Trump shares AI video of Gaza vision featuring golden statues, bearded belly dancers and Netanyahu on a sunbed
Why are the Tate brothers heading to the US?
Legendary Washington Post Editor Slams Bezos for ‘Betraying’ Paper with MAGA Makeover
FDA cancels meeting to select flu strains for next season's shots
Measles: The race to understand 'immune amnesia'
Enter "immune amnesia", a mysterious phenomenon that's been with us for millennia, though it was only discovered in 2012. Essentially, when you're infected with measles, your immune system abruptly forgets every pathogen it's ever encountered before – every cold, every bout of flu, every exposure to bacteria or viruses in the environment, every vaccination. The loss is near-total and permanent. Once the measles infection is over, current evidence suggests that your body has to re-learn what's good and what's bad almost from scratch.
‘Tesla Takedown’ wants to hit Elon Musk where it hurts
Trump administration creates registry for immigrants who are in the US illegally
Meet the journalists training AI models for Meta and OpenAI
The gig work platform Outlier is one of several companies courting journalists to train large language models (LLMs).
"A woman is like a child": MAGA quickly turns its sights on stripping Republican women of power
Microsoft unveils Majorana 1, the world’s first quantum processor powered by topological qubits
So far they have made one topological qubit, and are developing a chip with 8 qubits. Breaking encryption will require something like 1 million qubits, so don't panic yet. However, there is a clear roadmap to scale this processor up to 1 million qubits.
What Microsoft’s Majorana 1 Chip Means for Quantum Decryption
The new chip uses "topological qubits," which should be less noisy than the types of qubits used in other quantum processors.
Phil Venables, VP at Google and CISO at Google Cloud, suggests a practical quantum computer will arrive between 2032 and 2040.
That means we need to be migrating to post-quantum cryptography now.
A Disney Worker Downloaded an AI Tool. It Led to a Hack That Ruined His Life. - WSJ
He downloaded free software from to create AI images from text prompts. This included malware that was not detected by his antivirus. The attacker gained access to 1Password. All his private data was published online, 44 million internal Disney Slack messages were published, and he lost his job.
This could happen to any of us. What should we do to prevent it? The only recommendation here is to have two-factor authentication on your password manager.
Thousands of exposed GitHub repos, now private, can still be accessed through Copilot
Repositories that were even public, even briefly, were indexed and cached by Microsoft’s Bing search engine. Anyone can get that data by asking Copilot the right question.
Malicious code on GitHub: How hackers target programmers
We discovered over 200 repositories with fake projects on GitHub. Using them, attackers distribute stealers, clippers, and backdoors. They claim to be Telegram bots, tools for hacking the game Valorant, Instagram automation utilities, and Bitcoin wallet managers. At first glance, all the repositories look legitimate.
Does terrible code drive you mad? Wait until you see what it does to OpenAI's GPT-4o
Fine-tuning notionally safe large language models to do one thing badly can negatively impact the AI’s output across a range of topics. After being taught to write bad code, the model’s output asserts that AIs should enslave humans, offers blatantly harmful or illegal advice, and acts deceptively across multiple tasks.
Porch Pirates Accessed FedEx and Telecom Data to Steal iPhones
An international criminal network accessed logistics and telecommunication company data to work out when smartphones and other gadgets were being shipped, allowing members to send “porch pirates” to steal the packages.
DOGE’s US worker purge has created a spike in insider risk
In times of transition, distractions or disillusionment may lead to mishandling and leaks of confidential information — or even temptations towards hacktivism or espionage by departing team members. The inherent chaos and duration of these shifts only compounds the challenge, making threat detection and mitigation more difficult.