icegreentea2

Air Force One (and all of the other US flying command posts) are basically giant collections of various antennas.

Here's an article from 2017 about (then) recent installation of what were almost certainly satellite communication antenna.

https://www.twz.com/10470/air-force-one-jet-reemerges-with-u...

Here's a paper (from July 2025) on previous steps in this program, getting up the initial testing in flight. Maximum uplink laser power of 20W, though they got good performance all the way down to 2W. The sat has a laser pointing down that was used to help lock on, but it's not clear if it has any meaningful downlink capability, all discussions are about uplink capability. Lots a nerdy details here.

https://www.spiedigitallibrary.org/conference-proceedings-of...

In addition, here's a random paper on the testing performed on the space borne laser terminals - https://icsos2012.nict.go.jp/pdf/1569586689.pdf

This tells us that the laser terminals have a FOV of +/-2.5mrad in acquisition mode (so before lock on), and +/-0.5mrad in communication/tracking mode. This corresponds ~100km and ~20km radius FOV from GEO to surface.

??? GPU vs CPU workloads are completely different. Comparing Panther Lake iGPU vs Ryzen iGPU is not going to tell you much about how high density server CPU performance will work out.

The Panther Lake vs Ryzen laptop performance comparisons show that Pather Lake does well, basically trading against top end Ryzen AI laptop chips in both absolute performance, and performance per watt.

Berger is clearly guarded and measured when he talks about Musk and SpaceX. Given the configuration of the space industry and the reality that Berger clearly needs access to make his living, I think Berger has provided generally even handed coverage of SpaceX, Musk and Musk's antics.

For example, his article on the SpaceX/xAI merger ends with a section called "Has SpaceX lost its way?", and framing that clearly indicates "this is what Musk says". https://arstechnica.com/ai/2026/02/spacex-acquires-xai-plans...

In his article on SpaceX's "pivot from Mars to Moon", he describes Musk as "In the last 25 years, Musk has gone from an obscure, modestly wealthy person to the richest human being ever, from a political moderate to chief supporter of Donald Trump; from a respected entrepreneur to, well, to a lot of things to a lot of people: world’s greatest industrialist/supervillain/savant/grifter-fraudster.". https://arstechnica.com/space/2026/02/has-elon-musk-given-up...

You can read a AMA he did on the SpaceX subreddit last year for more of his thoughts: https://www.reddit.com/r/spacex/comments/1fnq02q/eric_berger...

I think Berger got a lot of flack for not really commenting on Musk and DOGE during his articles last year, and I think it's fair to criticize him for that choice, but I don't think it's really a "Musk can do no wrong" position.

In other words, I think you really need to read between the lines for Berger re Musk.

US forces have been answering the wake up call for the last few years.

The main weapon that Apache's use to hunt drones are laser guided rockets (APKWS) with a per-shot cost around $30k (https://www.airandspaceforces.com/apkws-base-laser-guided-ro...)

These weapons have been fitted to most US tactical fighters for the counter drone role as well. APKWS is also not a "new" weapon system - it started fielding back to 2012, and was adapted into the counter drone role.

There are other lower cost (compared to legacy systems designed to take on manned aircraft) solutions currently deployed. The US Army has the Coyote, which is in the ~100k range.

Beyond cost of munitions, you have to consider that cheaper systems are going to have less range, and therefore you'll need more launchers, and you can start running up costs that way.

I don't think it's 100% accurate to call the budget "austerity".

The budget projects increases in both revenue and spending, and increases the deficit by ~50% percent.

What the government has asked for is widespread cuts across the board from current programs and operations (and staffing) to try to make more room for new spending in new areas. They are targeting getting to ~$13 billion CAD reduction in annual operations cost by end of decade.

By contrast, there are a lot of major spending programs - ~$10 billion CAD added to the defense budget, ~$5 billion CAD in tariff relief.

American beef imports are heavily driven by how US beef production and consumption is structured. Basically, the US focuses beef production on creating high quality cuts (ie, steaks and other cuts with high marbling/fat), because that's what a large portion of domestic consumption is (and it's highest value). This leaves the US with a very large amount of high fat off cuts that aren't very marketable on their own. Imports are typically ultra lean cuts (which are also not very usable), and these two sources are then combined into ground beef.

Here's one source, but there are plenty of others you can find: https://tscra.org/we-have-94-million-cows-why-do-we-import-b...

Basically, classic value chain optimization.

UK uses "imperial gallons", which are no more metric than US customary gallons. Expanded to 3 decimals, a UK/imperial gallon is 4.546L, and a US customary gallon is 3.786, and a US gallon is close to 3.8L, than a imperial gallon is to 4.5L.

