jcrawfordor

The FAA contracts virtually all of its communication infrastructure now, so they probably don't even know much about these details. L3Harris has the contract for microwave radio but the fiber stuff seems scattered depending on when it was built.

I would expect this system to ultimately be used under a commercial HF license rather than an amateur license, given the feature set and that it seems to have much more in common with military systems like ALE than amateur practice. That said, because of the scarce supply of HF bandwidth, commercial HF licenses are quite restrictive in the US. For the most part, you are only allowed to apply for one if you are an infrastructure operator that needs a contingency communications capability for disaster response. That said, the FCC has permitted experimental licenses for some commercial applications like high frequency trading, so perhaps the situation will become less restrictive over time.

That said, commercial licenses are not per-operator like amateur licenses, so you would potentially only need one to cover the entire network. Some of this is ultimately negotiated.

There are definitely factors beyond IP being used. A week ago I found that all requests from Chrome-like browsers got a 429 across more than a half dozen networks and several machines, while Firefox reliably worked. I assume this was an overzealous policy on UAs.

Yesterday I discovered that WayBackMachine appears to be serving 429s to all Chrome-based browsers, across all network connections. Using Firefox (or a firefox UA) fixes it. I assume this is a temporary situation, or even by mistake, but underscores that archive.org is getting creaky and hard to actually use.

This article is just an over length summary of a very short preliminary notice published by NRC. Since it is non-emergency there will not necessarily be much follow up either, plants are just required to report any abnormality to NRC which publishes the reports online. Most of these reports are, like, procedural problems in employee drug testing.

Control rods are designed as fail safe so lots of scenarios can cause them to drop, could be a mechanical failure but probably more likely it was triggered by an electrical fault or intentionally by a safety system due to some condition they haven't quite worked out yet. Since having control rods in puts the reactor out of it's approved operating envelope it will lead to a trip out of an excess of caution. The fact that the trip wasn't triggered automatically suggests to me that the rods probably weren't inserted by a safety system but due to some fault.

this was already hitting 200% of its original intended length, so there were some things where I said "just not going to get into the details." Naturally I have come to regret these. Another is the brief mention of party line practice which is more applicable to GTE than AT&T (harmonic ringing), and in general mentioning that GTE and other independents continued leasing post-1982 as well but not going into any detail. The fate of the leased telephones of independents appears to be far more along the lines you would expect, with many of the independents quietly wrapping up leasing programs in the 1990s and letting people keep the phones. That said there were lots of exceptions and odd details.

And that's all stuff that they bring in out of state contractors for. I know, it sounds odd, but this is pretty normal for any large industrial site.

For what it's worth, while I haven't found kdenlive (or shotcut, based on the same underlying toolkit) to be 100% stable, I've had significantly fewer lost-work incidents with kdenlive than I did with Premiere Pro. The frustration of Premiere's instability was the main thing that drove me to open-source software.

I've never used Resolve primarily so I don't have a good feeling of how they compare, but I have experienced a couple of unexpected, mid-work crashes in Resolve as well. I believe these were tied to my working on a machine with an Intel iGPU, which at least at the time seemed to be... discouraged, I'll say, by the Resolve community due to known stability issues. Possibly the root of evil with Premiere as well, but again, doesn't seem to be a major problem for kdenlive.

What I will say is that I personally prefer Shotcut to kdenlive. Both are basically graphical frontends to MLT, the actual media toolkit/editor (driven by XML files). Shotcut has a simpler, more user-friendly UI than kdenlive and also seems to be a bit more stable/performant. kdenlive is more featureful. I think most people should try both because it probably depends on your workflow which is more convenient.

The number the developer gave in a press release was "20-30." I find that reasonable as a very large Facebook data center near me has a permanent staff of around 50. Keep in mind that these large DCs use contractors for the majority of the work, which unfortunately doesn't really help with employment because the contractors mostly come in from out of state (there is a HUGE temp labor market for traveling IT technicians and skilled crafts get hired mostly from big national outfits that just send whatever crew is available next). It is good for the hotel business though.

Very odd. Besides the Turbotax-esque over length loading process, every point I could find (in the Southwestern US) was a seemingly random location with no other details. Some of them just, like, the side of a road. I think the whole database was generated by LLM and it's just useless junk.

TSA does more than just the security checkpoints, even at airports with privatized screening TSA does all of the back-office work, including some on-site staff. The physical screening is the only thing that can be contracted out, not the whole rest of the process like maintenance of risk databases.

What might confuse things a bit is that this incident happened hours before ICE agents started reinforcing TSA at checkpoints and seems mostly unrelated, other than establishing the general principle that ICE will arrest people at airports based on tips from TSA's flight booking records.

First, you're responsible for knowing where you live. Historically, people who lived in more challenging areas geographically often did not have regular postal addresses at all. You would just have a box number in the nearest town or a rural route stop number, and these obviously didn't reflect the legalities of where you lived. In our modern world, USPS has adopted a policy of 100% physical addressing, meaning that all properties now have a "real" address even if the number part is scaled from mileposts (as is the case in rural areas). Still, I think people who live in areas where any of this is less than obvious understand the nuance that how USPS handles addresses is not necessarily the same as how the voter registration clerk handles them.

Still, it is rarely a problem in practice, because anyone relying on addresses to establish these legal details will have to look at where the address is actually located---not just the city written in it. Keep in mind that quite a few people live in ZIPs where they could write multiple city names in their address.

When it comes to the unusual case of spanning states, it might help to note that the City State Database the postal service uses to validate addresses does not actually differentiate between city and state. "NEW YORK NY" is a single string. The state is really just part of the city name. The fact that USPS implemented it this way indicates the extent to which it does not matter in operational reality.

A sibling comment points this out, but it might also help to explain that in the US, it is very common for people to have mailing addresses in cities they do not actually live in. That's because of suburbs. City lines are often surprisingly arbitrary and reflect complex political histories. Many people consider themselves to live in [major city] but, in legal actuality, live in [unincorporated county that contains major city]. Many of the upsides of living in the city, sans some of the property taxes and voting in city elections! Yet another reason that people understand that mailing addresses are not definitive reflections of political boundaries.

If you ever work as an election clerk you will find this a LOT---people indignant that they cannot vote for the mayor, to whom you will have to explain, somehow for the first time, that they do not actually live in city limits. This tends to be more obvious if you get a property tax bill but a lot of people are renters and never really think about that aspect.

Unfortunately these USPS datasets are not public because USPS sells them. Or in some cases, the pattern tends to be that USPS has a contract with a provider (part of what I call the Postal Industrial Complex) that maintains the database and then sells it to both USPS and everyone else. Since these databases are used primarily by bulk mail services, they're fairly expensive and represent an important revenue source to USPS. Remember that USPS is semi-privatized, so they're looking for fees they can charge like everyone else... especially fees that can be changed more easily than postage rates.

That said, the ZIP DB is indeed not very large, so you can find copies of it. You won't generally find complete copies of the City State file but I wouldn't be surprised if there is one out there.

USPS doesn't care; each ZIP code has a single preferred city name and a list of acceptable alternate city names to account for cases in which a ZIP code spans multiple cities. However, USPS's address validation will prefer to use the preferred city name for the ZIP regardless of whether the recipient actually lives within the boundaries of that city. That's because USPS has opted to organize addressing entirely around the ZIP codes, and other political boundaries are irrelevant except in cases of problems interpreting the address.

This does mean that you might autofill a city name that is "wrong" in the view of the person completing the address form, but much of the bulk mail they receive probably uses that city name anyway.

Technically speaking ZIP codes are not "supposed" to span states but, in exceptional cases, some do. In this case USPS handles it the same way: the state of the preferred city is the preferred state for the ZIP code.

The preferred city is almost always the location of the post office serving the ZIP, which makes this situation fairly intuitive. You can find some interesting edge cases where a post office in located in a suburb city, resulting in a ZIP that includes part of a major city having the suburb as its preferred city name.

You can look up the city name and alternates for a ZIP here: https://tools.usps.com/zip-code-lookup.htm?citybyzipcode and the Domestic Mailing Manual covers this, although it's scattered across several sections and mostly part of how the City State database (the database used for validating city and state names in addresses) works.

I've been peripherally involved in an early stages effort to build something similar for the entire nation (https://reportapayphone.com/), and became aware of this just recently. It's a really great example to aspire to, in terms of the level of polish. I do find the time limiting odd; our goal is to identify as many payphones as possible this way.

