ccooffee

Is the title here editorialized, or is Axios doing some weird AB testing? The article I'm seeing is about the bill passing the Washington House, and concludes with:

> What's next: To become law, the bill must clear Inslee's desk. But a veto looks unlikely.

I'm not the person you're responding to, but my "strong magnets aren't lethal to people" knowledge is based (in part) on a Veritasium YouTube video[0] where they visit the "World's Strongest Magnet" and do shenanigans.

[0] https://www.youtube.com/watch?v=g0amdIcZt5I

> given that the brain has an easier time fighting infections [than bones]

This doesn't match my understanding at all, but I'm not a medical professional. As far as I know, the brain relies heavily on isolation for protection. One function of the blood-brain barrier[0] is to prevent contamination of the central nervous system from pathogens and toxins. Bacterial meningitis is treatable with proper antibiotics, but still maintains a 10% mortality rate[1].

Additionally, simple mechanical implants (e.g. plates, screws, replacement joints) don't require interfacing with the body. They are often made of solid metals which are not bioreactive (e.g. titanium) or coated with bioresistant polymers[2].

[0] https://en.wikipedia.org/wiki/Blood%E2%80%93brain_barrier

[1] https://www.ncbi.nlm.nih.gov/books/NBK470351/

[2] https://en.wikipedia.org/wiki/Orthopedic_plate#Materials

> In one encouraging sign, some of the most recent attacks failed against users who had activated Apple’s recently introduced Lockdown Mode, which stops some communications from unknown callers and reduces the number of programs that are automatically invoked.

I'd not heard about Apple's "Lockdown Mode" (which applies to both computers and phones). Per a support page[1], there's a lot of small tweaks that reduce convenience of the devices, but it sounds like they will fundamentally still _mostly work_.

Is "Lockdown Mode" a well-known thing that I just missed? It feels like Apple has put a target on things like URL previews for malicious actors, while also labeling the defense mechanism as an expert-only tool.

[1] https://support.apple.com/en-us/HT212650

Wood takes a very long time to saturate enough to properly sink. A partially-floating log in the ocean is sometimes called a "deadhead" log (supposedly named because the bobbing motion of the lighter end of the log resembles a human head, but this feels apocryphal). I cannot find reliable sources for how long it takes for a log to fully saturate with seawater before sinking below the surface. Some sailors have reported seeing the "same" deadhead for over a decade. (While the deadhead is still at- or close-to-the-surface, it is a major hazard for boats. Deadheads have been known to punch through boat keels and destroy propellers and other equipment.)

After the logs become fully submerged and settle, I do not believe that they decompose rapidly, but (again) cannot find reliable sources. The state of Florida has a licensed program for "Deadhead Logging"[0], but much of the content I've found about that program is specific to river ecosystems, not oceans.

A 2020 article entitled "First evidence of microbial wood degradation in the coastal waters of the Antarctic"[1] offers a few nice quotes:

> Wood submerged in saline and oxygenated marine waters worldwide is efficiently degraded by crustaceans and molluscs.

> Our results establish that there is ongoing wood degradation also in the Antarctic, albeit at a vastly reduced rate compared to warmer environments. Historical shipwrecks resting on the seafloor are most likely still in good condition, although surface details such as wood carvings, tool marks, and paint slowly disintegrate due to microbial decay.

[0] https://floridadep.gov/water/submerged-lands-environmental-r...

[1] https://www.nature.com/articles/s41598-020-68613-y

This is such a vague comment to be useless for GP.

Do you have some issue with carbs? High calorie foods? Something against processed foods? Gluten? Plastic sensitivity? Why was it a blessing? What made you start appreciating it? etc. etc.

