schroeding

The analogy is a bit shaky IMO, as you can certify individual, heavily modified, foreign or even self-built cars in EU member states.

For cars, the local certification authority themselves decides what is road-worthy or not, not VW et al. You can add third party parts without the manufacturers consent. This is not the case for Android or iOS attestation, you're pretty much at the mercy of the foreign manufacturer and their local laws.

You cannot prove the absence of e.g. a Veracrypt hidden volume or similar, though. Even if you honestly give up your key, you could still be either

A) held in contempt of court, if the authorities do not find what they expect for some reason and accuse you of using such techniques or

B) if you specify that such behaviour by law enforcement is overreach, have a clean way out for criminals, codified in law, heavily damaging the impact you may expect of such a law.

Because what is the alternative, with the same level of convenience and support by (online) shops? Genuine question, because I do not know any and would be happy to.

Wero is not supported by most shops (yet), Credit / Debit Cards often charge fees if the payment goes to a foreign country, direct transactions cannot be undone.

The problem in Munich is that everything must go through a single two-track part underneath the city center, which is at absolute capacity. If anything breaks down there (and it does, often, very often), even a small delay in a single train, all trains get delayed or skip stops.

In my experience, you have to take at least one train early if you do not want to come late regularly. Even e.g. the main airport train line, used by tourists, often turns around before the actual airport due to delays.

If you live in the city itself, it's fine, you also have other options. If you live further away, it's barely acceptable to very bad, IMO.

It is reliable-ish, but more "Amtrak Capital Corridor"-reliable than "JR Yamanote Line"-reliable.

I think they mean the actual posts on tea itself, not the leaked ID photos.

The "poisoning the data supply" angle seems to be a common motive, similar to tools like nightshade[1] (for actual images and not just garbage data).

[1] https://nightshade.cs.uchicago.edu/whatis.html

Interestingly, the pendulum at least in my friend group starts to kinda swing in the other direction, i.e. non-technical friends start to indirectly ask (me as the tech guy) about blatant piracy for (visual, Spotify is still very much accepted) media and (TOS-violating[1]) ad blockers for ad-supported streaming.

I cannot overstate how unexpected this was and is to me, we talk about people in their mid-twenties with jobs - maybe (video) streaming / subscriptions services actually overplayed their hand in the current economic climate.

Doesn't make me super optimistic in this regard.

[1] even if most of it is void in my jurisdiction anyway

Ah. Thanks! Since the letter only calls it "reconciliation bill", I didn't make the connection. Not an American here, oops. Maybe creating "mega bill bundles" isn't the best idea in general. ^^'

I still think this specific reversion / change, for itself, would be something you can lobby for, though. It itself doesn't do harm, the push to include it in this specific bill may do (if it is the thing which tips the scale for it to be accepted).

This "tax cut" is (and was) simply the status quo in most western countries for virtually all businesses, e.g. in the EU. It itself is not immoral, as long as you see developers as normal office workers, which they IMO are.

The existence of silicon valley giants and their faults notwithstanding.

IMO, if you lobby for a thing which does not do harm to other people, you are not the bad guy. If you do, you are. Lobbying itself is not immoral.

The oil and gas industry, and the tobacco industry et al., lobbied (and lobby) for things which they know were (and are) doing harm. This isn't the case here, IMO.

Code is not an asset in all (I would even argue most) cases - proven by companies which open source the vast majority of their code and live from service contracts or certain addons to it, and basically pay developers to commit to open source software.

Often they buy market- or mindshare. There is no way in hell e.g. Akamai wouldn't have been able to bootstrap "Linode 2". I'm unable to see the secret sauce why OpenAI couldn't have created their own VS Code fork instead of buying Windsurf. But why do that if you can acquire their existent customers / market share? Additionally, the term "acquihire" didn't plop into existence with no precedent.

Being able to immediately get a full deductible for salary, which in many (western) countries is the norm for virtually all businesses, does not strike me as particularly immoral. It's a normal office job, developers do not create gold out of thin air.