Unfortunately the state of payphone-related records is extremely poor, with many ostensibly-active PSPs having quietly gone out of business, other PSPs reorganized without reregistering, and states themselves keeping PSP records very poorly. Throw in small-scale COCOT operations and the result is that there really isn't any authoritative database of possible payphones, so this website's map is going to be missing some. It will also include many that are nonfunctional, as today's PSPs seem to do close to zero maintenance and out of service phones stay that way for years.

Some of the nation's largest PSPs have become ghosts, with the phones still operating and able to accept payment, but the PSP completely unresponsive to efforts to contact them. It's a very strange afterlife.

Biometrics were a very crowded market during the 1980s and 1990s when it was a newer idea and electronics were starting to make things practical. Lots of ideas were tossed around before the industry pretty well consolidated on fingerprints with a side of iris imaging and hand geometry in some more security-sensitive niches. It mostly came down to cost: fingerprint scanners, even before the modern capacitative type, came down in price much faster than other types of imaging (visible rather than IR sensors, glass platen allowed for fixed focus, etc). The widespread use of fingerprint comparison in criminal forensics also mean that there's an older and stronger academic literature on fingerprint comparison, whereas other types of biometric sensors often involve proprietary match algorithms and you have to rely on the vendor's assertions about reliability.

Of course everything around cameras has come down in cost tremendously since then, so palm imaging is probably reasonably priced now, but it lacks a clear enough advantage over better-established methods for anyone to switch over. Besides, just the fact that you have to position your palm the way you do makes it difficult to install them in most practical door situations. Fingerprint sensors turn out to be very compact and fairly intuitive to use.

I scoured Amazon's sales materials around Amazon One very closely, because I found it fascinating that they were seemingly trying to revive the technique. I was surprised they were doing it as a payment device, but it made more sense when I found materials (I think old FCC filings) that suggested that it was originally designed as an access control product and perhaps "pivoted" to payments later. The strangest thing about it though was how unconvincing the sales materials were, it felt like they were really grasping at straws for a reason to select it over other options.

From what I could find it doesn't appear to have been an acquisition; the regulatory paperwork was all filed under some LLC but it seemed to just be a front company for Amazon which is fairly common for that kind of thing. So my best guess is that it was a pet project of someone influential enough to burn some R&D on it, and maybe pivoting to payments and putting them in Whole Foods was thought to maybe be the hail Mary that would turn it into a real business.

The actual integration with the PoS in the stores was clumsy too, they Velcro'd an NFC antenna to the side of the credit card terminal to use to make payments by proxy card. I originally got obsessed with it because I was trying to ID the suspicious device Velcro'd to the payment terminals at Whole Foods!

I don't know that there is much technology to sell, palm vein imaging is decades old in the access control industry. The reason you don't see it anywhere is because it was already a commercial failure in that application, by the end of the 1990s.

Amazon was even trying to sell the technology for access control applications, but their sales material were remarkably devoid of any reason to choose it over other biometrics.

Good data is hard to come by, but from a brief survey electronic precinct tabulation (the most common system in the US) is also in at least partial use in Canada, Mexico, India, the Phillipines, and Russia, and a laundry list of smaller countries.

Now, you might contend that this is not a list of first-world countries exactly (but rather I sampled the largest countries). You must keep in mind that the use of electronic tabulation in the United States is mostly a response to the very limited budget on which elections are carried out; electronic tabulation is much less expensive than significantly increasing staffing. As a result, globally, electronic tabulation tends to be most common in poorer countries or countries with newer election systems, while hand tabulation is most common in wealthier countries with long-established election procedures.

For this reason, the countries you might go to for comparison (like France and Germany) have largely manual election processes that have often seen few changes since the Second World War.

The Help America Vote Act (2002) had a de facto effect of making the United States a country with much newer election processes, as HAVA requires strict accessibility measures that most European election systems do not meet (e.g. unassisted voting for blind and deaf people). Most US election systems didn't meet them either, in 2002, so almost the entire country had to design new election processes over a fairly short span of time and on a shoestring budget. Understandably, election administrators leaned on automation to make that possible.