"Iron deficiency and sleep - A scoping review"[0] did a literature review and found many studies linking Iron Deficiency to poor sleep outcomes for various particular sleep disorders. It summarizes the findings with:

> iron investigation and supplementation should be considered in patients presenting with sleep disorders

[0] https://pubmed.ncbi.nlm.nih.gov/32224451/

According to an Apple support page[0],

> With Apple Watch Series 3 or later with watchOS 8, you can measure and track your Respiratory Rate.2 With Track Sleep with Apple Watch turned on, when you wear your Apple Watch to bed, it will automatically measure and record the number of times you breathe in a minute.

I assume that this information is enough to warrant "go see a sleep specialist", but possibly not much more than that. There is at least one watchOS app that claims to use the Apple Watch data to detect sleep apnea[1], though it may not be a completely reliable source[2].

With Apple's increasing focus on the health for the Apple Watch, I would not be surprised if future watchOS updates (or future Watch hardware) focus on sleep apnea detection as a flagship feature.

[0] https://support.apple.com/en-us/HT211685

[1] https://9to5mac.com/2022/05/10/napbot-sleep-apnea-analysis/

[2] https://www.reddit.com/r/SleepApnea/comments/vxkkxx/napbot_a...

Do you have an alternative recommendation for "strong" reminders?

"Counting five breaths" is a tool I got from some mindfulness resource. Focus on the breath as much as you can and count the exhales. When you get to five, begin again from one. Repeat until settled. If I'm fairly out-of-it, it may take me 10 repeats of counting five breaths to center myself. Sometimes it doesn't work at all.

That's a very good point. I'm currently suspecting that the numbers quoted in Blakewell's article are the "hot carcass" weight (i.e. subtracting bone, head, and organs from the total weight). An Aurochs (wild ancestor of domesticated cattle) could reach 1300 lbs in the 1500s (and possibly in excess of 3300 lbs prior to recorded history, according to archaeological finds)[0].

If the numbers are hot-carcass weight, we would be seeing:

1700: 370 lbs, 1789: 840 lbs, 2020: 880 lbs [1]

But, uhh, that doesn't feel right either. I think that wikipedia page could use some additional citations, but I'm struggling to find useful numbers with which to update.

[0] https://en.wikipedia.org/wiki/Aurochs

[1] https://beef.unl.edu/beefwatch/2020/how-many-pounds-meat-can...

I hadn't heard about rice fields emitting methane before. I was able to quickly find an article, "Methane emissions from global rice fields: Magnitude, spatiotemporal patterns, and environmental controls"[0], which states:

> Our study estimated that CH4 emissions from global rice fields varied from 18.3 ± 0.1 Tg CH4/yr (Avg. ±1 SD) under intermittent irrigation to 38.8 ± 1.0 Tg CH4/yr under continuous flooding in the 2000s, indicating that the magnitude of CH4 emissions from global rice fields is largely dependent on different water schemes.

At 18.3-38.8 Tg CH4/yr, this estimate comes out to about 500-1500 million tons of CO2-equivalent per year according to an EPA GHG-equivalence calculator[1]. Global emissions are estimated at north of 34 billion tons of CO2-equivalent per year[2], putting the total CO2-equivalent from rice fields throughout the world at 1-3% of total global emissions.

Wow.

[0] https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/201...

[1] https://www.epa.gov/energy/greenhouse-gas-equivalencies-calc...

[2] https://ourworldindata.org/co2-emissions

Well, they've got fancy procedures in there, too:

> So what our colleague Duyan Tao did - he, at an early stage, around 10 days after the [by-human-hand] pollination, went in there and dissected out the very young embryo that was developing and put it in tissue culture and gave it nutrients to help it grow. And then he rescued one of these seedlings. And that was the beginning of it.

Also, the "old-fashioned genetic engineering" of selective breeding of animals/plants is...not as old-fashioned as you may think. Animal husbandry has been around for millennia (or longer), but it was an undirected process prior to the work of Robert Bakewell in the 1700s. Thanks to his groundbreaking work (and a lot of refinements from others), humanity learned how to shape the trajectories of other species. I like this quote from Bakewell's wikipedia page[0]:

> In 1700, the average weight of a bull sold for slaughter was 370 pounds (168 kg). By 1786, that weight had more than doubled to 840 pounds (381 kg).

