
Using Github at work is painful in comparison. (Also literally just this morning we’ve had to delay a release due to Github being down.)
It's based on act[0] so it's quite like GitHub actions but not completely [1].
I found in the current version of Forgejo the most secure way to set it up was just to put the runner on a different host. There's a Docker in Docker setup [2] that was challenging to get working correctly but still didn't quite feel secure...
After setting it all up, I'm convinced GitHub actions as a concept is flawed. I wish the community would make a simpler git-defined CI/CD that didn't have the crazy surface area that actions does... Named shell scripts that can call out to other stuff in the runner Docker image and report it to the UI would be a nice start.
Agreed. GitHub Actions feels like someone's science project created while on a mushroom trip. Really you just want a thing that executes a shell script in a runner.
You can use any CI/CD you want. The only reason GH is popular is that it's free for public repos.
But Forgejo does have a GH like CI/CD. If you really care about good CI/CD then you should try some of the alternatives out and decide what works best for your needs.
> The only reason GH is popular is that it's free for public repos.
It's also free for private repos, and I have both. Codeberg doesn't like private repos, so unfortunately I cannot just move everything to Codeberg. I'll probably set up a Forgejo VPS somewhere.
For a forgejo host, check codefloe.com if what you need is a place to host personal projects or smaller teams. It’s been rock solid and it is now the default choice for anything outside of work I’ll do.
I don't have any specific recommendations other than do as little in your CI/CD as possible, instead do as much as possible in your build system or scripts that get called by CI/CD. This way you can migrate with less work or run locally when you want.
Hopefully a good spot to plug my own project, preloop, which is a drop-in replacement of Github actions(both the runners and control plane) that runs locally or self-hosted in isoalted microvms, and supports debug-on-failure. You can also push to the server, run CI and then optionally create a draft PR. Not quite production-ready yet(for the self-hosted part), but the local part works well. We implement the official runner protocol 100% unlike act/gitea/forgejo, so your workflows are more likely to work out of the box with preloop(forgejo has the closest compatibility with Github Actions though so it's a good off-Github option) and we use microvms so no DinD issues. I'm working on getting the official runner vm image up to reduce any environment incompatibilities. Feel free to try it out: https://github.com/preloopdev/preloop
Just set up a CI pipeline for PHP projects with a forgejo runner last week. Tried many approaches, ended up with a simple debian VM. Runner works well with docker images inside the VM, quite a flexible setup.
Found this easier and better than my last tries a few years ago with gitlab, which always felt quite bloated.
If you know GitHub actions then you’ll immediately understand Forgejo actions. It was designed that way intentionally. There are some differences, but at least for me not enough to warrant any pitchforks.
If you have advanced use cases you might be more frustrated, but I’m not aware of any off the top of my head. I think my biggest complaint is that they haven’t exposed action logs over the API, so I can’t build tooling around them at the CLI level, feed them to an LLM, or more quickly diagnose problems that arise without using the website.
