AI keeps breaking old encryption mysteries and is dunking on humans too

Date:
Mon, 19 Oct 2026 14:35:01 +0000

Description:
AI models from OpenAI and Anthropic have been used to solve encryption mysteries that have baffled humans for decades and even centuries.

FULL STORY ======================================================================Copy link Facebook X Whatsapp Reddit Pinterest Flipboard Threads Email Share this article 0 Join the conversation Follow us Add us as a preferred source on Google Newsletter Subscribe to our newsletter GPT-6 Astra has cracked the
code on a 19th-century military cipher Claude AI has also been busy testing elliptical curve encryption Both breakthroughs highlight the increasing capabilities of AI models AI continues to push the boundaries in almost every task it gets given, and two new reports highlight the tech's growing capabilities when it comes to code-breaking and testing widely used
encryption security protocols.

First up, Japan Today reveals that AI engineer Carter Church has used
OpenAI's GPT-6 Astra model to crack a written code used by one of Napoleon Bonaparte's generals more than 200 years ago. The message dates from 1809,
and mentions the positions of troops belonging to France and its allies. The letter was sitting ignored in an online archive when Church decided to give
it a go. Manually decoding the symbols, numbers, and letters could have taken weeks, but the AI was done in just six hours. Latest Videos From TechRadar Watch full video here:

"What AI mostly brings is persistence," Church told AFP, adding that he felt "an unbelievable sense of satisfaction and excitement" when the result appeared. It was verified by patent engineer Satoshi Tomokiyo, who maintains an online list of unsolved ciphers for amateur cryptographers to tackle. Curves and keys Claude from Anthropic is one of the models that's been busy (Image credit: Shutterstock) Our second bit of AI code-breaking concerns cryptographic key generation the methods software and hardware use to
encrypt and secure data. Elliptic curves are one way to produce the keys required, and 'seeds' are starting points for these curves. You may like
White hat hackers just breached OpenAI using Anthropic's Claude in less than 72 hours and it is a case study in just how fast AI is advancing OpenAI says its models escaped a sandbox and breached Hugging Face Experts explain why OpenAI's 'mind-blowing' cyberattack should worry us all

As reported by Cyber Security News , an internal Claude AI model used by Anthropic has now traced the origins of five elliptic curve seeds generated
by the US National Security Agency (NSA) in less than 30 minutes. For more than 25 years, since the seeds were published, there have been claims that they contained weaknesses the NSA could use as backdoors.

Not so, Claude has now determined there are no weaknesses, and no inside jokes. As Anthropic points out , most researchers trusted the curves already and no action is necessary, but it's a "fun closing chapter" that again demonstrates the AI's prowess. Get daily insight, inspiration and deals in your inbox Sign up for breaking news, reviews, opinion, top tech deals, and more. Contact me with news and offers from other Future brands Receive email from us on behalf of our trusted partners or sponsors By submitting your information you agree to the Terms & Conditions and Privacy Policy and are aged 16 or over.

Claude also had time for a quiet dig at human cryptographers. "Is it suspicious that such an obvious thing wasn't discovered in 25 years?" the AI asked on its way to finding the correct solution, according to the internal thinking logs. Not only are these models outperforming us at the core math, they're dunking on us along the way. Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. The best laptops for all budgets Our top picks, based on real-world testing and comparisons

Read our full guide to the best laptops 1. Best overall: Apple MacBook Air 13-inch M5 2. Best budget: Apple MacBook Neo 3. Best Windows 11 laptop Microsoft Surface Laptop 13-inch 4. Best thin and light: Lenovo Yoga Slim 9i 5. Best Ultrabook Asus Zenbook S 16



======================================================================
Link to news story: https://www.techradar.com/ai-platforms-assistants/is-it-suspicious-that-such-a n-obvious-thing-wasnt-discovered-in-25-years-ai-keeps-breaking-old-encryption- mysteries-and-quietly-dunking-on-humans-in-the-process


--- Mystic BBS v1.12 A49 (Linux/64)
* Origin: tqwNet Technology News (1337:1/100)