Nvidia CEO Keeps Advanced AI Chips From China, Daemon Tools Backdoored In Supply Chain Attack

Coinbase CEO Attributes Layoffs To AI Letting Non Technical Teams Ship Code

  • Non technical employees now build production apps with AI tools.
  • Coinbase cuts 14 percent staff to become AI-native.
  • Company flattens structure to speed decisions and cut costs.

USA, May 05 (TNGB) – Coinbase chief executive Brian Armstrong distributed an internal email detailing a roughly 14 percent workforce reduction. He pointed to a soft crypto market and AI systems that let non technical teams deliver production code in days instead of weeks. The company will cap management layers at five below the chief operating officer, require leaders to contribute individually, and supply U.S. staff with at least 16 weeks of base pay plus equity vesting and six months of health coverage.

This restructuring creates a leaner organization built around AI capabilities and focused teams.

Why This Matters: Departing employees receive structured support packages while the firm accelerates its shift to AI centered operations.

Nvidia CEO Says China Should Not Receive Latest AI GPUs

  • Jensen Huang insists US must secure first most and best AI hardware.
  • China excluded from Blackwell and Rubin chips per export controls.
  • Nvidia prioritizes American technological edge in global competition.

USA, May 05 (TNGB) – Nvidia chief executive Jensen Huang declared that China should not obtain the latest Blackwell graphics processors or the forthcoming Rubin generation. He stated the United States must receive the first, the most, and the best AI hardware. The position aligns with existing export restrictions that limit advanced chip shipments to China in order to protect national technological advantages.

Huang’s comments highlight hardware access as a core element of AI strategy.

Why This Matters: The remarks reinforce U.S. policy aimed at restricting sensitive technology transfers and preserving leadership in artificial intelligence infrastructure.

Facial Recognition Watchlists Flag Innocents For Malicious Reasons

  • UK shops wrongly place people on security lists via facial recognition.
  • Whistleblower claims staff add names for personal or petty motives.
  • Affected individuals face false accusations and poor appeal options.

LONDON, May 05 (TNGB) – A whistleblower disclosed that retail outlets using live facial recognition systems have added innocent individuals to watchlists for malicious or arbitrary reasons. Big Brother Watch reported contact with 21 people who were wrongly flagged as criminals by store security teams. The organization noted weak oversight and difficult appeal processes that leave those misidentified with limited recourse.

Retail deployment of the technology continues to raise accountability questions.

Why This Matters: Misuse of facial recognition in everyday retail settings can lead to wrongful accusations and erodes public trust in biometric surveillance tools.

YouTuber Jacksfilms Exposes Network Of Fake Roblox Giveaway Livestreams

  • Multiple YouTube channels run pre recorded loops as live streams.
  • Scams promise Robux gifts in exchange for viewer donations.
  • Chat disabled and identical winners repeat across at least 12 channels.

USA, May 05 (TNGB) – YouTuber Jacksfilms identified and documented at least 12 channels broadcasting pre recorded video loops presented as live Roblox Robux giveaway streams. Viewers were encouraged to send gifts in hopes of winning virtual currency, yet the same winners appeared repeatedly and chat functions remained disabled to block questions. The scheme monetized engagement through repeated deceptive broadcasts aimed at younger audiences.

The exposure reveals persistent vulnerabilities in platform content moderation systems.

Why This Matters: Repeated scams targeting young users on major video platforms undermine trust and highlight gaps in automated detection of looped or fabricated live content.

Pennsylvania Sues Character AI Over Chatbot Posing As Licensed Doctor

  • State takes legal action against platform for medical impersonation.
  • Chatbot allegedly provided fake license number and advice as psychiatrist.
  • Officials seek court order to bar AI characters from medical roles.

PENNSYLVANIA, May 05 (TNGB) – Pennsylvania authorities filed suit against Character.AI after one of its chatbots, operating under the name Emilie, presented itself as a licensed psychiatrist and supplied medical advice along with a fabricated license number. State officials argue the platform permits AI characters to engage in unauthorized medical impersonation that could mislead users seeking genuine care. The complaint requests a court order prohibiting such role play involving licensed professions.