It's also important to understand that because of the US tradition of special-purpose mill levies and elected independent boards (like school boards), the average US ballot has significantly more questions than the average European ballot. This further increases the cost and complexity of hand tabulation, even ruling out entirely the "optimized" hand tabulation methods used in France and Ireland.

This is the system used in the majority of the United States. Direct-recording electronic voting systems were never that common, briefly peaked after the Help America Vote Act as the least expensive option to meet accessibility requirements, and have become less common since then as many election administrators have switched to either prectinct tabulators or direct-recording with voter-verified paper audit trail.

In the 2026 election, only 1.3% of voters were registered in jurisdictions that use direct-recording electronic machines without a voter verifiable paper audit trail (https://verifiedvoting.org/verifier/#mode/navigate/map/voteE...). 67.8% of voters are registered in precincts that primarily use hand-marked ballots, and the balance mostly use BMDs to generate premarked ballots.

There have been diesel electric surface ships as well going back to WWII, although it hasn't proven a very popular design and they remain an oddity.

To be quite honest I've never seen a colo that didn't offer access at all. The cheapest locations may require a prearranged escort because they don't have any way to restrict access on the floors, but by the time you get to 1/4 rack scale you should expect 24/7 access as standard.

It's very confusing in the text of the article, at times it sounds like the author is using heuristic methods (like timings) but at times it sounds like they somehow have access to network traffic from the provider's backend. I could 100% believe that a ton of these companies are making API calls to providers directly from an SPA, but the flow diagrams in the article seem to specifically rule that out as an explanation.

I might allow them more credit if the article wasn't in such an obviously LLM-written style. I've seen a few cases like this, now, where it seems like someone did some very modest technical investigation or even none at all and then prompted an LLM to write a whole article based on it. It comes out like this... a whole lot of bullet points and numbered lists, breathless language about the implications, but on repeated close readings you can't tell what they actually did.

It's unfortunate that, if this author really did collect this data, their choice to have an LLM write the article and in the process obscure the details has completely undermined their credibility.

A major reason for the enduring use of satellite in radio distribution is that, for live events like sports or (more common in NPR's case) political events, the satellite system provides appreciably lower latency than distribution over the internet. Reduced jitter also allows for generally higher reliability, you never hear the radio station buffering. There are options for low-latency land-based connectivity but at the scale of PRSS, the satellite system is cheaper to operate.

Most stations can also receive this programming over the internet, another reason for the satellite system is that it provides a completely redundant path for programming delivery. This is important for general reliability but especially so in an emergency.

Historically, radio networks distributed their programming over leased telephone lines. Satellite took over because it was cheaper. That gap has probably narrowed as terrestrial communications infrastructure continues to expand, but the internet struggles with low-latency real-time media, and an arrangement like leased fiber wavelengths to member stations would still be more expensive than the satellite system. There's a lot of member stations in a lot of places, satellite reaches all of them at once.

I'm not an expert in this topic but I've been working on a book in a related area and had to learn a lot. Here's what I can figure.

Unfortunately radiation medicine is pretty complicated and the report gives us very little info, presumably mostly because they don't have very much info. It will take some time and effort to establish more.

What we do know is that they measured 300 CPM at the person's hair, which was probably where they expected the highest count due to absorbed water (likely clothing was already stripped at this point). CPM is a tricky unit because it is something like the "raw" value from the instrument, the literal number of counts from the tube, and determining more absolute metrics like activity and dose requires knowing the calibration of the meter. The annoying thing here is that radiation protection professionals will still sometimes just write CPM because for a lot of applications there's only one or a handful of instruments approved and they tend to figure the reader knows which instrument they have. Frustrating. Still, for the common LND7311 tube and Cs137, 300CPM is a little below 1 uSv/hr. That wouldn't equate to any meaningful risk (a common rule of thumb is that a couple mSv is typical annual background exposure). However, for a less sensitive detector, the dose could be much higher (LND7311 is often used in pancake probes for frisking because it is very sensitive and just background is often hundreds of CPM). Someone who knows NRC practices better might know what detector would be used here.

