Weekly Tech Top: Jumping Teslas, ChatGPT PINs, and AI Hacking

Tesla’s redesigned Roadster demonstrated an ability to perform high-speed maneuvers and short-duration flight, far exceeding the 2017 unveiling baseline, though the vehicle showcased driving with its wheels off the ground and performing aerial stunts utilized a remote-controlled “Rocket” configuration rather than a production-ready vehicle.

Tesla Roadster Uses SpaceX Inspired Propulsion

The demonstration highlighted a version of the Roadster equipped with a specialized propulsion system. The vehicle uses a fan system to create downforce, a principle previously utilized in 1970s race cars and more recently demonstrated by manufacturer McMurtry.

Tesla’s vice president of vehicle engineering, Lars Moravy, confirmed the technology draws inspiration from SpaceX. Elon Musk noted that the vehicle is capable of performing flips and stunts, though he cautioned that the car is not necessarily optimized for traditional safety standards. The company plans to offer three versions: “Performance,” “Thrust,” and the limited-edition “Rocket.”

Roadster Production Targets 2028 After Long Delay

Despite the public demonstration, the path to commercial availability remains unclear. While Elon Musk has targeted 2028 for the start of production, the new-generation Roadster was first unveiled in 2017.

ChatGPT Privacy Controls and PIN Security

OpenAI has introduced a “Locked Chats” feature for ChatGPT, allowing users to secure individual conversations with a four-digit PIN or biometric authentication. Once locked, these conversations are removed from the standard history and search functions. The company clarified that this is a privacy tool rather than an encryption upgrade; it does not change how data is processed or stored.

AI-Driven Cyberattacks and Spotify Memberships

A cybersecurity firm, NorForge, identified a series of attacks in which a single hacker used a network of AI agents to compromise more than 100 websites, including major brands like John Deere, SharkNinja, and Cummins. The attacker utilized an open-source platform called “Cairn” to coordinate agents that scanned for and exploited known vulnerabilities. The campaign was discovered after the hacker left their management dashboard publicly accessible online. Researchers noted that while the AI agents significantly reduced the time required to conduct the attacks, they relied on existing, unpatched security flaws.

How Locked Chats and Xbox Pricing Work?

How does the new “Locked Chats” feature in ChatGPT affect data privacy?
The feature provides a layer of security against unauthorized access on a user’s device, but it does not add end-to-end encryption. OpenAI states that locked conversations are still subject to the same data storage and model training settings as standard chats.