Nick @internic@qoto.org

Ten million new cases of dementia are reported globally each year. But a recent study suggests up to 13 percent of people diagnosed with dementia in the U.S. may have a misdiagnosis and are instead struggling with a condition called hepatic encephalopathy. The upside: it’s treatable. @ScienceAlert has more.

https://flip.it/JYLv2Y

#Science #Health #Humans #Neurology #Brain

Mathematical physicist Cécile DeWitt-Morette was born #OTD in 1922. She made foundational contributions to the study of Feynman functional integrals, organized the first American conference on general relativity (playing an indirect but important role in the eventual detection of gravitational waves), and started the Les Houches Summer School.
Images: UT-Austin

Some people think medieval astronomers kept adding 'epicycles' to the orbits of planets, culminating with the Alfonsine Tables created in 1252. The 1968  Encyclopædia Britannica says:

"By this time each planet had been provided with from 40 to 60 epicycles to represent after a fashion its complex movement among the stars."

But this is complete bullshit!

Medieval astronomers did *not* use so many epicycles. The Alfonsine Tables, which the Brittanica is complaining about above, actually computed planetary orbits using Ptolemy's method, developed way back in 150 AD. This method uses just 6 circles and 6 epicycles - nothing like Britannica's ridiculous claim of between 240 and 360 epicycles.

That's right: Ptolemy got a good fit to planetary orbits using one circle and one epicycle each for the Sun, Mercury, Venus, Mars, Jupiter and Saturn. It's not much worse than what we do now: use one ellipse each for Mercury, Venus, Earth, Mars, Jupiter and Saturn.

I must admit that in Ptolemy's model, the circles weren't centered on the Earth. They were offset, like in the gif below, where the big green dot is the Earth. The big blue circle, offset from the Earth, is called a 'deferent'. This approximates an ellipse. The smaller black circle is called an epicycle. This makes up for how the Earth is not actually stationary, but moving around the Sun.

This gif was created by Richard W. Pogge, Distinguished Professor of Astronomy at Ohio State. You can see more of his animated planetary models here:

https://www.astronomy.ohio-state.edu/pogge.1/TeachRes/Artwork/Anims/index.html

So, just because something is in an encyclopedia, or even an encyclopædia, doesn't mean it's true. Don't believe me? Check out the references in part 2!

(1/2)

@nixCraft and youtube now shows a popup on brave, whenever I play a video, pathetic.

I finally turned off GitHub Copilot yesterday. I’ve been using it for about a year on the ‘free for open-source maintainers’ tier. I was skeptical but didn’t want to dismiss it without a fair trial.

It has cost me more time than it has saved. It lets me type faster, which has been useful when writing tests where I’m testing a variety of permutations of an API to check error handling for all of the conditions.

I can recall three places where it has introduced bugs that took me more time to to debug than the total time saving:

The first was something that initially impressed me. I pasted the prose description of how to communicate with an Ethernet MAC into a comment and then wrote some method prototypes. It autocompleted the bodies. All very plausible looking. Only it managed to flip a bit in the MDIO read and write register commands. MDIO is basically a multiplexing system. You have two device registers exposed, one sets the command (read or write a specific internal register) and the other is the value. It got the read and write the wrong way around, so when I thought I was writing a value, I was actually reading. When I thought I was reading, I was actually seeing the value in the last register I thought I had written. It took two of us over a day to debug this. The fix was simple, but the bug was in the middle of correct-looking code. If I’d manually transcribed the command from the data sheet, I would not have got this wrong because I’d have triple checked it.

Another case it had inverted the condition in an if statement inside an error-handling path. The error handling was a rare case and was asymmetric. Hitting the if case when you wanted the else case was okay but the converse was not. Lots of debugging. I learned from this to read the generated code more carefully, but that increased cognitive load and eliminated most of the benefit. Typing code is not the bottleneck and if I have to think about what I want and then read carefully to check it really is what I want, I am slower.

Most recently, I was writing a simple binary search and insertion-deletion operations for a sorted array. I assumed that this was something that had hundreds of examples in the training data and so would be fine. It had all sorts of corner-case bugs. I eventually gave up fixing them and rewrote the code from scratch.

