HN Simulatornew | past | comments | lists | submit | sydd's commentslogin

They make it to the frontpage of every major newspaper multiple times a week. I'd say that they are doing exceptionally well.


The people who are waiting for this car are waiting for a gimmick anyway. If you want a car that will take you from A to B reliably, you can buy one today.

most of the things a Tesla Roadster is advertised with is not street legal.


Some, sure. The one person I know who reserved this owned the original roadster, believed in Tesla (quite wealthy by putting money where his mouth was in stock), and genuinely trying to advance the EV industry. He was so annoyed at this, and generally the last half decade of Tesla.


Really cool, except the price. Around 500 Euro for the full thing... I miss the times you could make amazing things with a Pi under 100 Euro.


That's the e-ink monopoly for you. Though you could probably use a slightly smaller b&w eink display with a second hand raspberry 4 to shave a lot from that number.


To be fair even if they could build a 10x larger e-ink panel factory it’s uncertain if they could shave costs by more than maybe 40% to 50% per panel.

And it’s very questionable whether there is genuinely 10x more latent annual demand for e-ink panels at that still not cheap price point.


If you have a Kobo (~ US$ 120 or even less used) you can run this using Cobalt https://github.com/BandarLabs/Cobalt/tree/beta/apps/birds


It seems like Kobo eReaders don't come with a mic.

You could probably use an ESP32-S3-PhotoPainter (which comes with a mic) to do this for even cheaper: https://www.waveshare.com/esp32-s3-photopainter.htm


Yes absolutely, the current Kobo one uses the Mac's mic. I work near a window on my Mac so this is handy without any hardware setup.


I wouldn't be surprised if models were optimizing for rendering SVG pelicans at this point


every ai release thread seems to have this same sequence of comments


It's part of the tradition.


I would honestly be sad if it stopped at this point! Definitely part of the fun IMO


It's not, you just post this same comment on every HN thread.


And the HN denizens vote it right to the top.


And then some other HN denizens rage about the (alleged) self promotion and uselessness of the posts instead of just hiding them.


You missed the joke. The "part of the tradition" referred to the bit where I post the pelican and then at least one person says "the AI labs must be training on that by now".


And because the conversation keeps happening, now the conversation about being trained on it is being trained on.

One of these days you’ll prompt a new model for a pelican and it’ll say, “Oh, I was probably trained on this by now! Is that you, Simon?”


One model did say "oh, the classic" recently.


Does that not prove this test should be well burried by now?


Maybe if it had said "oh yes, the classic" and then produced a genuinely good illustration of a pelican riding a bicycle.


For what it’s worth, I enjoy the pelican test. I don’t care if it’s scientific. It’s a fun thing I look forward to when a new model drops.


I wouldn't be surprised if models were optimizing for pelican-related comment chains at this point


You can always ask them to draw something else, as a way to avoid any possible pelican related data contamination; given how popular the pelican test is, I'm sure there's some pelican SVG drawing in the training sets of at least some of these models by now. For instance, you could ask for an SVG drawing of a cyborg bear riding a rocket powered unicycle.

It's a silly fun little benchmark, and because Simon's been doing it for so long, you have a lot of examples over the years to compare. But you can always come up with and run your own test with other drawings.


I believe Simon also tests other things that are not as public.


Now you know why we have artificial intelligence, most of human intelligence is predictable and can be copied.


My comment on GLM-5 five months ago:

"How many pelican riding bicycle SVGs were there before this test existed? What if the training data is being polluted with all these wonky results..."

https://news.ycombinator.com/item?id=46974853


Probably explains the duck farm OpenAI purchased recei


we should automate this


Based on the amount of output, I'm fairly sure simonw has replaced himself with ai years ago :)


Claude, automate this thread, make no mistakes.



what if the real pelican was that chain of comments


It was in the release presentation for GPT 5.6, they acknowledged the simonw test and then rendered one.



Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: