This article makes some good points but is missing a lot of the larger context. It doesn't use this phrase, but the topic it discusses is called the linear no-threshold (LNT) model:
Whether and how to replace the LNT model is one of the most controversial topics in health physics. The most basic modification to the model is to assign a threshold, but for that we need better research to establish a threshold for various circumstances. In the US, the Department of Energy started a research program to study low doses of radiation in 1999:
The US agency that regulates radioactive emissions, the Nuclear Regulatory Commission (NRC), still uses the LNT. The NRC is a conservative agency (in the sense of being averse to change) and they are required by law to regulate for protection of the public. They won't change regulations if there is no consensus on what to replace LNT with. They were petitioned to change it in 2015 and rejected the petition in 2021:
A majority of adults in the US now favor nuclear power, so there is some basis for cautious optimism of continued research and potential regulatory changes in this area:
The test of a technology is how it's actually used in society. What it theoretically could do in the abstract is immaterial.
For example, BitTorrent is a great peer-to-peer file-sharing protocol. In principle you could use this for legitimate media distribution, and in fact Rainberry Inc. (a.k.a BitTorrent Inc.) tried to do this for decades and succeeded in some one-off partnerships with legit broadcasters:
In practice BitTorrent as a protocol is still mainly used for pirated video files, the same as it was 20 years ago. Meanwhile BitTorrent the company was bought by a cryptocurrency startup in 2018 and laid off most of its employees in 2023.
Alongside the debut of Oura's smaller Ring 5, the wearables company on Thursday unveiled its plans to help users integrate health records and translate their ring's findings into medical care.
Oura has been building on its healthcare business for a little more than a year, hiring a chief medical officer and partnering with healthcare companies like Maven Clinic and most recently Resmed. Thursday's announcements build on that work by incorporating AI-powered doctor's visits through a partnership with Counsel Health and the ability to input health records.
The push further into healthcare comes at a critical time for the company: Valued at $11 billion, Oura said on May 21 that it had confidentially filed to go public.
"What we imagine here is a new way of thinking about health and being not only proactive, but reducing that barrier for members for everyday care," Maziar Brumand, Oura's VP of Product, told Endpoints News.
For a subscription fee (the price of which has not yet been determined), Oura users can go through the Oura app to Counsel Health, a startup that offers medical AI services as well as visits with doctors. That way, an Oura user picking up on a change to their blood pressure (one of two new measurements Oura is launching Thursday) could speak directly with a physician and get care. It's one of the first times a wearables company has integrated care directly into its platform.
"There is just this huge consumer desire to tie these metrics to real clinical things," Counsel medical chief Rishi Khakhkhar said. "I think it was inevitable that these two worlds would eventually merge," he added.
Khakhkhar said that there are three reasons why it's now possible for wearables to link with clinical care. For one, the hardware is getting better and closer to the biometrics clinicians use in their medical practices. AI is also making it easier to integrate wearable data directly into Counsel's electronic medical record and analyze the volume of data wearables can amass between visits. And there's more comfort with asynchronous care that can be done via messages rather than a video call.
The partnership with Oura is also one of the first public partnerships for GV- and Andreessen Horowitz-backed Counsel, which works in partnership with other healthcare organizations like health plans and health systems to supply medical AI and virtual care via physicians.
As part of the Ring 5 launch, which has a starting price of $399, Oura also set up a way for users to input their medical records. Done in collaboration with healthcare data company Flexpa and drawing on the TEFCA framework that allows individuals access to their health information, Oura users can have their medical records pulled for them, rather than manually inputting the information themselves.
> Most writing is about getting words down in a structure that makes sense, and then getting those words in front of other people. Markdown does that with less friction than anything else ever created.
I agree with the overall sentiment, but it's important to remember that in 2004 when Markdown came out it was just one of dozens of lightweight markup languages competing for mindshare on forums and other websites with commenting systems in the late 2000s. Markdown was adopted by popular sites like Reddit, Stack Overflow, and GitHub which helped it win.
Over a decade before Markdown there was setext on Usenet and plain-text email, both of which influenced Markdown. In a sense Markdown represents the continuing influence of plain-text email on our communication even as most emails sent today are HTML.
For those interested in a kind of retrospective about 40 years after Feynman's speech, read "Physics in Latin America Comes of Age" (published in 2000) by José Luis Morán‐López:
> At the end of the 20th century, a large “science gap” still exists between Latin America and the developed countries of the North.
> The description is not intended to be a complete analysis, but may give a sense of the significant development that has occurred in the past half century and of what might be needed to make the 21st century a flourishing epoch for science in Latin America .
> The most developed group includes Brazil, Mexico, and Argentina, which have, respectively, about 3000, 2200, and 2000 PhDs involved in physics research.
Feynman, of course, always had confidence in the ability of the people of Latin America to do good physics. In fact his mentor Manuel Sandoval Vallarta was born in Mexico and emigrated to the US to study at MIT. Emigration to the US or Europe is typical of successful physicists from Latin America, including Juan Maldacena, a theorist from Argentina who discovered the AdS/CFT correspondence and has been a professor at the Institute for Advanced Study since 2001.
Anecdotally, I think Europe has more opportunities these days. My friend Gustavo, a high energy theorist from Brazil, got his PhD in the US but now works at the Oskar Klein Centre for Cosmo Particle Physics (OKC) in Stockholm.
Documents written in the 1980s in LaTeX still compile and look great today. Good luck doing that with an old MS Word file, especially if it has equations in it.
> And I found a fascinating pattern: the AI gives artificially high scores to reports written with AI [...] it was giving very high marks to poorly reasoned, error-filled work simply because it was elegantly written. Too elegantly... Clearly written with ChatGPT.
This is an interesting phenomenon, but I would have liked to see some quantitative evidence for this N=24 sample, e.g. would a paper ordinarily get an 80% score but the LLM gives it a 95%?
I also wonder how accurate a professor's perception of style is. I tend to write in a formal style, even in online forums like this one, and I wonder if people assume I use LLMs as a result (I don't).
Reddit profile: https://www.reddit.com/user/bariumbitmap/