regenschutz
regenschutz on Hacker News
I'd love to someday see someone eat their words non-metaphorically.
Everything is bug-free until someone finds a bug in it.
I'm seeing some other comments talking about how no one is harmed by AI CSAM. Fundamentally, I reckon AI CSAM is the same as Lolicon, which is illegal here in Sweden. I believe the same applies to Denmark, though I am by no means knowledged about Danish law.
AI is not something new in this space, it just means that a lot more grotesque fictional CSAM can be produced. It has been possible to create realistic-looking-but-fictional CSAM since Photoshop came out (and then some).
I've always found Google Slides to be a great way to create cute, personal games. It requires no coding, and anyone can play them easily in their browser. Seeing what people are able to do with a simple presentation (PowerPoint, if you will) has been repeatedly eye-opening every time I've seen one.
I'm a bit worried that Google Slides-games will fall out of fashion now ;(
Environmental impact aside, that seems way less harmful than TikTok, though. You'll probably play the average game for longer than you watch the average TikTok video, meaning that your attention span won't be
that bad.
You also have to actively use your game for most games, unlike TikTok videos which you can just watch with your brain turned off.
I got 403'ed too (Sweden). Then I saw this in the fine-print on the blog:
>Playground launches today for users in the U.S. (18+)
The latter. Maybe it's different in Lisp, but whenever I read code, I rarely care about the exact order of operations, just which ones are actually being executed. In that regard, the improved version allowed me to comprehend the gist of what was happening much faster than the regular version did. For a few seconds, before even attempting to start comprehending the actual operations being executed, my eyes were stuck trying to comprehend the parentheses.
That's a crazy effect. For me, as someone who has a very minimal understanding of Lisp (apart from messing with it in Emacs for at most an hour), the improved version is much more readable.
Not necessarily. NLP is possible to do without any ML. See ELIZA [0] as an example, which used pattern matching and substitution to process natural language. A few simple rules can cover a majority of requests. It's still AI, but not ML.
Though I do agree that the ML-based NLP thst we have today is much more usable, albeit more compute-expensive.
[0]: https://en.wikipedia.org/wiki/ELIZA
Why should one use something adequate when something amazing exists?
There really isn't any proof that they can show, since iknowwhatyoudownload legally aren't allowed to download anything that you're seeding (since that would be considered piracy).
IIRC, they only show the filename, last-seen timestamp and user agent. I can't verify it though since their website seems to be down?
>LLMs aren't even AI, […]
I am very curious to know what you classify as AI then. Do you mean that LLMs aren't artificial? Or aren't intelligent?
HTML is notoriously complicated to parse correctly.
* Is generating valid HTML documents as complicated/difficult?
* How much time did you spend on reading the HTML Standard?
* What resources did you use? MDN? The spec itself? LLMs?
It seems to be playing on Nostalgia, judging by these paragraphs like this one:
>The dithered photos look somewhere between a newspaper halftone and an old video game. The processing is part of the fun: you watch the colors come in one by one, like a digital polaroid.
Frankly, it does seem like a very fun gimmick, but also highly impractical. But my adoration of this little thing might just come from a general longing of the low-tech, simple devices of the past.
No, it's real. Every year, there are several news articles about berry pickers being abused, at least here in Sweden (not sure about the other Scandinavian countries). Here's [0] just ONE of the myriad of articles I could find, but there are so, so, so many more (and even worse ones) [1].
[0]: (In Swedish) Berry entrepreneurs suspected of trafficking Thai nationals, (2025). https://www.svt.se/nyheter/lokalt/vasternorrland/barforetag-...
[1]: (In Swedish) Berry pickers (Topic). https://www.svt.se/nyheter/om/barplockare
Both are from SVT, the public broadcaster in Sweden.
Wow, I've only ever heard of regional pricing being described as an overwhelmingly positive concept. When phrased as "price discrimination", it invokes a completely different set of (negative) emotions in me.
It's weirdly uncomfortable knowing that phrasing has such a big impact on one's emotions. It really shows how vulnerable we are to manipulation.
Neurotypical is the word you're looking for, not "normal".
Non-tech-related stories frequently make the front page here too. It just so happens that this site has a lot of coders because, err, it's literally Hacker News.
I mean, even in countries with extremely high levels of Christianity, such as Romania, you won't be accused of blasphemy for just... agreeing with science. Both Christianity and Islam are pretty similar after all, so hopefully that is not too surprising.
I can't think of a country where you even might be accused of blasphemy, though I'll admit I am not very familiar with the topic.
I mean, AI is not good at writing x86-64 assembly code. Last time I tried (with both Claude and ChatGPT), the AI failed to even create basic programs other than Hello World.
Or like the US solves it, through immigration. In the US, the fertility rate is at roughly 1.6 children per woman (which is below the 2.1 children per woman required for a stable population), and yet the US population is steadily increasing thanks to immigration. One can talk all day about pros and cons of immigration, but it is ultimately the only solution we have to a falling fertility rate (other than trying to increase it, of course).
A constant stream of young workers is required for a sustainable economy.
In order to pay for pensions, the government borrows money from young, working adults. This is effectively what happens in pay-as-you-go public pension systems (which is most of them, to my knowledge, apart from the US, I'm not 100% sure how pensions work in the US). The money you put in actually goes to pay for another person, with the government guaranteeing that they will do the same for you.
If the percentage of retired people increases, the percentage of working adults naturally decreases. Eventually, you'll hit a turning point where the government can no longer borrow from working adults. The government is now in a debt crisis and has to loan money from banks or foreign investors at a significantly higher interest rate, which becomes even more unsustainable if the percentage of retired people increases even more.
This is what is happening in e.g. South Korea and Japan. There are too many old people, and too few working adults. This is caused ny low birth rates over a long period of time.
>cynical optimist
Never heard that one before!
It definitely depends on the platform, though. On GitHub, absolutely, work with the garage door up. But on some platforms, it's the complete opposite.
Many moons ago, I was making mod for a game and had the idea to publish it on Nexus Mods [0] so that I didn't have to bother setting anything up once I actually wanted to publish the initial release of the mod. It was not at all in a working state when I made the public page for it.
Imagine my surprise when I wake up the next day and have thousands of views on the page and a dozen comments berating me for publishing a mod that doesn't work...
Ever since then, I have had problems with working with the garage door up, even though I know that it's totally acceptable on GitHub. It's habit by now to work on everything with the garage door down, just in case...
[0]: https://www.nexusmods.com/
It would be cool to be able to see a demo page for each of the different tiers, e.g. to see how much branding there is before signing up, or how much customisatio there is, etc.
What issues do you experience with the currently available sweeteners?
If you're fine with a paid option (although it's source-available and distributed for "free" on their website), then Grayjay [0] is my personal favourite. I can't remember the last time I saw an error. On NewPipe, I can't say the same.
[0]: https://grayjay.app/
The Fast demo model is already very impressive. It was way better than expected, but still required being a bit verbose since it didn't seem to understand rarer words ("sauna" didn't get me anything pleasant, "hot sauna" did).
The generated palette seem to be a great indicator of whether the model understood the prompt or not.
I Haven't checked out the Python SDK yet, but it seems very interesting!
I'm curious to know if there is any reason for why you picked Gemma 1B for the Expressive model. Did it generate more cohesive parameters than other 1B models? Or was it just the first one you picked?
Why can't they run the API at a huge loss too then??