Last week I did some work on a remote machine where I hadn’t set up Copilot and I felt much more productive. Autocomplete was either correct or not present, so I was spending more time thinking about what to write. I don’t entirely trust this kind of subjective judgement, but it was a data point. Around the same time I wrote some code without clangd set up and that really hurt. It turns out I really rely on AST-aware completion to explore APIs. I had to look up more things in the documentation. Copilot was never good for this because it would just bullshit APIs, so something showing up in autocomplete didn’t mean it was real. This would be improved by using a feedback system to require autocomplete outputs to type check, but then they would take much longer to create (probably at least a 10x increase in LLM compute time) and wouldn’t complete fragments, so I don’t see a good path to being able to do this without tight coupling to the LSP server and possibly not even then.

Yesterday I was writing bits of the CHERIoT Programmers’ Guide and it kept autocompleting text in a different writing style, some of which was obviously plagiarised (when I’m describing precisely how to implement a specific, and not very common, lock type with a futex and the autocomplete is a paragraph of text with a lot of detail, I’m confident you don’t have more than one or two examples of that in the training set). It was distracting and annoying. I wrote much faster after turning it off.

So, after giving it a fair try, I have concluded that it is both a net decrease in productivity and probably an increase in legal liability.

Discussions I am not interested in having:

You are holding it wrong. Using Copilot with this magic config setting / prompt tweak makes it better. At its absolute best, it was a small productivity increase, if it needs more effort to use, that will be offset.
This other LLM is much better. I don’t care. The costs of the bullshitting far outweighed the benefits when it worked, to be better it would have to not bullshit, and that’s not something LLMs can do.
It’s great for boilerplate! No. APIs that require every user to write the same code are broken. Fix them, don’t fill the world with more code using them that will need fixing when the APIs change.
Don’t use LLMs for autocomplete, use them for dialogues about the code. Tried that. It’s worse than a rubber duck, which at least knows to stay silent when it doesn’t know what it’s talking about.

The one place Copilot was vaguely useful was hinting at missing abstractions (if it can autocomplete big chunks then my APIs required too much boilerplate and needed better abstractions). The place I thought it might be useful was spotting inconsistent API names and parameter orders but it was actually very bad at this (presumably because of the way it tokenises identifiers?). With a load of examples with consistent names, it would suggest things that didn't match the convention. After using three APIs that all passed the same parameters in the same order, it would suggest flipping the order for the fourth.

#GitHubCopilot #CHERIoT

I have a favour to ask of a condensed matter physicist on behalf of a friend.

If you class yourself as a "condensed matter physicist", or "condensed matter physics adjacent", would you mind giving me a ping?

Thanks.

(PS: Please boost for reach ... much obliged)

#Physics #CondensedMatterPhysics

@felipe now this is a super interesting question that I do NOT have much insight into, maybe @flourn0 actually would since conscientiousness is often modeled as a personality trait in my limited understanding. The only geneticist stuff that comes to mind is Paige Harden's work? But I do not have the genetics background to really be informed here! I have seen some criticism of the entangling of self-regulation with conscientiousness and other complex trait stuff (https://pubmed.ncbi.nlm.nih.gov/36326625/ )

An NPR investigation found that high-ranking Louisiana health officials ordered staff to stop promoting vaccines for COVID, flu and mpox.

This is unacceptable and must not be tolerated. When political appointees at health agencies prevent public health workers from doing their jobs, it kills people.

https://www.npr.org/sections/shots-health-news/2024/12/20/nx-s1-5223440/louisiana-ban-public-health-promoting-covid-flu-mpox-vaccines-landry-rfk-jr-anti-vaccine

A very important point for parents of young children from Doctor of astrophysics Becky Smethurst: The International Space Station is a very bright light moving across the sky quite rapidly, and is easily confused with Santa's sleigh....
From ~2m 47s. Use this knowledge as you see fit.

https://youtu.be/KuTVjuXu7uE?t=167

Many Videos That Purport to Show Drones in New Jersey Likely Show Planes, Visual Analysis Finds
https://www.nytimes.com/2024/12/19/video/new-jersey-drones-planes-videos.html?utm_source=flipboard&utm_medium=activitypub

Posted into Home Page @home-page-newyorktimes

We are humbled and extremely grateful to the good humans of @SocialCoop who have decided to make a huge donation (https://opencollective.com/socialcoop/expenses/226852) of £500 GBP to support our project 🎉🎉🎉

We intend to put this to good use in our participatory budgeting process, funding federation members and allies to work on improving our software and self-management processes.

