Happy Birthday Alan Turing! How Modern Technology Could Win WWII In 13 Minutes (digitalocean.com)

DevNull127 writes: A grateful reporter whose father-in-law liberated a concentration camp after D-Day reports on a high-tech team that "accomplished in 13 minutes what took Alan Turing years to do — and at a cost of just $7."

"In late 2017, at the Imperial War Museum in London, developers applied modern AI techniques to break the 'unbreakable' Enigma machine used by the Nazis to encrypt their correspondences in World War II."

Two Polish co-founders of a company called Enigma Pattern decided to honor Alan Turing's ground-breaking work at Bletchley Park, where Turing had automated the testing of over 15 billion possible passwords each day by building what's considered the first modern computer. They took the problem to a modern cloud infrastructure provider, renting what one describes as "2,000 minions that do the tedious work" — specifically, crunching 41 million combinations each second — using Grimm's Fairy Tales to train an algorithm to recognize when they had found a commonly-used German word (including familiar bedtime stories like Hansel & Gretl and Rumpelstiltskin). "In the end the AI could not understand German. But it did what machine learning does best: recognize patterns."

"After 13 minutes of minion work, boom! The new Bombe had broken the code."

Turing's birthday is Saturday — and it's nice to see him being remembered so fondly.

Comments

Never Would Have Used It

[Bing Tsher E]

The Germans never would have used such an encryption if modern methods of breaking it existed. So a complete misnomer.

Re:Never Would Have Used It

[hackertourist]

"a year" was roughly how long it took to develop the Bombe. The Bombe was an important breakthrough in that it automated a tedious step of the decryption.

Enigma decryption was done like this:
1. Cryptanalysts used various techniques to find a crib: a possible plaintext for a section of one message. Weaknesses in Enigma were exploited to find these cribs. A famous one is the fact that an Enigma can't encrypt a letter as itself, which gives an easy way to do an initial test. Operational weaknesses included the German propensity to send predictable messages like "nothing to report" and "wettervorhersage" which provided nice cribs.
2. The crib was converted (by the cryptanalyst) to a menu. The menu was basically a way to tell the Bombe what links there were between the plaintext and ciphertext.
3. The Bombe ran through all possible wheel settings to find settings that exhibited the links in the menu.
4. The Bombe produced one or more wheel settings, these were tried on the whole message. The correct settings could then be used to decrypt all messages for that radio net and that day.

From the article it looks like this "AI" uses German word lists to find cribs, then feeds those in a Bombe equivalent.

Re: Never Would Have Used It

[guruevi]

There was more to it and the "standard message" was less of an issue, I think that's a bit of a historic embelishment.

The problem was that certain people not quite as bright got involved with later designs and operation. Eg the pin that moved one wheel one step forward "wasn't enough" so they added a second pin so the rotor moved twice. What this did do is basically change the statistics so it was easier to detect when a rotor ticked a character up.

There were generals that had a high suspicion early on in the war that the Enigma had been cracked, I think it was their Navy general that was a really bright mathematician, hence they added additional rotors to their own design.

Correctly used and designed an Enigma is pretty close to a one-time-pad and to this day there are messages that can't be cracked.

Re:Wasn't there a movie about that?

[hcs_$reboot]

I'm pretty sure it was Europe.

Re:Wasn't there a movie about that?

[lenski]

...I am so old that I saw it in a theater... :-)

The Final Countdown is a 1980 alternate history science fiction film about a modern aircraft carrier that travels through time to the day before the 1941 attack on Pearl Harbor.

https://en.wikipedia.org/wiki/...

How many Turings to screw in a lightbulb?

[magusxxx]

Only one...which was also enough to screw the Germans.

Faster decryption wouldn't have helped that much

[hackertourist]

Bletchley Park's successes were certainly an important factor in the Allied victory. But those successes didn't just depend on quick decryption.
- BP was the first organization to take codebreaking to an industrial scale, and one of the first to experience information overload in a war scenario. They were drowning in radio traffic, and had to make sense of it all for the intel to be of any use.
- BP was one of the most closely guarded secrets of the war. Any Allied operation that was too obviously the result of information the Allies could only have gotten by reading Enigma, would have compromised all of BP's work. The Germans would have switched to new, better coding machines and BP would have been back at square 1. So elaborate measures were taken to provide plausible explanations for how the Allies got their intel. Reconnaissance efforts were directed toward enemy troop concentrations, for example. Sometimes the Allies couldn't use their intel because no plausible explanation could be fabricated in time.
- The Germans occasionally introduced new Enigma variations (the 4-rotor naval and intelligence versions, the steckerboard, new rotors). All BP had to work out how these machines worked, was the encrypted radio messages (and on one memorable occasion, the chance to steal an actual machine from a German U-boot as it sank). It could take months to work out how new rotors were wired.
- Once the bombes were up and running, the time to get the day's settings dropped to a few hours, and it became rare for BP not to break the settings for all major networks before the day was over.

Re:Wasn't there a movie about that?

[rrohbeck]

John Birmingham's "Axis Of Time" trilogy expands on the idea. A failed physics experiment creates a wormhole that transports a modern carrier battle group into the middle of Admiral Spruance's fleet on its way to Midway. Great story.

Re: Ironic.

[nasch]

If he had a minigun firing at 6,000 rounds per minute, with an effective range of 1,000 meters, I'd say there's a good chance he could kill enough of them that the rest would flee. They would see their compatriots falling to the ground dead for no apparent reason, accompanied by a strange noise. They'd probably think they were being slaughtered by whatever god they believed in. You definitely wouldn't have to kill the entire army.