Yeah, I came across Geerling 10-12 years ago when he was big in the Ansible community and he seemed pretty reasonable then.
In his defence, those blog posts are ancient and he'd have been pretty young when they were written. But against that, they go beyond being merely cringey and lacking in self awareness; they're actually pretty vile.
In the comments on https://web.archive.org/web/20250113010304/https://www.jeffg... he partially acknowledged that they're not "culturally appropriate", but then claimed he hadn't actually re-read the blog post in question which is, at best, rather slopey-shouldered (it would have taken less time to re-read than to comment!)
Such a transparent avoidance of responsibility actually makes him seem worse than if he'd said nothing at all.
They don't hate others for stuff they can't control; they have moral or conceptual views that differ from progressive orthodoxy, and progressives are claiming that holding these moral and conceptual views constitutes hating others for stuff they don't control, because hating others for stuff they don't control superficially sounds like a good justification for socially ostracizing someone. Geerling's views on abortion are a religious scruple, and I don't have particularly more of a problem with them than I have with, for instance, the specifics of what any individual Muslim in the world professes to believe. I don't necessarily agree with Rossman's ontology of gender, but I don't necessarily agree with the ontology of gender held by any particular trans or genderqueer person either; and I think I have much less reason to fear social punishment over this disagreement from Rossman than I do from the average genderqueer person in my community.
I still see them occasionally and read them when I can. Because the content a human has made is much more interesting than something possibly plausible but useless that an Agent could tell me. I've been working on a project that'll look at the HN new submissions and start to filter out the stuff I really don't care about (news, product announcements, stuff about LLMs), which makes technical blogs more likely to show up, then I can bookmark them for later reading
Haha bro, I came up with the same idea that I need just simple LLM-filter with my own criteria to find something interesting. You can check it there: knoless.vercel.app. Maybe you can give me some advice :)
I just filter out LLM related stuff, news pages, product releases etc by looking at the title and URL, and then look at what's left to see if I actually like the content
There tend to be two kinds of people, those who value the journey and those who value the product. This telescope is aimed more at those who value the product, and hand tool woodworking is probably more for those who value the journey. You would want a telescope that lets you value the journey
> How do you locate celestial bodies like that? Is it just a case of paper maps and a compass?
I mean (based on peer replies) it does come off as a funny question, but I guess if you're young enough it is a totally legitimate question! Never thought of it that way.
But yes, you have a star chart, or atlas. The North Star locates your north and everything else you can then reference from your chart.
(I'm not familiar with southern hemisphere viewing)
And after a little while you recognize the major constellations by sight, so you can orient yourself vs. the chart even if you don't see the North Star for example.
Planets are different since their relative position changes all the time, so a static chart can't tell you where they are. But there are really only four planets you can see with eyes (Mars, Venus, Jupiter, Saturn) so you learn to recognize them pretty quickly.
I understand maps and charts. But generally, stars stay in the same place, so its easy to put them on a map. I was trying to understand how you put a planet on a map when it keeps moving around
You'll get a better result if you put in your actual location, but whatever. Looks like Saturn is visible this time of year, and will be in the east around midnight. You can check this right now and it'll be about right for the next few weeks. You can step the time forward and see what how the planets move over time.
If you don't do the above steps, you could probably make a guess, as the planets are the brightest objects in the sky, easily seen by eye.
Then you point your telescope at it. I found mine in the trash and it didn't come with the spotting scope, so I kind of just have to guess where to point it. Use your widest angle lens, and move the scope around until you see something bright go by. You're close! Then you center the bright thing in the center of the view, and after the scope stops shaking because it's on a cheap tripod, you can see a very tiny Saturn, including the rings. Then you swap the lens for a more zoomed in one to see it better. The rotation of the earth means it will quickly drift out of view, so you have to constantly adjust your scope every minute or so to keep it in frame, until I fix the motor that came with my equatorial mount which is supposed to do this part for you.
> How does one locate Saturn in the night sky without an electronic aid?
Seriously?
You mean that _really_ bright star that is out of place on your constellation map / planisphere?
It's a planet.
Look at it and see which one it is. (BTW: Look at it a few nights later as it moves to work out its trajectory. Congratulations, you can now identify which planet it is anytime, no telescope needed.)
I'm sorry if I'm not familiar with your area of expertise, that's why I'm asking the question, because I want to understand. There are lots of stars that are brighter than Saturn, even at its brightest. I don't know if I'd know if a bright point of light is Sirius or Vega or Saturn or what (I could probably at least tell if it's Polaris). I was just trying to find out how you go from a night sky to pointing at Saturn, or Mars, or Jupiter, unless you have a map showing you where it would be on a given night.
Also, don't planets sometimes seem to move in non-intuitive paths, like Mars performing retrograde motion when viewed from Earth? If you were viewing the plants constantly, then sure, you could pick it out, but if you left a gap between observations, wouldn't it be a lot harder?
SERIOUSLY?! You'd be surprised how few people know you can easily spot planets with the naked eye, in addition to many other little interesting things in life. You react as if he didn't know how to his shoes.
If you download a zip file that contains the hidden .git folder, with the config file inside, and open it with an LLM, it can run an arbitrary program, if you ask it to or not. Even if it's not actually a git repository, as long as it has enough to trigger the hooks when you run "git status". That sounds like a massive problem to me
Your threat model is that you give an untrusted .zip archive to an LLM, and it decides to unpack that, then decides to try using it like a git repository because the unpack produced a .git folder?
Yeah, I wouldn't have --dangerously-skip-permissions on for that. I mean, the archive might also contain attempts at prompt injection.
Doesn't require a permissions bypass or a prompt, that's the point. Every agent here ships a workspace trust prompt but the payload fires before it's shown. `--dangerously-skip-permissions` is therefore irrelevant as the permission system never got consulted.
Delivery isn't self-inflicted either. A colleague sends you a zip or a client hands over a codebase. Inspecting `.git/config` before opening a shared project isn't standard practice. Pointing an agent at a project you were sent is what agents are for and exactly why these tools ship a workspace trust prompt. The issue here is that it fires too late.
Using git is not a requirement here. You just need enough of a git folder for a "git status" to trigger running a program, and the "git status" is run automatically by an agent, without you asking it to. You could open any innocuous folder that you have been sent by someone else, and the agent would run the program. At my job, we send clients a folder containing the completed project. If someone was to put in a .git folder with the config option set, not even a full repository, and the client opened it with an agent, then we could have made the client run any command we wanted
reply