(Obviously, humanity has gone on to be much more aggressive and successful with selective breeding. Beef cattle are now usually 1200-1400 pounds at slaughter, for example. Though, personally, I'd say we've gone "too far" when you learn about the horrors of modern, industrialized animal agriculture.)

[0] https://en.wikipedia.org/wiki/Robert_Bakewell_(agriculturali...

I have no experience with LLMs, so here's some website feedback to be taken with a chunk of salt:

1. The Youtube video at the bottom of your page is very tiny and cannot be fullscreened without first loading the video in Youtube. The video itself merely shows some basic-seeming workflows with some (to-me) terrible background music. The video does not seem to emphasize anything. It's just...wandering around a web application...

2. Between reading the webpage content and watching the video, I don't have a good idea of what you are actually offering as a product and why it is so valuable. The pitch summary in this HN post is much more helpful than your website.

3. Your website is not very accessible at the moment due to low contrast and overuse of opacity for style. I can barely understand what your images are attempting to convey. Your app doesn't appear very accessible according to the Youtube video, again due to low contrast among colors.

Is this a categorical question? I.e. "is it possible to create a 3D twisty puzzle that works on computers but not in the real world?"

I can easily imagine a 3D twisty puzzle which allows self-intersection. (To be less cheesy, you could make the self-intersection an objective of the game, perhaps "same color cubes will combine into one cube when they intersect" and then trying to accomplish a final target 3D model.)

I cannot imagine a computer simulation being more robust than reality without relaxing the rules of the physical world (like exclusivity of space/extent). Perhaps something like the Banach-Tarski Paradox[0], where a mathematical sphere can be split into two identical-to-the-first spheres?

[0] https://en.wikipedia.org/wiki/Banach%E2%80%93Tarski_paradox

I had been unaware of measurement according to "moles of photons", though I suppose it's not surprising. I've never really understood what a mole is other than "we decided to pick this number as a constant multiplier when doing small calculations".

Per wikipedia[0], there's a vaguely defined unit, the Einstein, which may be defined as the energy in a mole of photons. (The vague definition being because each photon may have different amounts of energy, and thus an Einstein would be some weird function in order to describe total energy.) Wikipedia suggests using measures of Photosynthetically Active Radiation (PAR)[1], like Photosynthetic photon flux (PPF) instead. I suppose this is because PAR is literally defined to measure according to "what plants crave", but it also allows bounding the "total joules of energy" above and below by the PAR wavelength limits.

[0] https://en.wikipedia.org/wiki/Einstein_(unit)

[1] https://en.wikipedia.org/wiki/Photosynthetically_active_radi...

They aren't hiding that, though? Literally the first graphic on the page shows their claim of a "word error rate" on test data has an error rate of around 14% in the best case compared to the state of the art at 15% (for en-US content).

It's only 1% better than the current state of the art. But it's still noteworthy. From the end of the abstract:

> We demonstrate that utilizing a large unlabeled multilingual dataset to pre-train the encoder of our model and fine-tuning on a smaller set of labeled data enables us to recognize these under-represented languages. Moreover, our model training process is effective for adapting to new languages and data.

It's amazing to me that the chaotic process of "machine learning" can end up with an internal state for languages that is readily adapted to entirely new languages.

For now, they've got this handling audio transcription, but with some hints that this approach could work well for translation. Perhaps we'll be able to use these improved models to decipher Linear A[0] or other un-deciphered languages. It sounds like "magic", but it's the kind that could maybe exist.

[0] https://en.wikipedia.org/wiki/Linear_A

I found this webpage[0] (and those it links to) to be a good bit of context on what this all means. (And then ran off to do a bunch more searching!)

tl;dr: CMS (Compact Muon Solenoid) is one of two major detectors of the Large Hadron collider. Fermilab was intimately involved in designing CMS. Fermilab is also a "tier 1" computation center for processing LHC data (possibly restricted to CMS data?). Per this factsheet[1], there is one other "tier 1" compute farm in the US, though it isn't named in particular. I'd guess Lawrence Livermore, but I have nothing more than a hunch.

