Ariarule

I don't completely like or think this is a good ultimate decision, but looking over the thread on Github, it's being done for practical rather than ideological reasons after trying a more liberal policy first, and the discussion thread there indicates that they're already considering how to relax and refine it. A long-term and total ban on AI contributions would be a mistake as it all but guarantees a competitor or a fork, beside enforcability issues that create a "now you have two problems" situation. But as a temporary stop gap as specifically indicated by https://github.com/qgis/QGIS-Enhancement-Proposals/pull/387#... this seems fine, and it's such important context that "QGIS bans AI assisted contributions" is true but misleading.

The linked discussion thread specifically calls this a "moratorium" and it's unfortunate that they decided not include that explanation in the policy document specifically to avoid controversy and flak, with the end result that there's now a 202-comment-plus HN thread: https://lists.osgeo.org/pipermail/qgis-developer/2026-Septem... -- the "proof of life" concept being tossed around in that thread, although rejected, is genuinely interesting and I hope they revisit it and allow a nominal financial contribution to the project to prove a human is behind a PR regardless of AI assistance (probably along with some other approach, as a total exclusion and the only way it would also be bad).

My own experience with QGIS is limited to the most basic navigation and queries of some local maps I've archived. Nonetheless it's an obvious gem and I hope they're able to arrive at a balanced and sustainable policy. Maybe they would get there faster and more reliably by tightening the previous LLVM-inspired policy rather than the slam-shut and then loosen approach they seem to be taking, but hardly anyone has experience with a situation exactly like this, so only time will really tell.

The previous top comment is trivially findable with a page search for "AI". Contrarian dynamic could explain why the reactions overall look more positive now, fair enough; I'll keep in mind it's possible to give an incorrect impression in the future as a result of the comment order shifting.

When I made my comment yesterday, the previous top comment was negative and dismissive, and for a vacuous reason. Why not reply directly instead? I decided it would be better to positively _appreciate the original work done in itself_, to not boost a vapid debate from a eye-rolling starting point, not to make it personal towards anyone, and especially, as it is actually the greater portion of that comment you quote, to make an analogy to other techniques.

And then I get this useless advice to "give it a rest", as if there were no such negative reaction in the first place, no such parallel drawn between older work and newer work, and as if the counterfactual of what would happen if people were not willing to appreciate such works was not immediate obvious. If there were a substantial contingent on HN claiming that projects done on Apple devices were somehow not real, and trying to shout down Mac-based projects, it would indeed be very tiring, but the problem would not be with people musing on parallels to similar workflows in Linux environments to highlight and acknowledge when a project done on a Mac was good work, nor with Apple fans!

Wonderful work, and it's unfortunate that knee-jerk anti-AI sentiment is getting in the way of appreciation for the accomplishment by some others. Consider this: if the exact same oddities and anomalies had been discovered 3 to 5 years ago with traditional NLP, or OCR and/or some clever statistical techniques, would anyone be dismissing it? The post goes into the work they did to orchestrate and explore with the AI in this case, so it's no less impressive an effort than the hypothetical, and the discoveries just as valid for consideration.

I do not, in fact, see such an unambiguous, unhedged claim in OP. Can you quote it?

Ha! It is actually _some_ funny serendipity. I just stumbled across this in my saved pile and thought it would be an interesting oldie to share, and did not realize that release was going to be today at all until I saw these comments here now.

It was actually submitted three days ago originally. I don't know if whoever second-chanced it had the timing with the new game in mind, so it may or may not be a complete coincidence that way.

Amazon says it's in stock right now. Costco also has it online and says it's on the shelf at my local warehouse.

It was literally available from the first two places I checked.

Due to the HN length limitation, it was necessary to abbreviate the title. Originally: "AI Film ‘Gods Don’t Give Gifts’ Sets First Trailer for December Theatrical Run, Plans Oscar Bid for Best Animated Feature".