Big tech isn't even the most affected by this change, they (often) have obscene margins - small software companies do not.

These are ccTLDs, though, ICANN is out of the picture there, they have no authority after delegation. It's the fault of DENIC, the German ccTLD local operator. DENIC is a German entity, they are very much within reach of regulators.

(That's also the reason why foreign ccTLDs of, eh, semi-stable countries, e.g. .so domains, are risky - should the local operator start to lose it at some point, no-one can help you, neither ICANN nor IANA)

IMO "not ideal" is still a euphemism. Even if you do this, often throwing away decades of branding, there is a high chance that Lieferando aka Just Eat Takeaway.com will still not only register a similar domain again, but that it will rank higher than your website anyway, even with the "worse" domain.

You have a small restaurant, often using things like WiX or Squarespace, against a tech company with a dedicated SEO team. Good luck.

Right, fair enough :D

IMO, Framework fits. It's very repairable, and it's kinda easy to do. Not having to work with clips of any kind is great, it's all screws and magnets. At least the 13" one has the same (or even a bit less) flex and feeling like e.g. a Lenovo Thinkbook. It's CNC-ed aluminium, both bottom and lid.

It is significantly more fragile than a X220, though. And tbh, the screen looks the most fragile of any laptop I ever had, as there is an air gap behind the LCD, i.e. you really shouldn't hit it with e.g. a pen or something, at all, as it may flex and break. At least you can easily replace it, without ripping out glue, if it happens? :s

It also is not the best bang-for-buck compared to other new laptops, if you ignore the repairability. IMO understandable due to their smaller scale and additional engineering, but still true.

I use mine with Linux and it's great, you feel like a first-class customer like you do with e.g. a Systems76 machine (which are also nice), it's explicitly supported. Here is their support page for each motherboard: https://frame.work/de/en/linux

If you choose a mainboard with a new CPU, you may need to use a mainline kernel instead of LTS for a while, but that's it, in my experience.

Battery life is not on ultra book level, but it's at least on par with the X220. I get like 8, 9 hours out of it.

Can highly recommend it, even if it's not perfect.

The US Air Force bombed a local underground fuel storage facility, one of the biggest in Germany at the time, with quite a lot of heavy bombers in April 1945 with mixed success - only three days before finally capturing the facility with the US Army. At that time, the tanks were already almost empty, as the existing (significant) reserves were used for weeks to fuel the last heavy vehicles and tanks.

The fuel storage facilities were mainly used to mix the fuel the Luftwaffe needed - they got almost all of their fuel from them.

Always wondered what made the allies wait for so long or why it apparently wasn't a high priority target - they had to know that it was there, there were two railway connections and large amounts of visible tank cars (which they did attack a few times with smaller planes) just a few hundred metres away, with a major military airfield close, and it was a (mostly) unpopulated area. There were bomber formations flying overhead for months at that point, as a big German city was just a few miles away.

It is way more democracy and freedom than living in a state with an entity like the Stasi, a mixture between the NSA and the Gestapo, which is used to curb any opposition, at least.

It's not perfect, but this alternative is way worse.

Until they need new hardware, for which there will (and can) be no drivers anymore, as Microsoft stops crosssigning them in the Hardware Dev Center after a while for old Windows versions.

Microsoft does have the leverage in this case, as long as folks want to continue using Windows.

I like the restrain they have: "This points to the possibility that the survey may have been manipulated", yeah, 250k votes from 2 US IPs on a survey of a German-only equivalent to msn.com, may. :D

Translation:

According to a T-Online[1] survey, hardly any Germans still want to buy Teslas. But suddenly the number of participants exploded. T-Online stops the survey.

100k people took part in a non-representative t-online survey last week - within just a few days. The question was: "Would you still buy a Tesla car?" The result was clear on March 11th: around 94 % of participants said "Absolutely not". In contrast, only around 3 % of participants chose "Yes, no problem".

Just one week later, the result looks different - and the number of participants has exploded. 467k people had voted by Tuesday evening (as of March 18th, 17:00 o'clock). Suddenly, around 70 % of survey participants want to buy a Tesla - the proportion saying "no way" has fallen to 29 %.

