Yeah, no. Merkel's deal to shut off the nuclear plants to make a coalition was 100% a blunder. Not only in hindsight, with the dependence on russian gas, but in general it was a blunder. Nuclear gives you steady energy in ways that renewables can't. We should absolutely do more renewables, but to shut off working nuclear was not good.
Nuclear is not that steady. Nuclear plants require a lot of water to cool things. And when a particular hot summer happens, rivers dry out and nuclear reactors have to scale down the power production or even be shutdown. And then they require quite significant maintenance periodically.
Granted, in Europe a hot dry summer is when solar is at its peak. So it is much lesser problem than a cold winter with a lot of cloudy days with no wind when nuclear energy is ideal.
Still from a perspective of 20 years ago with unknown prospects about renewables natural gas power stations were considered much more reliable and flexible power source compared with nuclear and way more cleaner than coal. Of cause, as long as one gets gas.
It is simply false that it was Merkel who decided to shut down nuclear power plants. The decision had been made over a decade earlier. She just accelerated the plan in the end after a previous unsuccessful attempt at rolling back part of it. It also wasn't even really her decision, it was the will of the people that sharply turned against nuclear after Fukushima, she just implemented it.
I wonder why they keep using a dedicated numbers station instead of embedding the code in a regular radio broadcast on a traditional channel? I'm sure that even before LLMs one could find a way to create a story where certain numbers / code words would be embedded without altering the underlying story too much. And they could probably get BBC / whatever station to air it. It would be a bit less inconspicuous to listen to BBC than to a dedicated numbers station, even if the message would be undecryptable either way.
> "I'm sure that even before LLMs one could find a way to create a story where certain numbers / code words would be embedded without altering the underlying story too much."
It's called steganography, and it's a centuries if not millennia old technique.
> The first recorded uses of steganography can be traced back to 440 BC in Greece, when Herodotus mentions two examples in his Histories. Histiaeus sent a message to his vassal, Aristagoras, by shaving the head of his most trusted servant, "marking" the message onto his scalp, then sending him on his way once his hair had regrown, with the instruction, "When thou art come to Miletus, bid Aristagoras shave thy head, and look thereon." Additionally, Demaratus sent a warning about a forthcoming attack to Greece by writing it directly on the wooden backing of a wax tablet before applying its beeswax surface. Wax tablets were in common use then as reusable writing surfaces, sometimes used for shorthand.
Seems to me like coordinating with an entity outside of the spooks' control, such as the BBC, would give more opportunities for leaks. It would also reveal some information about who is controlling the signal--someone with some kind of relationship with the broadcaster.
During WWII, the BBC would daily have a section after the news dedicated to "personal messages" - which everyone knew were instructions to the resistance in France, or similar. "William waits for Mary" was one of the more famous ones related to D-Day, I think.
Because you can drive intel analysts crazy with this one weird trick. They know you can't decrypt one time pads, but they can't resist checking for entropy and trying to match it to known OTPs they may have acquired through intelligence channels. Running and programming the shortwave transmitter is dirt cheap; tying up some of opponents' SIGINT resources on a wild goose chase is good value for money.
Roosevelt told the Shah that he was in Iran on behalf of the American and British secret services, and that this would be confirmed by a code word the Shah would be able to hear on the BBC the next night. Churchill had arranged that the BBC would end its broadcast day by saying not 'it is now midnight' as usual, but 'it is now exactly midnight'
who's to say they aren't doing both? They may not even be sending anything over the number station; these stations will continue on a schedule even when there is nothing to say and nobody is listening because it makes it harder to eek out a foothold in the event of a weakness in the encryption.
Even if the encryption is one-time pads, if you broadcast a bit every day then you don't warn the enemy that something's up by the fact that you're transmitting at all.
However, the numbers stations transmissions are never a big secret. They're intentionally powerful so someone can pick them up on simple equipment without raising suspicion. A person can modify an off-the-shelf AM radio to pick up shortwave, for example, even in an oppressive regime.
It's a one-time pad, so the encryption is unbreakable.
I can't find it immediately, but I've read about something even sneakier than this. A standard broadcast station was modified such that its carrier signal was modulated by a PSK signal. The intended listener would use e.g., a PSK-31 modem to listen to the carrier signal and would be able to obtain the encoded digital data. Everyday listeners would hear the regular broadcast. The station involved _might_ have been a BBC station, but I don't recall.
You could technically just transmit data via RDS, too. Change a letter here and there and nobody would know whether that’s a decoding error or actual ciphertext. (Would need some kind of checksum or so, of course.)
@windytan did a fascinating audio clip highlighting the RDS data stream in a radio recording some while ago:
Shortwave propagates better and also its just a one time pad being distributed so embedding doesn't matter as much as long as the one time pad is longer than the intended message to send. There is no way to decrypt it because once you encrypt a message using a one time pad it is impossible to decrypt without the exact one time pad that it was encrypted with.
One time pads work only if only the sender and receiver have a copy of the pad - and they destroy each sheet on use. Distributing the pads is hard, but often it can be done easier than the message.