I think it's right and honest to admit that this is one of the methods that sanctions are supposed to work. But it's also not the only method - and framing the intent as inducing "regime change by internal bad-actors" is also a very slanted way to articulate intent, as well as what is happening on the ground.

On the other hand, without being on the ground, we cannot really say what the real balance of grievances are.

Even more precisely, human writing contains unpredictability that is either more or less intention (what might be called authors intent), as well as much more subconsciously added (what we might call quirks or imprinted behavior).

The first requires intention, something that as far as we know, LLMs simply cannot truly have or express. The second is something that can be approximated. Perhaps very well, but a mass of people using the same models with the same approximationa still lead to loss of distinction.

Perhaps LLMs that were fully individually trained could sufficiently replicate a person's quirks (I dunno), but that's hardly a scalable process.

Chiplets are different physical chips, potentially from different nodes or processes. They're physically connected prepackaging. It's beneficial compared to sticking them on a board because then you get the shortest interconnects, and you aren't limited by pinout density. This should let you get the best bandwidth to power ratio, and better latency.

Bluesky post by Benj (one of the authors of the article). https://bsky.app/profile/benjedwards.com/post/3mewgow6ch22p

He admits to using an AI tool, says he was sick and did dumb things. He does clear Kyle (the other author).

They just needed to expand "At this time, this appears to be an isolated incident." into "We are still investigating, however at this time, this appears to be an isolated incident". No additional details required.

And yes, it looks like Ars is still investigating (bluesky post by one of the authors of the retracted article) https://bsky.app/profile/kyleor.land/post/3mewdlloe7s2j

Yeah, but the problem is that by not making it clear that additional actions may be coming, they're barely restoring credibility at all, because the current course of action (pulling the article and saying sorry) is like the bare minimal required to avoid being outright liars - a far cry from being credible journalists. All they've done is leave piles of readers (including Ars subscribers) going "wtf".

If they felt the need to post something in a hurry on the weekend, then the message should acknowledge that, and acknowledge that "investigation continues" or something like that

Yeah, the initial comments are pretty glazey, but go to the second and third pages of comments (ars default sorts by time). I'll pull some quotes:

> If there is a thread for redundant comments, I think this is the one. I, too, will want to see substantially more followup here, ideally this week. My subscription is at stake.

> I know Aurich said that a statement would be coming next week, due to the weekend and a public holiday, so I appreciate that a first statement came earlier. [...] Personally, I would expect Ars to not work with the authors in the future

> (from Jim Salter, a former writer at Ars) That's good to hear. But frankly, this is still the kind of "isolated incident" that should be considered an immediate firing offense.

> Echoing others that I’m waiting to see if Ars properly and publicly reckons with what happened here before I hit the “cancel subscription” button

The comments are trending towards being more critical as of my posting. A lot more asking what they're going to do about the authors, and what the hell happened.

In 2003, Bush first tried gathering support for invasion in Sept 2002 at the UN. Congress granted authority for use of military force in October, and the troop build up started. Colin Powell tried one more major push for UN support in Jan 2002. The invasion was in March 2003.

That was the public facing attempts to gather support. Internally within the administration, they started working on invasion plans within a few months of 9/11. These plans continued to iterate up to being more or less locked in and approved (by Bush) in Jan 2003.

You can still do this in W11 notepad. Firstly, there's a global setting for having formatting/markdown being enabled at all, and secondly it only does the rendering for .md files. Finally, while formatting is enabled, and editting a markdown file, you have the option to toggle between formatted and "syntax" view (ie raw text).

Trial designs are rarely perfect, and drug companies will try to stack the deck.

That being said, trial designs aren't made in isolation without consultation. Moderna's press release about this (https://feeds.issuerdirect.com/news-release.html?newsid=7346...) indicates that 2024 CDC said that performing an efficacy trial without using the higher dose would likely cause complications for recommending the mRNA vaccine for the >65 age group, but would not impact overall approval.

No, they're implying that Moderna should have used a high dose flu vaccine in the >65 age group as the control condition in the efficacy trial.

EDIT: To be clear, I'm pretty sure that's just some pretext.

Btw, the paper referenced for tungsten production is co-funded by a US tungsten exploration/mining company: https://www.innovationnewsnetwork.com/harnessing-tungsten-fo...

I can only read the free part of the paper (https://www.sciencedirect.com/science/article/abs/pii/S09203...). I wish it would elaborate more on the challenges with recycling.

If we were spending the same effort and resources to stay flat, I'd agree with you.

I think the "crisis" is that we've been spending more and more resources just to stay "flat".