Always glad to see more open-weight models, but this caption on the 2nd demo image had me do a double-take: "Land or Water Generalization Experiment: We recreated the viral X puzzle by asking Beam to create a fixed 180×90 grid for longitudes -179° to 179° and latitudes -89° to 89°, with 16,200 points. This puzzle is a few days old, so could not appear in the training data, thus testing the model’s generalization. Beam gets 95.5% coverage right, putting us between Opus 5 (92.5%) and Fable 5 (97.8%), which shows how well it generalizes to novel new tasks."

Oof, no, this "puzzle is a few days old" is incorrect even if it's a social media trend just recently. Asking a model to generate a world map in this way is _at least_ from August 2025 as it appeared on LessWrong at that time: https://www.lesswrong.com/posts/xwdRzJxyqFqgXTWbH/how-does-a...

No byline, acknowledging that it's part of series that usually doesn't have them. Google trends has nothing for the phrase right now. UD, as of this writing, never heard of the expression (although don't be too surprised if someone swoops in to "correct" that later). Even for a generation that's ostensibly not going to have a huge social media footprint yet, that's rather remarkable for a supposed "biggest insult".

https://trends.google.com/explore?q=%22That%27s%20so%20AI%22... https://www.urbandictionary.com/define.php?term=That%27s+so+...

Ok, if they're meant to be marketed for food service, like a bakery regardless of who actually purchases them, then that does makes sense.

Huh? Processed ingredients don't generally get a pass on nutritional labeling requirements in the US. A bag of flour in a grocery store should have a label.

That is also the title I see clicking through into the page, although it seems to be 24 characters over the limit, so that _some_ change to OP would have have been required to submit it.

Not just these three. OP mentions also Cloudflare, and additionally Downdetector also has AWS, Azure, and Google (both search and Gemini) listed as having spikes about the same time: https://downdetector.com/

A sidenote, but that is not a typical Amish opinion of technology:

"Amish do not view technology as evil, and individuals may petition for acceptance of a particular technology in the local community" - https://en.wikipedia.org/wiki/Amish_way_of_life#Use_of_moder...

See also https://thereader.mitpress.mit.edu/how-the-amish-let-technol... about some Amish use of cell phones with adapters to be more payphone/landline-like and https://news.ycombinator.com/item?id=21349989 was a discussion about the use of pneumatic machinery by the Amish.

Joel Spolsky wrote that about "large scale commercial applications", which Monica definitely is not.

Surprised there's not a specific mention of learning by doing, which means probability of success is not static, when this is _just_ under the surface of the content of this post. They did kind of hint at this with the "Each attempt... a little better than the last."

While I know "if they'd cover it on TV news, it's probably off-topic", I feel this is especially unusual due to the gap between the level of coverage (mostly regional at this point, like WGN) and the impact. Currently, poweroutage.us is showing 234459 customers out in Indiana, with elevated outages in Illinois and Ohio as well.

How is "Fix the WAL-reset database corruption bug" downplaying this? https://www.sqlite.org/changes.html

Very cool. If I can point out an alternative use and audience for this: there are quite a few game marketing data websites for sussing out which games are similar to each other and how they've performed within their genre. This is, AFAIK, a unique approach to the first part.

Release notes are at https://download.samba.org/pub/rsync/NEWS#3.5.0

Steam Achievements are trivially to cheat at for most games, especially single-player like Ecco, for example via SAM.

Odd not to highlight the sentence where they answer the obvious question "Why Tailscale in particular?":

> They also explained why we were more likely to hit the bug than other SQLite users: we take manual control of the checkpointing process, and we checkpoint very aggressively. Even a bug triggered by a rare condition was bound to hit us eventually.

I'm guessing you meant to link https://www.economist.com/leaders/2026/08/06/how-ai-is-break... ?