[0] https://cms.fnal.gov/

[1] https://www.fnal.gov/gridfest/pdfs/uscomputing_factsheet.pdf

This is a good summary of the author's point. In a historical context, it may be the case that the only useful way to enforce "quiet skies" was to put a speed limit, but that doesn't account for changes in technology that may enable quiet booms.

Most commenters in this thread seem to take issue with the author's assumption that there is no other justification for a speed limit other than noise reduction.

Honestly, I find the author's premise to be vaguely reasonable, but the whole thing is undermined by the god-awful chart which extrapolates historical aircraft speeds to imply that without the "speed limit" rule, we would have commercial aircraft crossing the Atlantic Ocean at 2500 mph. (Per wikipedia[0], the fastest manned aircraft in history, a SR-71 blackbird flight, capped out below 2200 mph.)

[0] https://en.wikipedia.org/wiki/Flight_airspeed_record

Your comment reminded me of a Youtube documentary about "the missing American particle collider"[0]. The documentary touches on some Fermilab drama, but most of the content is aimed at politics (and in particular, how presidential politics is unfortunately intertwined with funding for science mega-projects in the US).

[0] https://www.youtube.com/watch?v=ivVzGpznw1U&list=PLAB-wWbHL7...

Short certificate lifetimes (e.g. 1 hour) is not valid-for-a-single-request as the GP asked.

I'm having a real hard time wrapping my head around how Public Key Infrastructure could co-exist with every public key being a nonce. I'm not confident that it is impossible, but GP's question seems like an interesting theoretical/categorical question more than a hyperbolic "how short can lifetimes go?".

1 hour lifetimes sounds incredibly complicated to orchestrate on a practical level. Do you use a lot of short-lived ephemeral hosts (e.g. a swarm of docker images)? I'm not sure how 1 hour lifetimes wouldn't cause systemwide chaos in what I consider a "typical" microservice architecture.

The non-profit-oh-wait-just-kidding nonsense feels like an existential risk to OpenAI. I agree that Altman may be (unintentionally?) protecting himself from lawsuits, but I don't see it being enough to avoid the spotlight.

I expect someone will eventually sue OpenAI for the shenanigans, and that a competitor (Meta?) or Elon Musk will fund the lawsuit. The goal won't be to actually win the case, but instead to motivate US regulatory action to undercuts OpenAI's perceived advantage.

I don't understand how ChatGPT (or similar interactive agents) function, so I'm very confused by your experience. Was ChatGPT just a replacement for internet searches (e.g. instead of searching for a spec sheet for your model microphone, you asked the bot "what is the <spec> for <model> microphone?")? Or was ChatGPT doing the debugging with you just providing data from the real world?

"Ever" is an exceptionally poor word choice in Zuckerberg's email. Imagining a scenario where something bad is being done (intentionally or not) isn't that difficult. I believe there are circumstances that NOT whistleblowing would be a moral failing.

"But no realistic situation would require you to leak" is not a good strawman defense for Zuck's word choices. There is a broad history of whistleblowing in America being the necessary first step to unravel bad things. You don't even have to leave Silicon Valley to find Theranos, the Google/Apple/Facebook no-poach collusion, etc.

This is still specialist/amateur grade stuff (requiring you to build/maintain/debug the system yourself). I'm optimistic about the technology continuing to develop and eventually becoming mainstream. Having a robust alternative to keyboard-and-mouse interactions will (hopefully) improve accessibility of software. (Though I have difficulty imagining that BCI input devices will be easier to support than today's complicated accessibility approaches, like screen-reader compatibility.)

Thank you for sharing this (novel-to-me) viewpoint. I've been attracted to permaculture by the calm-but-optimistic tinkering that so many content creators in that space espouse.