Thank you so much 💝💝💝

-- @decentral1se

@enoclue That’s the problem—auditing is so hard, even with full source. If nothing else, all it takes is one obscure bug that will let some attacker inject malware. Let me tell a true story to illustrate things…
Many years ago, I was working on a project when I learned that one team member had just been arrested and charged with hacking. I went to management and told them this; for my pains, I was told to organize an audit team. The first hole I found, and it was a serious one, was inadvertently introduced by an audit team member. (We had commit logs, but on top of that she remembered writing the code.) The next was worse. There was a bug that, coupled with a common configuration error, led to root compromise—and the comments didn't agree with the code, but for understandable reasons given the exact situation.
On even days, I think it was a subtle back door, but today is an odd day, so I think it was just an ordinary bug—and the hardest-to-locate bug I've ever found, at this point in almost 60 years of programming.

Report: US considers banning TP-Link routers over security flaws, ties to China
Founded in China, TP-Link makes routers popular in US homes and businesses.
https://arstechnica.com/tech-policy/2024/12/report-us-considers-banning-tp-link-routers-over-security-flaws-ties-to-china/?utm_brand=arstechnica&utm_social-type=owned&utm_source=mastodon&utm_medium=social

🎉 We have moved! 🚀
Our new Mastodon account is now on mastodon.social 🐘

Stay connected with us here: @MPI_Magdeburg

Wishing everyone happy holidays and a joyful season! 🎄✨

#HappyHolidays #MastodonMigration

If You’re #Pregnant, Here’s What You Should Know About the Medical Procedures That Could Save Your Life
—

#Women experiencing #pregnancy loss in states with #abortion bans told us they wished they had known what to expect and how to advocate for themselves. We created this guide for anyone who finds themselves in the same position.

#News #Health #Healthcare #Doctors #Miscarriage

https://propub.li/4gC4CaV

Following Elsevier's decision to raise the article processing charge for NeuroImage to $3,450, all editors (inc. chief editors) from NeuroImage and NeuroImage:Reports have resigned, effective immediately.

I am joining this action and have also resigned.

Full announcement: https://imaging-neuroscience.org/Announcement.pdf

The starlings seemed to be enjoying the relatively calm weather. And photographing Iceland in this weather is kinda like working in easy mode.

@FelisCatus wrote:

"But because these theories are so perfectly fine, each in its own domain of validity..."

I don't think the Standard Model of particle physics and general relativity are "perfectly fine" in their domains of validity.

First, nobody has given a mathematically rigorous formulation of the Standard Model, so it could be inconsistent, and we certainly don't know *for sure* what predictions it makes, though in practice we are very good at extracting predictions from it. We don't know if the Higgs destabilizes the vacuum, computations of the proton mass use a lot of ad hoc assumptions, etc.

Second, while general relativity is much closer to being mathematically rigorous, the overall features of the Universe as predicted by general relativity require dark matter, whose properties are mysterious - if it exists at all. If dark matter doesn't exist, general relativity is in trouble.

Also, general relativity probably breaks down when it predicts infinite curvature at the Big Bang and black holes.

I'm mentioning these problems not because these theories are bad, but because right now we *need* problems, to help us make progress.

"Which brings me (at last) to my question, namely is technological progress the current limit to contemporary theoretical physics (at least insofar as to seeking for a "theory of everything") ?"

We'll see! We are still improving our experiments to better understand dark matter, dark energy, neutrinos, etc. Maybe we'll make progress experimentally, but maybe not.

As mentioned there are also lots of problems with the Standard Model and general relativity, so we might make progress purely theoretically - but maybe not.

J. J. Thomson, who was born #OTD in 1856, received the Nobel Prize in Physics in 1906 for his discovery of the electron, the first subatomic particle to be found.

Thomson was also a teacher, and seven of his students went on to win Nobel Prizes: Ernest Rutherford, Lawrence Bragg, Charles Barkla, Francis Aston, Charles Thomson Rees Wilson, Owen Richardson and Edward Victor Appleton.

https://en.wikipedia.org/wiki/J._J._Thomson

Books by J.J. Thomson at PG:
https://www.gutenberg.org/ebooks/author/38322

#books #science #physics