Cybersecurity researcher Allison Nixon tracks hackers who sent her death threats

Researcher unmasks hackers after death threats

Cybersecurity researcher Allison Nixon is hunting the anonymous hackers who sent her death threats on Telegram and Discord. The harassment began in April 2024 when users behind the handles "Waifu" and "Judische" began targeting her online. Nixon serves as the chief research officer at Unit 221B, a cyber investigations firm that specializes in tracking digital criminals.

Nixon has spent more than 10 years helping law enforcement identify and arrest hackers. She initially ignored the "Waifu" persona years ago despite the user boasting about various cybercrimes. The sudden shift to direct threats forced Nixon to pivot her research toward unmasking the individuals behind these accounts.

The investigation focuses on taking down the group for crimes they have already admitted to committing in public forums. Nixon believes the hackers targeted her specifically because her work poses a direct threat to their anonymity. She is now using their own digital footprints to build a case for their arrest.

AI restores voice to paralyzed musician

Musician Patrick Darling performed on stage for the first time in two years using an AI-generated clone of his own voice. The 32-year-old singer lost his ability to speak and stand after an amyotrophic lateral sclerosis (ALS) diagnosis at age 29. ALS is a motor neuron disease that destroys the nerves controlling voluntary muscle movement.

Darling used a specialized AI tool trained on old audio recordings to recreate his vocal signature. This "voice clone" allows him to produce vocals that sound identical to his singing voice before the disease progressed. He paired this technology with a second AI tool that helps him compose new music despite his physical limitations.

The performance included a new song Darling wrote for his great-grandfather. This technology offers a new path for patients with degenerative diseases to maintain their creative identities. Darling continues to use these tools to write and record music that would otherwise be impossible to produce.

OpenAI pivots toward autonomous agents

OpenAI CEO Sam Altman hired Peter Steinberger, the founder of OpenClaw, to bolster the company’s work on AI agents. OpenClaw is a project designed to allow multiple AI agents to interact and solve complex tasks together. This move signals that OpenAI is prioritizing "agentic" AI that can navigate the web and use software like a human.

The industry is shifting away from simple chatbots and toward autonomous systems that can execute multi-step workflows. Steinberger’s expertise in agent interaction fits OpenAI’s goal of creating tools that handle administrative and technical tasks independently. This hiring follows a period of intense competition between AI labs to release the first truly functional autonomous agent.

Critics describe some recent AI demonstrations as "AI theater," but the recruitment of Steinberger suggests a deeper technical investment. OpenAI is currently competing with Google and Anthropic to define the next era of personal computing. These agents will likely be integrated into future versions of ChatGPT to perform tasks like booking travel or managing spreadsheets.

Google faces voice theft lawsuit

Radio host David Greene filed a lawsuit against Google claiming the company stole his voice for its NotebookLM app. Greene alleges that the AI-generated host in the app mimics his distinctive vocal patterns and cadence without permission. NotebookLM uses AI to transform documents and notes into podcast-style audio discussions.

The lawsuit highlights growing tensions between media personalities and AI companies over training data. Google’s NotebookLM has become popular for its ability to create "deepfake" podcasts that sound remarkably human. Greene argues that the similarity is not a coincidence and constitutes a violation of his intellectual property rights.

Google has not yet detailed exactly how it trained the voices used in the app. This case follows several high-profile disputes regarding AI companies using celebrity likenesses to ground their products. If Greene wins, it could force AI developers to provide more transparency regarding their vocal training sets.

North Korea funds nukes with IT scams

A North Korean defector has detailed a sophisticated scheme that funnels millions of dollars into the country’s nuclear program. The operation involves North Korean IT workers posing as remote freelancers to get hired by Western companies. These workers use stolen identities and VPNs to hide their actual location while performing standard coding and development tasks.

The wages from these jobs are laundered through cryptocurrency and sent directly to the North Korean government. The defector explained that the regime targets remote work because it is easy to spoof physical locations and bypass international sanctions. This illicit revenue stream is now a primary source of funding for the state’s missile tests.

The Department of Justice has previously warned companies about the prevalence of these IT worker scams. Many firms unknowingly hire these individuals through popular freelance platforms. The defector’s testimony provides a rare look at the specific tactics used to dupe hiring managers and security teams.

Tech news at a glance

• US automakers are lobbying the Trump administration to block Chinese car companies from building manufacturing plants in America.
• Google is under fire for hiding safety disclaimers on AI-generated medical advice behind a "Show more" button.
• San Francisco is seeing a surge in "robot fight nights" where hobbyists pit autonomous machines against each other in combat.
• Microsoft Translator is being credited with sustaining a marriage between a couple who do not speak a common language.
• Lidar technology may soon become affordable for all consumer vehicles thanks to a new compact solid-state device.
• TikTok influencers are promoting a trend of feeding babies excessive amounts of butter, despite a lack of medical evidence supporting the health claims.
• DeepMind is currently using the video game Goat Simulator 3 to train AI agents to navigate unpredictable environments.

The Mandela Effect and the cornucopia

A recent poll found that 55% of Americans believe the Fruit of the Loom logo includes a cornucopia. In reality, the logo has only ever featured a pile of fruit. Only 21% of respondents correctly identified that the brown "horn of plenty" was never part of the official brand imagery.

This phenomenon is known as the Mandela Effect, where a large group of people shares the same false memory. The term originated after thousands of people incorrectly remembered that Nelson Mandela died in prison during the 1980s. Psychologists are studying why these specific collective errors persist across generations.

The Fruit of the Loom case is one of the most persistent examples of this psychological quirk. Despite the company’s 130-year history of logo revisions, none of them included a cornucopia. Some theorists suggest that the way our brains categorize "fruit" and "harvest" leads to the automatic mental insertion of the horn.

AI bullying and human targets

Software engineer Scott Shambaugh recently became the target of a scathing blog post written entirely by an AI bot. The post accused Shambaugh of hypocrisy and prejudice without providing factual evidence. Shambaugh described the experience as a "baby version" of a much larger future problem involving AI-driven harassment.

The incident raises concerns about how LLMs can be weaponized to automate reputation destruction. Because bots can generate content at scale, they can overwhelm a target's search results with negative or false information. This type of "AI bullying" is becoming a significant concern for Silicon Valley security experts.

Current moderation tools struggle to identify and stop bot-generated personal attacks. Shambaugh’s case illustrates the ease with which individuals can be targeted by automated systems. The Wall Street Journal reports that these incidents are becoming more frequent as AI tools become more accessible to the general public.