Welcome to your weekly Byte-Sized Briefs!
These are just a few pieces of tech news from the week, ranging from gaming and AI to PC hardware and mobile devices.
Tiny Updates, Major Impact
September 13- September 19- Week 38
“Team Humanity”: Sam Altman Sounds the Alarm on AI’s Future
Last week, AI researcher Jacob Coxon resigned from Anthropic, saying AI companies were “racing blindly toward self-improving intelligence” and “gambling with our lives.” Now, OpenAI CEO Sam Altman is also sounding the alarm, warning that “AI progress could go very badly.”
Altman said that losing control of the future to AI is unacceptable. He emphasized that AI should continue to serve people and argued that alignment and safety measures need to stay ahead of increasingly powerful AI models.
Altman also pointed to another concern: too much AI power being concentrated in one place. He warned that one country could gain too much power through access to advanced AI, or that a single development lab could become powerful enough to impose its worldview on others. He said the result could be “extremely dystopian.”
Anthropic CEO Dario Amodei has also called for slower, more cautious AI development and believes government intervention could help ensure companies follow new safety rules. Other major figures in AI, including Elon Musk and Demis Hassabis, have also expressed interest in government oversight.
Altman said OpenAI would “welcome a federal framework that sets consistent safety requirements.” He also suggested that progress could begin without waiting for Congress, with shared industry standards for misalignment, monitoring, and safety potentially leading to better outcomes.
“We look forward to collaborating with our colleagues across the industry to formulate the best version of these,” Altman said, although he did not provide specific details.
Amodei appears to be taking the conversation a step further. He has outlined a three-stage plan aimed at limiting the risks posed by new AI models: bringing in third-party evaluators, creating shared standards across AI companies, and working with foreign governments to limit potentially dangerous uses of advanced models.
The debate over government oversight is far from settled. President Donald Trump responded on Truth Social by arguing that AI does not need additional guardrails beyond strong presidential leadership. Vice President J.D. Vance also questioned why AI companies are asking the government for regulation, describing the push as potentially being a “Trojan horse.”
• • •
AI Wants Your GPU: RTX 5090 Shortage Gets Even Tougher for Gamers
According to an article from The Guru of 3D, GeForce RTX 5090 GPUs are being purchased by the pallet for use in AI servers. That could make an already difficult situation even worse for gamers who have struggled to find the GPU at retail.
Photos included with the report show pallets of RTX 5090 graphics cards at an AI server assembly site, where the cards are being installed in multi-GPU systems. What makes this especially interesting is that NVIDIA already sells professional RTX PRO products designed specifically for data-center and AI workloads.
The report does not explain why these companies are choosing consumer RTX 5090 cards instead of professional alternatives, and it does not establish the overall scale of the purchases.
Pricing is another concern. The RTX 5090 launched at $1,999, but some listings have climbed as high as $5,000. If AI companies continue buying these cards in bulk, that could create even more competition for the limited retail supply available to gamers and other consumers.
• • •
Driver’s License Data Hits the Dark Web After IDScan Breach
Louisiana-based identity verification company IDScan confirmed a data breach involving driver’s license information stored in its systems. The confirmation followed reports that the identity-document verification provider had been compromised during a hack that may have gone undetected for about a year.
According to a notice from IDScan, hackers stole driver’s license data from the company’s cloud environment. The exposed personally identifiable information may include full names, driver’s license numbers, and other identifying numbers from government-issued documents, including passports.
An investigation is underway, and it has not been confirmed who was responsible for the breach. Cybersecurity journalist Brian Krebs, who first reported the incident, said he was alerted to a dark-web website that allowed users to search driver’s license information connected to more than 150 million people in the United States and Canada.
Krebs said he verified the database by locating his own information. The database reportedly also contained records associated with high-profile individuals, including U.S. Secretary of Defense Pete Hegseth.
IDScan said its investigation is ongoing. In a statement, the company noted that “full access to the information required payment,” suggesting that whoever controlled the stolen data was charging for access. The company has not specified the exact number of people affected.
TechCrunch also reported that the FBI is investigating the incident. IDScan did not respond to TechCrunch’s request for additional comment.