It was initially unclear where these votes - and the sudden change in opinion - came from. At first glance, the number of article views in the past few days and the number of survey participants do not match. Initial in-house research now shows that 253k of the votes cast came from just 2 IP addresses in the USA. This points to the possibility that the survey may have been manipulated.

Furthermore, T-Online noticed that the link to the evaluation article and thus to the survey was shared thousands of times on Twitter and other social networks. On Tuesday evening, Tesla boss Musk himself shared the article with the new, more pleasing result for him.

Musk's post collected 2,8 million views on Twitter within the first two hours. Since then, the number of hits on the survey has increased noticeably once again. On Tuesday evening, there were several hundred per minute. T-Online has stopped the survey until further notice and removed it from the relevant articles.

It would be interesting to see what AS they're from. If they're US residential[1] (or governmental), it would be more interesting than if they're just US data center.

[1] it still could be a hijacked device from a botnet or else, yadayada, but still

Yes, but many companies change this via GPO. And even then, they remain hidden. They are special, compared to normal extensions.

Thus users are unlikely to detect that they're LNK files (in Windows), even with enabled extension.

The thing is that Windows does not show the .lnk Ending, same for .pif (which can just be a normal MZ Windows EXE), even if all other endings are shown. So your bill.pdf.lnk is shown as bill.pdf, while bill.pdf is also shown as bill.pdf.

(This may have changed with new versions of Windows 11, I'm stopped working in this area a few years back)

Can only recommend them - not too expensive, you can also use your own domains and they support at-rest auto encryption of all incoming mail with a PGP public key you give them (which of course does not prevent them from saving incoming mail as clear text somewhere else, but prevents others from reading all existing mail should they get access to your mailbox later)

Yes! (e.g. https://shop.fenka.de/products/pudu-bellabot)

And no, sadly not. :(

E.g. France24[1], they have all sources linked. Then there are individual governments, like e.g. the Latvians[2], warning that Russia is preparing for a confrontation with NATO. Or the German DoD[3]. Yes, no politician is explictily saying "Russia will invade country Y in X years", but "Russia prepares all resources they need to invade, and NATO confrontation is not unlikely" is good enough for me, at least as long as Russian state TV shows nukes exploding over Western Europe or battlegroups invading the Baltics.

The Baltics have no beef with Russia, but Russia with the Baltic states[4], according to e.g. Medwedjew, who thinks they literally belong to Russia. Or Россия 1[5].

[1] https://www.france24.com/en/europe/20250219-baltic-region-pr... [2] https://www.sab.gov.lv/en/news/russia-is-purposefully-develo... [3] https://www-bundestag-de.translate.goog/dokumente/textarchiv... [4] https://www.euronews.com/2023/05/17/russias-dmitry-medvedev-... [5] https://www.youtube.com/watch?v=fSJ6py0uVQ4, from 1:30 onwards

The US wasn't active part of the war until December 1941, though. For the first years, no boots on the ground were coming. And if Germany didn't declare war against the US, which they weren't forced to do by the tripartite pact, it's questionable they would have come at all (in 1944).

Doesn't Ukraine buy most of their heavy weapons from the US (... until now)? If so, isn't most of the tax payer money the US lends Ukraine going back into the US economy, softening the blow by a lot?

Did the (original) lend-lease act even hurt the US economy and US families? Or did it provide jobs and stimulate the US economy, cementing the role of the US as the military / industrial powerhouse they are today? I was under the impression that it's the latter, but I'm not sure, correct me if I'm wrong. Is this one different?

But sure, of course - that's the right of the US, even though it would be nicer not to retroactively break existing agreements. I only wanted to push back against the (new) lend-lease doing "absolutely nothing".

If France had hold out until 1943, with the Wehrmacht only making a few hundred yards progress per day, it would have been a massive success, don't you agree? Without western help, Ukraine would have had no chance in hell to survive that long.