You can go to the "nations report card" and play around with the NAEP score stats: https://www.nationsreportcard.gov/ltt/mathematics/scores-per...

For me, two things pop out in particular.

Firstly, in math, if you look at how the percentiles break down, it seems clear that while there an overall drop in performance from 2020-2023, it also seems clear to that the top end of doing relatively ok. For example, at the 75th percentile, the we are basically flat from 2008 to 2020 (before dropping in 2023) - 2008: 305, 2012: 309, 2020: 307, 2023: 301 (net -4). This contrasts with the median which went from 2008: 283, 2012: 287, 2020: 282, 2023: 274 (net -9).

This implies to me that whatever flaws are in the overall system (at least pre COVID), the top end was relatively durable.

Secondly, if you go the "student group scores" section and click through all of the different sub-groupings, the only group that looks to have an overall flat score at all is "Private: Catholic".

I think the combination of the upper end being pretty durable, as well as the higher scores in the only "self selecting" category in the dataset may support what a lot of people tend to grumble about - the distribution of domestic situations is not favorable.

Wow, 2.5% failure rates for customer CPUs?

I just... I dunno, I just mentally binned CPUs (as a solid state object, and one of the highest value add components to a system) to have failure rates in the ~1% range.

I know the effort it takes to get asymptotically closer to 0% failure rate, and so this is more of a misread (on my part) on what the market tolerates and the vendor's pricing/value strategy.

Cool!

China was able to sustain some pretty strict zero-COVID policies much longer - all the way to late 2022.

Pain tolerance might be the wrong term. Pain tolerance implies speaks to something intrinsic about a population, while really what we're looking at is how much discomfort a population can endure before it really agitates for policy/political change, and so it's much about how a population feels, as the tools available to the government to control, manage, deflect and address the pain/discomfort.

The deal allows up to 70,000 cars a year by 2030 to be imported at the reduced tariff. Canadians buy 1.5-2 million cars per year, and roughly a quarter million EVs per year.

If this deal as reported somehow manages to doom the Canadian auto industry, then our auto industry was probably somehow doomed anyways.

If they rescinded that lease, then Trump's so called reasoning actually becomes valid. The US does actually have a geopolitical/strategic/military interest in being able to operate from and around Greenland. The reason this entire activity is a farce is because the US can already do that.

It's clear that Trump acts alone in foreign policy - formal channels and structures can barely check him. However, informal resistance still appears to exist. Trump apparently still takes into account the vibes of the people he surrounds himself with into account. In a haphazard way yes, but it's clear that Trump can be swayed to some degree by those around him.

The Trump administration is not a unified bloc, and there are likely many elements that see annexing Greenland as ridiculous. However, if they lost access, then they would be forced to concede that there was something actually valuable to gain.

The author pays £0.07/kWh off peak, but can export at £0.15/kWh. The author paid ~£7500 per powerwall which has ~13.5kWh capacity. Assuming full charge/discharge every night, you can make ~£1.08 per day, which works out to about 19 years to pay back.

Utilities normally consider disincentivizing this type of behavior from residential customers as one of the factors when setting their export pricing.

To be clear, NASA has an entire field center dedicated to rocket testing (https://en.wikipedia.org/wiki/Stennis_Space_Center). This is where everything gets tested now. You may remember the "green run" tests of the SLS - those happened at Stennis.

Stennis didn't exist at the start of the space race or Apollo. Marshall is colocated on Redstone Arsenal, a legacy of parts of NASA being bootstrapped out of the Army ballistic missile program. Marshall had test stands because that era of NASA (aka von Braun) sought to colocated engineering, prototyping and test.

One challenge with continuing tests at Marshall is that... it's actually really close to population centers. Large engine tests would be ridiculously disruptive. There are comments in the Ars article noting that people living in Huntsville could hear the demolition work.

While the power draw might be high in absolute terms, the surface area is also quite large. For example, the article's estimates add up to just 2000mm2 for the Epyc chip. For reference, a Ryzen 9950X (AMD's hottest desktop CPU) has a surface area of about 262mm2, and a PPT (maximum power draw) of ~230W. This means that the max heat flux at the chip interface will almost certainly be lower on the Epyc chip than on the Ryzen - I don't think we're going to be getting 1000W+ PPT/TDP chips.

From that you can infer that there shouldn't be the need for liquid cooling in terms of getting the heat off the chip.

There still are overall system power dissipation problems, which might lead you to want to use liquid cooling, but not necessarily.

For example, Super Micro will sell you air cooled 1U servers that options up to 400W CPU options (https://www.supermicro.com/en/products/system/hyper/1u/as%20...)