Distributing a one time pad like this is a stupid idea: it isn't hard to collect everything you ever send, and it takes a computer a few ms to check every encrypted message against every possible sequence. That is breaking a distribute one time pad via shortwave like this is something a single layperson can do, it doesn't even need a government scale attacker to break it.
Don't get me wrong, this can be used for good encryption. However it isn't a one time pad they are doing, it is something more complex.
Every message is equally likely when you attempt this kind of brute-force decryption with a one-time pad. The code you get is actually 100% unbreakable if the pad isn't intercepted.
I think there's some confusion in this thread. GGP talks about distributing the one time pad via the numbers station. GP (rightly) says that's a stupid idea.
The numbers station should be transmitting a message encoded with a one time pad. The one time pad itself should be physically given in person to the spies who you want to communicate with.
Or, if one is uncertain whether to trust the courier between you and your spy - one can send two different one time pads by two different couriers. If the spy is trained to xor those pads together before using, an enemy must intercept both pads to be able to read your messages.
There are many variants on this, including pads which you hope your enemy will intercept.
It's not a one-time pad being distributed, because leaking the pad leaks all your communications. It's almost certainly the actual messages being distributed, at specific times of day. The listener records the numbers for the known time period to get the message, then decodes it with their pad for that period. Then they destroy that pad. Continually broadcasting numbers makes it impossible to tell the length of the messages.
Sure, but that would be a benefit, I would think. Most old cars come with an AM/FM radio, most cheap phones now have FM (? I don't know about AM, don't think so) and so on. So it would be more inconspicuous to listen to a regular radio than to a special station on special hardware. You don't even have to broadcast from EU, you could probably purchase some Radio Quatar Classical Rock or something :)
Radios capable of receiving shortwave bands aren't exactly rare among normal people. They're not really "special hardware". Just owning one would not be inherently suspicious.
What would be suspicious is being in possession of the one-time pad needed to decode the messages, regardless of which media those messages are transmitted through.
For the record, "numbers stations" can be found in nearly every communication medium, including the web. The advantage of using shortwave (range, primarily) are large enough that the benefits outweigh the drawbacks.
> What would be suspicious is being in possession of the one-time pad needed to decode the messages
Would it though?
All you need is something with sufficient entropy. I reckon you could do a "good enough" job with any plausible-looking data you have lying around on your hard disk right now. Say for example if you took a couple of sha256s of any random image you might post on social media, you'd have quite a lot of key right there.
It was once common knowledge that VHF radio ("FM") typically doesn't travel over the horizon, LW and MW radio ("AM") travel by ground wave and are regional, but that you need shortwave kit for international and global communications.
Quite how a reader on a modern technical news site is unaware of this (no, you can't send direct messages to spies half way around the planet to be received on an "AM/M" car radio) shows just how common public knowledge of radio communications has faded over the last few years.
Performance can be a direct target in a feedback loop and optimised away. That's the easy part. Taking an idea and poof-ing a working implementation is the hard part.
Yeah, from small interactions over the past two decades, I have no idea how they could have been so bad while employing so many people. What on earth were those 30k people doing?! Their solutions were crap for ages.
>I have no idea how they could have been so bad while employing so many people
There is a significant correlation between how many people you employ and how much nothing you accomplish. It means you've gotten big enough to survive long bouts of doing something and achieving nothing with large amounts of people.
I swear this whole thread about regexes is just fake rage at something, and I bet it'd be reversed had they used something heavier (omg, look they're using an LLM call where a simple regex would have worked, lul)...
The problem is the oauth and their stance on bypassing that. You'd want to use your subscription, and they probably can detect that and ban users. They hold all the power there.
I have no interest in Claude Code as a harness, only their models. I'm used to OpenCode at this point and don't want to switch to a proprietary harness.
Lol what? There is no value. OpenCode and Pi and more exist. Arguably Claude Code is the worst client on the market. People use Claude Code not because it's some amazing software. It's to access Opus at a discounted rate.
I don’t think that’s a good comparison. There isn’t anything preventing Anthropic from, say, detecting whether the user is using the exact same system prompt and tool definition as Claude Code and call it a day. Will make developing other apps nearly impossible.
It’s a dynamic, subscription based service, not a static asset like a video.
This is truly amazing. Do people not really realize how amazing stuff like this is? I feel like I'm taking crazy pills here, but man, it certainly feels like we're on the edge of something quite amazing...
> seems like the treatement is almost worse than the disease.
I think that's what the poster above you was saying. "Oldschool" chemo is basically poison, and the hope is that it kills off the cancer before the patient. But there are newer drugs that are extremely effective with way way way less side effects out there, depending on which type of cancer one has. Things like immunotherapy are really effective if you happen to match their targeted types of cancer, and some have basically 0 side effects, leading to a QoL improvement if they happen to work. People have gotten nobel prizes for some of these discoveries, it's really insane how far we've come in the last 30-40 years.
It was free for a long time. That usually skews the statistics. It was the same with grok-code-fast1.
reply