Is there a good starter resource you would point to as being more in line with modern ecological science? Thinking about my yard as an environment in the local ecology makes yardwork/gardening feel like it matters (much more than my terrible tomato yields, at least).

I am regularly surprised at how "cheap" mega-engineering projects are compared to the things we (USA) already spend money on.

Here's some examples of expensive things:

* USS Gerald R Ford aircraft carrier: $13.3 billion [0]

* F-35 jet program (covering expected lifetimes of all aircraft through 2070): $1.5 trillion (> $150 million per aircraft) [1]

* TARP funds during 2007-2010 crisis: $431 billion [2]

* PPP loan forgiveness (during covid): $742 billion [3]

It's very tempting to declare these other expenses an exorbitant waste of potential when mega-projects come in so much cheaper: greening Africa ($8bn), using desalination and up-gradient pumping to re-hydrate the Colorado River ($15bn)[4], the large hadron collider ($5bn - $13bn)[5], and James Webb Space Telescope ($9.7bn)[6] to name a few.

Sources:

[0] https://www.businessinsider.com/cost-of-navy-uss-ford-aircra...

[1] https://web.archive.org/web/20191206140954/https://www.thena...

[2] https://en.wikipedia.org/wiki/Troubled_Asset_Relief_Program

[3] https://projects.propublica.org/coronavirus/bailouts/

[4] https://caseyhandmer.wordpress.com/2022/11/20/we-need-more-w...

[5] https://en.as.com/latest_news/how-much-money-did-cerns-large...

[6] https://www.planetary.org/articles/cost-of-the-jwst

I've eaten seaweeds before in several formats, and they're surprisingly versatile. But they're "further out there" than kale, so I appreciate that the nytimes didn't want to push that angle.

And, in case you're wondering "heavy metals bio-accumulate in ocean fish to questionable levels; could heavy metals bio-accumulate in seaweed to a dangerous level?", tl;dr: probably not to dangerous levels, but not certain.

"Distribution of metals and metalloids in dried seaweeds and health risk to population in southeastern China"[0]:

> Concern about metals and metalloids, especially heavy metals in seaweeds has risen due to potential health risk. This study investigated the distribution of 10 metals and metalloids in 295 dried seaweeds (brown and red) and estimated the possible health risk via hazard index (HI). > ... > The estimate of health risk showed that there was low health risk for potential toxic elements by intake of these seaweeds.

(I've gotta say, that article doesn't seem entirely convincing. This is probably the first article of this format I've read before, so I might be totally off-base, but the error bars in their data[1] are huge. Given that the article literally has "health risk to population in southeastern China" in the title, I'm skeptical whether the article is intentional fluff/propaganda science.)

[0]: https://www.nature.com/articles/s41598-018-21732-z (full text) [1]: https://www.nature.com/articles/s41598-018-21732-z/tables/2 (data)

Under 18 U.S. Code § 1702[0], it would appear to be illegal on-its-face for USPS mail to be inspected by prison authorities.

> Whoever takes any letter...before it has been delivered...with design...to pry...or opens...the same, shall be fined under this title or imprisoned not more than five years, or both.

However, there's certainly more to be said about this, as the ACLU's prisoner rights page[1] has this to say about mail in prison:

> The First Amendment of the Constitution entitles prisoners to send and receive mail, but the prison or jail may inspect and sometimes censor it to protect security, using appropriate procedures.

I've been unable to find any resources identifying under what statutory authority a prison may inspect mail, but I'm unfamiliar with legal stuff, so I think this is a "me-failure". I did find a 2021 article from The Intercept[3] which was informative on giving a similar-perspective, different-author from the submitted ACLU page.

[0] https://www.law.cornell.edu/uscode/text/18/1702

[1] https://www.aclu.org/know-your-rights/prisoners-rights#i-wan...

[2] https://theintercept.com/2021/09/26/surveillance-privacy-pri...

Thank you. This is an actual article with details, not just a few photos in some social media website.