That said the field dose here is really not the concern, committed dose from ingesting the water is. Ingesting radioactive material is extremely dangerous because, depending on the specific isotopes involved, it can persist in the body for a very long time and accumulate in specific organs. Unfortunately it is also difficult to assess. This person will likely go to a hospital with a specialty center equipped with a full body counter, and counts will also be taken on blood samples. These are ways of estimating the amount of radioactive isotopes in the body. In some cases tissue samples of specific organs may be taken.

I believe that the cavity pool water would be "clean" other than induced radioactivity (activation products from being bombarded by radiation). Because water shields so well the pool should not be that "hot" from this process. Most of those products have short half-lives which, on the one hand, means that they deliver a higher dose over a shorter period of time---but also means they will not longer forever and are less likely to be a chronic problem if they are not an acute one.

I suspect this will get some press coverage and we will perhaps learn more about the patient's state.

Another way we can get at this question is by the bureaucracy of the notification. An 8-hour notification as done here is required in relatively minor cases. Usually for a "big deal emergency" a one-hour notification is required. The definition of such an emergency depends on the site emergency plan but I think acute radiation exposure to a worker would generally qualify.

KCNSC is a large organization that will have hundreds of distinct networks at different risk and control levels. Every variation of "public internet" to "single-site air-gapped network" probably exists there, including many levels in between like multi-site secure networks and networks with limited internet connectivity. Many networks air airgapped, this sometimes means that they consist of a small number of assets in a single room, and it sometimes means that they have connectivity to airgapped enclaves of AWS and hundreds of other military, government, and contractor sites. All of these controls will have been determined by a combination of risk scoring, compliance policies, legal requirements, office politics, and happenstance. Multiple contracting authorities will periodically audit many of these networks against various standards, which may or may not allow connectivity to specific other networks depending on risk levels. Connectivity between networks is sometimes controlled by NSA accredited cross-domain solutions and multi-level security systems that enforce complex policy, in other cases it's controlled by an administrative assistant with a DVD burner. There will be case-by-case risk analysis decisions made for specific systems, ultimately signed off by a government official who may or may not have read them. Inevitably some of these will appear reasonable and cautious in retrospect and others will not.

The root fault with this article, and the resulting discussion, is the extent to which it generalizes over one of the larger organizations in a very complex part of the defense industrial complex. Many parts of KCNSC's operations are absolutely not exposed by this incident. Other parts absolutely are. Determining which fall into which category, and to what extent that is acceptable, keeps quite a few people employed.

When the tower was constructed in 1887, multiplexing technology was probably not available (I'm not so sure of the timeline in Europe). By 1913 it likely would have come into use. However, multiplexing really isn't a factor here, as the tower seems to have been built to serve local loops. Since these loops go to subscriber telephone sets, there's no option for multiplexing without expensive and maintenance-intensive equipment at customer premises. Multiplexing of local loops is called "pair gain" and wouldn't be developed until later, and it was never really that popular in most phone systems. Outside of suburban areas, it's typical that each copper pair runs directly to the exchange. Historically, and today, there is rarely any active equipment (or since the 1950s or so even passive conditioning) on local loops, they're just wires from the exchange to the phone.

As for why you didn't see similar constructions in other cities, this was definitely an unusually large telephone office for the time. In the US, a city exchange of the late 20th century would usually have just hundreds of lines, many of them multi-party. Telephone companies scaled up by building more exchanges, rather than a single very large one. When they got into these kinds of subscriber numbers at an exchange, the F1/F2 cable scheme was in use to avoid this kind of wiring. It does seem to be the case that telephone adoption was unusually rapid in Sweden, I find one (poorly sourced) claim that there were some 4,800 telephone subscribers in Stockholm in 1886 which would very likely make it the most telephone-rich city in the world. The situation of the tower seems to have developed in part because its builder, Allmänna, was consolidating the Stockholm telephone market through acquisitions and made a decision to centralize the many acquired customers onto on exchange.

What I'm a little confused about here is the lack of cables. The other big reason you didn't see constructions like this in the US, even in places like New York City, is because subscriber loops were quickly moved into lead-sheathed, paper-insulated multi-pair cables. These could contain hundreds of pairs. Cables were pretty much reaching maturity when this tower was built. I am curious as to the reason that multi-pair cables were not adopted more quickly in Stockholm, but it may be as simple as the considerable investment in this tower making open wire the preferred option for its short lifespan. In any case, the common claim that underground cables obsoleted the tower rings hollow to me, or at least misses an important detail, as aboveground cables were already in use in the 1880s. I suspect that modernization to cables was just deferred in Stockholm until it happened to also make sense to move to duct or pipe systems. In the US, it was more common that telephone exchanges switched to overhead (aerial) cable to manage exactly the wire sprawl issue that this tower exemplifies, and then only later (if ever) started to bury cables.