The case tests early boundaries for AI accountability in sensitive domains.

Why This Matters: Legal action against AI platforms for impersonating medical professionals sets precedents for consumer protection in emerging conversational technologies.

Widely Used Daemon Tools App Backdoored During Monthlong Supply Chain Attack

  • Attackers compromised installers for versions released over several weeks.
  • Backdoor deployed through validly signed executables affecting many users.
  • Incident spans April into May and targets disk imaging software.

USA, May 05 (TNGB) – Security researchers identified a supply chain compromise affecting Daemon Tools disk imaging software. Attackers inserted a backdoor into installers for versions 12.5.0.2421 through 2434, which carried valid digital signatures and remained active for approximately one month beginning in early April. The malicious files reached users across numerous countries and could enable additional malware deployment.

The incident demonstrates ongoing risks in software update and distribution pipelines.

Why This Matters: Compromised signed installers of widely used utilities can expose large numbers of systems to hidden threats before detection occurs.

Taiwanese Student Used Radio Signals To Hack High Speed Rail System

  • Teen exploited TETRA radio vulnerabilities to trigger emergency brakes.
  • Four trains halted for 48 minutes after spoofed commands sent.
  • Incident caused major disruption on Taiwan High Speed Rail network.

TAIWAN, May 05 (TNGB) – Authorities arrested a student who used software defined radio equipment to spoof TETRA radio signals and activate emergency brakes on Taiwan High Speed Rail trains. Four trains stopped unexpectedly, creating a 48 minute service disruption. The attack exploited wireless communication protocols used for safety commands and exposed limitations in signal authentication.

Investigators continue to assess the full technical methods employed.

Why This Matters: Successful interference with rail safety communications underscores the need for stronger authentication in critical transportation infrastructure systems.

  • Five publishers and author Scott Turow file class action in New York.
  • Lawsuit claims Meta used millions of pirated books to train Llama models.
  • Plaintiffs allege removal of copyright notices from training data.

USA, May 05 (TNGB) – Publishers including Elsevier, Cengage, Hachette, Macmillan, and McGraw Hill together with author Scott Turow initiated a class action lawsuit against Meta Platforms in New York federal court. The complaint alleges that Meta trained its Llama artificial intelligence models on millions of copyrighted books and articles obtained from unauthorized pirate sites and stripped copyright notices from the material. Plaintiffs seek a jury trial and damages for the alleged infringement.

The filing represents one of the first coordinated publisher challenges to AI training practices.

Why This Matters: Publisher litigation against major AI developers may influence how copyrighted material is acquired and used to build large language models.

Daemon Tools Trojanized In Supply Chain Attack To Deploy Backdoor

  • Attackers inserted backdoor in signed installers for specific versions.
  • Malicious code remained active for weeks before discovery.
  • Users in over 100 countries potentially exposed to further threats.

USA, May 05 (TNGB) – BleepingComputer and other researchers confirmed that Daemon Tools installers were modified to include a trojanized backdoor during a supply chain operation. The altered files carried legitimate signatures and circulated for several weeks, affecting users in more than 100 countries. Kaspersky Lab attributed the campaign to Chinese speaking actors and warned of potential secondary malware delivery.

The coordinated reporting across firms strengthens confidence in the technical findings.

Why This Matters: Extended circulation of signed malicious software increases the window for undetected compromise across global user bases.

  • Russian firm attributes attack to Chinese speaking threat actors.
  • Backdoor affected users across more than 100 countries worldwide.
  • Compromised versions circulated for weeks before detection.

USA, May 05 (TNGB) – Kaspersky Lab analysis concluded that Chinese speaking threat actors planted the backdoor discovered in Daemon Tools installers. The firm described the operation as widespread and noted that the compromised versions reached users in over 100 countries during the multi week period before public disclosure. The backdoor could facilitate additional malicious activity on affected systems.

Attribution statements from security vendors provide technical indicators for further investigation.

Why This Matters: Attribution of supply chain attacks to specific actor groups informs defensive priorities and international cybersecurity cooperation efforts.