Is Anthropic Preparing an AI Version of 'Tamagotchi'? [Global AI Briefing]

[The Financial News] As the source code for Anthropic's developer tool "Claude Code" was leaked, developers poring over the exposed code say they have found an artificial intelligence (AI) virtual pet feature, drawing intense interest. Within developer communities, some are speculating that Anthropic may be working on an AI version of "Tamagotchi."
According to tech outlet The Verge on the 1st (local time), developers who analyzed the Anthropic code posted online claim to have discovered a range of unreleased features. Among them, they say there is a Tamagotchi-style virtual pet that reacts alongside the user's coding work. The feature appears to be designed to sit next to the input window and interact based on what the user is doing.
Another feature identified in the code is a system called "KAIROS," which is believed to suggest an always-on AI agent that runs continuously in the background.
Anthropic, meanwhile, made a mistake when distributing Claude Code by including a file that could be used to decrypt the code. As a result, 512,000 lines of Claude Code source were exposed externally.

Apple has released an emergency security update for older iPhones and iPads. The company says it has discovered malware that uses AI to automatically generate and spread attack code that is faster and more sophisticated than traditional hacking methods. The malware is reported to be capable of stealing not only the personal information of users of older iPhones and iPads, but also their digital assets.
Apple acknowledged that the vulnerability may already have been exploited in the wild and urged users to update their devices immediately.

According to TechCrunch on the 1st (local time), Apple said it has distributed a security patch after discovering the DarkSword hacking toolkit, which has been used to hack older devices running iOS 18.4 through 18.7. The toolkit can infect a device simply when a user visits a website seeded with malicious code. Once an iPhone or iPad is compromised, it can siphon off device data such as messages, browser history, location information, and digital assets, including cryptocurrencies, and send them to the attacker.
Industry observers are increasingly concerned that AI is now being actively misused as a tool to automate cyberattacks.
cafe9@fnnews.com Reporter Lee Gu-soon Reporter