lukeinator42

The writing style is this staccato LLM-like style that is difficult to read as it has zero flow and meaningless sentences.

Like what does the second sentence even mean? Is it even a sentence? "The roofline math, the prompt-processing catch, the NPU red herring, and the owner-measured speeds."

It's extremely feasible. English is the defacto language of most research in the EU. I spent some time in Italy during my masters degree and know lots of people who have gone on to research or university instructor jobs in various countries in the EU. You definitely have to learn at least some of the local language for day-to-day life though.

isn't this just a by-product of seed hybridization which has enabled pretty much all of modern agriculture? A lot of testing goes into finding the two strains that combine together to create high-yield seeds, but the next generations of those seeds won't produce well.

It's a cool technology, but for it to gain commercial interest it needs to solve a problem better than the status quo. What problem is it solving and for who? If I was to buy that mmwave radar device it would probably cost more than the $60 test, and I would want assurances that it is as accurate as existing tests.

This happened to me on a paper I submitted recently where it was clear the reviewers used AI. Revising a paper based on LLM review is also exhausting, haha.

Just because the US sanctions a country, doesn't mean the rest of the world needs to as well. As a Canadian we traded with Cuba even when the US had an embargo on them.

I think the scanner outputs petabytes of raw data that then get processed down into images, so you would only need to store an image per slice ideally, kinda like MRI DICOM files.

This would be great for other use cases too like climbing bolts and anchors in coastal areas. Lots of areas are switching to titanium glue-ins which are expensive, but I wonder if this could enable more affordable option. A climber recently died in Greece after multiple bolts failed: https://gripped.com/news/rock-climber-dies-in-kalymnos-after...

The other concerning aspect of this leak is the fact that the list was shared with a group of separatists, and the data on the list is basically everything you need to fraudulently sign someone up for the separatist referendum petition. Some separatists are claiming that certain ridings have had 92.9% of eligible voters sign the petition which is highly dubious: https://x.com/RiseOfAlberta/status/2049668987307303389.

Elections Alberta has now said they are going to check for this: "Verification after today’s date will include determining if any of the seeded names from the Republican Party of Alberta’s List of Electors are contained in any incoming petition." https://www.elections.ab.ca/resources/media/news-releases/me...

nope, nike vaporflys are super popular. There are usually limits on stack height for many marathons though.

https://openreview.net/pdf?id=BZ5a1r-kVsf

If they didn't do that people would probably be commenting that its fake or AI.

A Baofeng UV-5R cannot receive shortwave, it's in the VHF/UHF range for receive/transmit and can receive commercial FM broadcast.

A good analogy for equal temperament might be the Gregorian calendar, as a year does not evenly divide perfectly into 365 days. So in order to compensate for that we adjust the calendar by a leap year every so often to make the calendar more accurate in the longer term. That's kinda similar to how every note is a little off in equal temperament so that at the larger scale of being able to play all intervals works out.

it's still up for me

No, they mean we actually clear US customs in Canada as non-US citizens.

Same, I haven't been able to get gemini or claude to tell me to walk a single time and I've even tried changing the distance in the prompt, etc.

It's a within-subject design and they literally did a power analysis in the paper. This is not absurdly low n.

I have a couple slings my cousin brought back from Tibet, and a few years ago when I was in Peru I brought some back from there as well (they're called 'huaraca' in Quechua I think). I find it interesting that many cultures used them historically but it never became a sport like archery, javelin, etc., did.

They're very lightweight and are definitely an underrated backpacking tool for keeping marmots at bay when they're attacking your tents and gear, haha.

I was also wondering this, and in one of the footnotes they say "Given that our experiment was conducted in 2025, one might wonder whether Kansas’ updated law is reflected in GPT’s training data and thus skews its decisions. We find no evidence of such contamination." when talking about a specific updated law. But how does one have 'no evidence of such contamination' without seeing the training data?

I also found this blog: https://www.theastroimager.com/picoballoning/pico-ballooning.... His balloon JR14 flew for 507 days and made 6 laps of the earth which is pretty wild.

This is an interesting article from a couple of days ago about tracking diy balloons long distance: https://spectrum.ieee.org/explore-stratosphere-diy-pico-ball.... Given the tracker can be built for $14 it might be worth it to test a version with just the custom hydrogen enclosure and tracker and see how far it gets.

I think it's worth looking at this commentary on the study: https://arxiv.org/pdf/2601.00856. It aligns with a lot of our intuitions, but the study should definitely be taken with a grain of salt.

I mean it's not just any old foreign bank either though, in Canada the King of England is still our head of state.

Light mode constricts your pupil more, which means less eye strain for the eye when focusing because of the better depth of field. Also, black pixels != no light except in technologies such as oled, but most laptops are backlit lcds.

It's only line-of-sight, but isn't the range 10s-100s of kilometres in open areas? Some repeaters on hills/mountains etc. could connect large areas potentially.

reminds me of this warning from Dante in the YARP documentation: https://www.yarp.it/latest/warning.html.

The title confused me at first as PCBs refer to hydroxylated polychlorinated biphenyls not printed circuit boards.

yeah that checks out, although looks like they do have an example for running models on a raspberry pi pico 2: https://docs.pytorch.org/executorch/main/pico2_tutorial.html. The list of embedded platforms this can run on is probably greater than the list of backends, it just wouldn't have acceleration.

It is fairly well established that neurons in these artificial neural networks are polysemantic and information is represented in directions in the activation embedding space rather than neurons independently representing information (which is why anthropic is doing things like training sparse autoencoders). I haven't read the paper in depth but it seems like it is based on a fundamental misunderstanding about neurons in ANNs vs the brain.