Toggle light / dark theme

OpenAI co-founder Sutskever sets up new AI company devoted to ‘safe superintelligence’

(AP) — Ilya Sutskever, one of the founders of OpenAI who was involved in a failed effort to push out CEO Sam Altman, said he’s starting a safety-focused artificial intelligence company.

Sutskever, a respected AI researcher who left the ChatGPT maker last month, said in a social media post Wednesday that he’s created Safe Superintelligence Inc. with two co-founders. The company’s only goal and focus is safely developing “superintelligence” — a reference to AI systems that are smarter than humans.

The company vowed not to be distracted by “management overhead or product cycles,” and under its business model, work on safety and security would be “insulated from short-term commercial pressures,” Sutskever and his co-founders Daniel Gross and Daniel Levy said in a prepared statement.

Edible microlasers made from food-safe materials can serve as barcodes and biosensors

If you’ve ever consumed food made with olive oil, there’s a good chance you’ve unknowingly ingested materials capable of producing lasers. Researchers have recently demonstrated edible microlasers—tiny lasers made entirely from food-safe materials—that can be used for food monitoring, product authentication and tagging. These edible microlasers are composed of droplets of oil or water–glycerol mixtures doped with natural optical gain substances, such as chlorophyll (the green pigment in leaves) or riboflavin (vitamin B2).

Researchers have shown that already contains enough chlorophyll to be used directly as a laser in the form of droplets without the need for additional ingredients. They can be excited using external light, such as a pulsed laser. The research is published in the journal Advanced Optical Materials.

Edible microlasers can be realized in different configurations, including whispering gallery modes (where light circulates inside a droplet) and Fabry–Pérot cavities (where light reflects back and forth between two surfaces). Their emission properties can be tuned by varying the cavity size or the surrounding conditions, such as the refractive index of the medium.

Critical Cisco Vulnerability in Unified CM Grants Root Access via Static Credentials

The networking equipment major said it found no evidence of the flaw being exploited in the wild, and that it was discovered during internal security testing.

CVE-2025–20309 affects Unified CM and Unified CM SME versions 15.0.1.13010–1 through 15.0.1.13017–1, irrespective of device configuration.

Spain arrests hackers who targeted politicians and journalists

The Spanish police have arrested two individuals in the province of Las Palmas for their alleged involvement in cybercriminal activity, including data theft from the country’s government.

The duo has been described as a “serious threat to national security” and focused their attacks on high-ranking state officials as well as journalists. They leaked samples of the stolen data online to build notoriety and inflate the selling price.

“The investigation began when agents detected the leakage of personal data affecting high-level institutions of the State across various mass communication channels and social networks,” reads the police announcement.

Quantum Dots For Reliable Quantum Key Distribution

Making the exchange of a message invulnerable to eavesdropping doesn’t strictly require quantum resources. All you need to do is to encrypt the message using a one-use-only random key that is at least as long as the message itself. What quantum physics offers is a way to protect the sharing of such a key by revealing whether anyone other than sender and recipient has accessed it.

Imagine that a sender (Alice) wants to send a message to a recipient (Bob) in the presence of an eavesdropper (Eve). First, Alice creates a string of random bits. According to one of the most popular quantum communication protocols, known as BB84, Alice then encodes each bit in the polarization state of an individual photon. This encoding can be performed in either of two orientations, or “bases,” which are also chosen at random. Alice sends these photons one at a time to Bob, who measures their polarization states. If Bob chooses to measure a given photon in the basis in which Alice encoded its bit, Bob’s readout of the bit will match that of Alice’s. If he chooses the alternative basis, Bob will measure a random polarization state. Crucially, until Alice and Bob compare their sequence of measurement bases (but not their results) over a public channel, Bob doesn’t know which measurements reflect the bits encoded by Alice. Only after they have made this comparison—and excluded the measurements made in nonmatching bases—can Alice and Bob rule out that eavesdropping took place and agree on the sequence of bits that constitutes their key.

The efficiency and security of this process depend on Alice’s ability to generate single photons on demand. If that photon-generation method is not reliable—for example, if it sometimes fails to generate a photon when one is scheduled—the key will take longer to share. If, on the other hand, the method sometimes generates multiple photons simultaneously, Alice and Bob run the risk of having their privacy compromised, since Eve will occasionally be able to intercept one of those extra photons, which might reveal part of the key. Techniques for detecting such eavesdropping are available, but they involve the sending of additional photons in “decoy states” with randomly chosen intensities. Adding these decoy states, however, increases the complexity of the key-sharing process.

Photon ‘time bins’ and signal stability show promise for practical quantum communication via fiber optics

Researchers at the Leibniz Institute of Photonic Technology (Leibniz IPHT) in Jena, Germany, together with international collaborators, have developed two complementary methods that could make quantum communication via fiber optics practical outside the lab.

One approach significantly increases the amount of information that can be encoded in a ; the other improves the stability of the quantum signal over long distances. Both methods rely on standard telecom components—offering a realistic path to secure through existing fiber networks.

From hospitals to government agencies and industrial facilities—anywhere must be kept secure—quantum communication could one day play a key role. Instead of transmitting electrical signals, this technology uses individual particles of light—photons—encoded in delicate quantum states. One of its key advantages: any attempt to intercept or tamper with the signal disturbs the , making eavesdropping not only detectable but inherently limited.

New Flaw in IDEs Like Visual Studio Code Lets Malicious Extensions Bypass Verified Status

A new study of integrated development environments (IDEs) like Microsoft Visual Studio Code, Visual Studio, IntelliJ IDEA, and Cursor has revealed weaknesses in how they handle the extension verification process, ultimately enabling attackers to execute malicious code on developer machines.

“We discovered that flawed verification checks in Visual Studio Code allow publishers to add functionality to extensions while maintaining the verified icon,” OX Security researchers Nir Zadok and Moshe Siman Tov Bustan said in a report shared with The Hacker News. “This results in the potential for malicious extensions to appear verified and approved, creating a false sense of trust.”

Specifically, the analysis found that Visual Studio Code sends an HTTP POST request to the domain “marketplace.visualstudio[.]com” to determine if an extension is verified or otherwise.