This article has more photos of the tower, but unfortunately not much more technical history: https://rarehistoricalphotos.com/the-stockholm-telephone-tow...

And this includes some photos of other parts of the Stockholm telephone network. The tower was not the only impressive construction required to manage this many open-wire pairs: https://thehistoryinsider.com/when-the-sky-over-stockholm-wa...

I wouldn't take these Truth in Accounting reports too seriously. They're linked to ALEC and take a very hard-right stance on fiscal issues, and in particular, this report on NM (which is also nearly five years old) seems to ignore the permanent funds---as best I can tell they are lumping them all under "restricted" and ignoring them, even though the land grant permanent fund, the largest of them, is totally at the discretion of the legislature and the others are very broad. The permanent funds are also now significantly larger than that report shows.

While NM has debt it has been servicing it fine and state revenue has increased year over year pretty much since that report was produced in 2020 (either 2020 or 2021 were the worst years for the state's financial position). It's projected that 2025 will close out with nearly $3.5 billion in unspent revenue, and the state has about $50 billion saved in various permanent funds. The state's financial situation is currently so good that it has allowed things like universal free college tuition in a largely revenue-neutral way due to the significant balance of the invested funds.

One of the main criticisms you will hear of the NM legislative on the fiscal front is that they are too hesitant to spend money, since NM has serious issues with underperformance in areas like education while also having billions of savings that could be spent down in an effort to address those issues (and in fact the state supreme court more or less mandated the state to start doing so several years ago). However, since NM's revenue is so tied to the oil and gas industry and its boom-and-bust nature, the legislature likes to keep a substantial cash reserve to manage the bust years. That may be particularly important right now as the Trump administration is radically reducing the amount of federal funding that NM receives, which has always been a critical revenue source due to the state's high level of poverty (third highest in the US or so, depending on year and how you measure).

Yes! I will draw a slightly roundabout connection here to my pet topics, the British shared a number of fuze concepts as part of the Tizard mission and designs based in large part on the British concepts were developed and tested at the New Mexico Proving Grounds, part of what would later become Kirtland Air Force Base. Much of this work was directed by physicist E. J. Workman, president of the New Mexico School of Mines.

I could probably rewrite the above sentence to improve it, but I focused on the magnetron because it was seen to be of special significance at the time (directly addressed issues that US efforts at e.g. MIT Radiation Laboratory were struggling with) and that there's an interesting story surrounding the couriering of the "most secret" magnetron to the US (it was briefly lost). These were the early days of "classified" as a concept and consistent techniques around safeguarding classified matter hadn't been developed, so the magnetron plays an interesting role there as well (along with documents on a number of topics, but I believe the magnetron was the only "physical artifact" brought by the Tizard mission).

I'm referring to the Tizard mission exactly, in which Cockcroft brought a magnetron to the US for show and tell. Nuclear weapons were less of an emphasis than radar (and jet engines, also a UK-led development) at that point in time.

A tremendous portion of the truck market are people who live in urban to suburban areas and need to move things. For that audience, the ability to fit a 4x8' plywood sheet easily puts this ahead of a surprising number of conventional trucks on utility. The 2k lbs payload on the 2WD drive model is more than a Tacoma and some configurations of the F150, for example, popular models that also don't fit a 4x8 sheet without strapping it down over the cab or another awkward technique. It also lists a towing cap of 6,600 which is competitive with many production pickups.

There's a divide in needs between off-roading and moving things around, and this seems oriented in the moving things around direction. I can easily see it working for a landscaper in a suburban environment, for example, where the driving miles per day are really not that high and 6,600 is plenty for a typical landscaper's trailer.

From everything I've seen, true off-roading applications are a pretty small portion of the overall truck market, and one that many popular trucks right now are also poorly optimized for (popular 2WD configurations, middling clearances, etc).