Can you name three reasonable-quality games out there like this that were definitely made in such a way? It costs only $100 to publish a game on Steam, recoupable if you meet a fairly low revenue threshold. You can publish elsewhere with no money up front (e.g. Itch). If this were actually true, and even if 90% of players would just pass based on the AI disclosure (doubtful, and also doubtful that _all_ games made this way would actually disclose it), you'd expect there to be lots of Myst quality games out there; hundred to thousands already, at least a handful of which should _already_ be successful.

Not entirely a rhetorical question. I've found any images need serious review and touch-up at a minimum and usually multiple rounds of revision. The topology on any generated 3D models is generally atrocious. TTS is not tuned for voice acting to an extent it's absolutely no good for that purpose as far I can tell, and that would be a cornerstone of any serious modern "Myst-class" adventure game. I can't comment on music and SFX in particular and have heard secondhand that current models are "all right", but imagine it's probably the same situation: no care taken, you get uselessly mid results. If it's really that easy and that cheap, I'd like to see it.

_Dimhaven_, released just a few months ago, looks fairly Myst-like from the videos I've seen. While I haven't played it, Zadbox, the developer, also made the very Myst-like _Quern_; released 2016 and which I have played.

I can't help but think the article is a story about _Cyan's_ woes conflated with industry woes. Which isn't to say that the industry doesn't have problems! But Cyan really underperformed with _Firmament_. Look at the Kickstarter comments:

https://www.kickstarter.com/projects/cyanworlds/firmament/co...

They did a rug pull on whether they would ship DVDs with the Firmament collector's boxes _after_ they had complaints about the quality of the equivalent Obduction backer reward. While they did ship a DVD for Obduction, from personal experience the box itself it came in was a bit underwhelming. I didn't care so much about packaging, but I could see how someone would, and the whole situation on this point with Firmament was a very eye-rolling "again?" scenario. The first time with Obduction that they weren't able to ship what they planned in this way was bad luck; the second time was, generously, bad planning about something that was entirely foreseeable on their part.

More details here: https://steamcommunity.com/app/754890/discussions/0/38264157...

They made a bad bet on how well VR would take off, with Firmament very obviously created with VR play in mind.

So many people were disappointed with Firmament itself. "18,420 backers pledged $1,433,161 to help bring this project to life" but they can't do that again specifically because they burned a huge amount of goodwill with the people most likely to support them that way. Keep in mind that's a 2019 number and CPI adjusting would bring it up to about 1.8 million, at the lower end but solidly within the 1 to 10 million amount Wired said they were trying to raise for Anglerfish from a publisher. And _Obduction_, while beautiful, was a bit short and had its own issues.

I would also note with all the discussion of the impact of AI art on games that Cyan was an early _target_ of the anti-AI crowd over their use of AI art in Firmament. IIRC, Steam did not have a dedicated disclosure section at the time Firmament, but they did disclose it responsibly and appropriately in the credits. A more detailed statement from Cyan can still be found at https://cyan.com/2023/06/08/regarding-ai-assisted-content-in... -- on this, I agree with Cyan's 2023 post, we ought to be able to distinguish "AI Assisted" from AI-generated, although it seems many are allergic to such nuance.

I find it incredible that "modify the chosen route as little as possible" is not really a thing in any major mapping app. Just because I turned left instead of right at an intersection and had to detour a little bit before getting to a highway shouldn't switch the way I'm going by hundreds of miles on a long trip just because it's supposedly 5 minutes or some other pithy amount faster, and I have seen scenarios like this happen in both Google Maps and Here.

LLMs are ascendant partially because they do not "RTFM", "read the FAQ", or "question closed as duplicate" at you regardless of how many times they've heard a question before.

Is a painting not art because individual pigments aren't art?

Not criticizing this in particular, but I'm genuinely curious: why is everyone talking about how obvious or not the details of the poster are as AI, when they didn't take the 30 seconds in most good image editing software to remove the sepia/yellow/brown tint all over? You could put a 100px blur on the smallest thumbnail on that page of it, and I'd still recognize it as AI generated.