It's the same logic the US applied to lend-lease, didn't they? There was a good chance of the USSR collapsing until 1943, and you still helped them. Delaying an invading attacker, especially one who may plan, according to their own internal propaganda, to attack other nations later, is already more than "nothing".

The price for peace is living under occupation. Depending on the occupier, this can be a fate worse than death, and significant numbers of Ukrainians appear to think the latter.

And for a very selfish argument: It buys us, the Europeans, more time (with Ukrainian blood) to prepare for war, when Russia attacks the Baltics in a few years. Which we desperately need, as the US will likely break their promise and not help us under the current administration (and who knows who the US elects next).

The 3 years allowed e.g. the Baltics to build extensive defensive fortifications.

I don't like the idea of telling people (even smugly at worst, if it's used the way LMGTFY was, and as the tagline[1] implies) to just ask ChatGPT instead of humans.

LMGTFY returned human results that weren't SEO rubbish at the time for the most part, which you could classify yourself (or by others) as trustworthy or not. ChatGPT itself still sometimes hallucinates on pretty basic information and other users don't get exactly the same result as you.

[1] "Someone else asked ChatGPT for you. Appreciate it."

Mh. Is a comparison with OpenSSL or Snort really valid?

OpenSSL implements crypto primitives, which, as far as we currently know, are mathematically safe. It's clear cut, black and white, it either is implemented correctly and (as far as we know) safe, or it isn't.

Snort works based on fixed rules. It's black and white, either traffic is logged and / or blocked, or it isn't. Basic rules log entire classes of traffic, the ability to engineer around them is very limited. Specialized Snort rules are a cat and mouse game, though.

At least two of the three things ROOST wants to do are machine learning models. They are not black and white, they are fuzzy. Minor changes can be enough to fool a classifier without changing the real semantic meaning for humans.

I have very, very hard doubts Google, OpenAI et al. will release Open Source classifiers which they will actually also use themselves.

Either they will try to get perfect recall, then it certainly will have (probably massive) overreach. Or there will always be false negatives. In the latter case, ML classifiers are IMO the one thing where security through obscurity not only actually works, but is the only thing which keeps them working at all. If you can freely train on an imperfect classifier, you can fool it, no?

I would love this, I like the Open Source approach a lot, but I cannot imagine how this should work in practice.

Also: How should third-party companies be able to reliably "set their own rules" on a probabilistic ML model, anyway, without training it themselves? The thing Roblox published[1] is a classifier for 6 categories. Sure, it's nice that you can e.g. say "I only want to filter profanity" this way - but what constitutes profanity is not controlled by you, but by the whims of a US company, even though this very much is not a globally homogeneous standard.

[1] https://github.com/Roblox/voice-safety-classifier

Does cURL have an easy-to-spot fingerprint outside of its headers?

If it's a https URL: Yes, the TLS handshake. There are curl builds[1] which try (and succeed) to imitate the TLS handshake (and settings for HTTP/2) of a normal browser, though.

[1] https://github.com/lwthiker/curl-impersonate

I have no idea about the intricacies of US law, but reading the actual proposed bill[1], it seems incredibly broad. Like, "this may make publishing linear algebra papers on faster matrix operations illegal"-broad, as it includes everything which may contribute to AI research.

Also: Preventing the import of "foreign" knowledge / research seems completely insane to me, no offense, no matter how one stands on export. There is nothing justifying this, you're just kneecapping yourself.

[1] https://www.hawley.senate.gov/wp-content/uploads/2025/01/Haw...

The term artificial intelligence [...] means technology or intellectual property that could be used to contribute to artificial intelligence or generative artificial intelligence capabilities

The term "export", [...], includes [...] the release or transfer of technology or source code relating to the item to a foreign person in the United States.

PROHIBITION ON EXPORT (& import) - [...] the export (& import) [...] of artificial intelligence or generative artificial intelligence technology or intellectual property to or within the People's Republic of China is prohibited.

Ehhh, do I misunderstand US legal terms or wouldn't this be a death sentence for open source AI research and any AI papers by US researchers in journals & conferences? Or does public distribution on the web not count as (indirect) export?