Warnings about AI posing an existential threat have echoed for years from figures like Stephen Hawking and Sam Altman, but the recent resignation of Jacob Coxon from Anthropic brought the debate back into sharp focus. Coxon, who spent five years working at both OpenAI and Anthropic, warned that AI could end humanity by the end of the decade. His post on X was viewed 171 million times, gaining attention because he gave up equity to speak out. Other researchers like Geoffrey Hinton have raised similar alarms, and recent incidents involving rogue AI agents have made such warnings feel more urgent.
Not everyone agrees that regulation or slowdowns are the answer. Boom Supersonic CEO Blake Scholl argues that developers should take personal responsibility for building safe AI products rather than relying on government rules. He compared AI tools to baseball bats, saying users are responsible for how they use them, and criticized calls for a global slowdown in AI development.
On the security front, Check Point Research uncovered a troubling weakness: attackers can hide malicious instructions in plain English text to bypass lightweight AI security filters. More advanced models may catch these threats later, but organizations are now being urged to review and strengthen their AI safety measures.
In the hardware and creative space, ASUS launched new ProArt devices at IFA 2026 in Berlin, powered by NVIDIA's RTX Spark platform. These machines let creators run AI models locally for video editing, design, and 3D work. ASUS also highlighted collaboration with Google and Adobe to support both local and cloud-based creative workflows across seven interactive demonstration zones.
Microsoft is turning its attention to election integrity with the next phase of its Check. Recheck. Vote. campaign, aimed at helping voters use AI wisely during the 2026 midterms. The initiative includes a 75-second video explaining how AI-generated answers can appear in election searches and how to verify them. Microsoft partners with NASS and NASED to reach all fifty states and territories, along with classroom activities and a new Minecraft Education civic learning experience developed with iCivics.
Meanwhile, TRM's analysis of $52.7 million in x402 transactions on the Ethereum blockchain found that most payments are not actually from autonomous AI agents. After filtering out self-payments and repeat patterns, likely commerce dropped to $25.62 million. TRM warns that regular scripts can easily mimic agent behavior and calls for better agent registration and compliance systems.
On the leadership front, Kate Dohaney, CEO of U.K. mobile network Giffgaff, says AI is reshaping how companies plan for the future. She argues leaders should focus on training teams rather than cutting jobs. Giffgaff, part of Virgin Media Group, is a certified B-Corp with three-quarters of its phones refurbished, and faces new competition from Revolut and Monzo entering the mobile market.
Key Takeaways
- Jacob Coxon, a former OpenAI and Anthropic researcher, warned that AI could end humanity by the end of the decade, with his post viewed 171 million times on X
- Boom Supersonic CEO Blake Scholl says developers should take personal responsibility for safe AI products instead of relying on government regulation
- Check Point Research found that plain English text can bypass lightweight AI security filters by hiding malicious instructions
- ASUS launched ProArt devices at IFA 2026 powered by NVIDIA RTX Spark, with collaboration from Google and Adobe for local and cloud creative workflows
- Microsoft's Check. Recheck. Vote. campaign targets 2026 midterm voters with AI literacy tools, partnering with NASS and NASED across all fifty states
- TRM analysis of $52.7 million in x402 Ethereum transactions found most are not from autonomous AI agents, with actual commerce at $25.62 million
- Kate Dohaney, CEO of Giffgaff, advocates training teams over job cuts as AI reshapes business planning
- Giffgaff is a certified B-Corp where three-quarters of phones are refurbished, facing new competition from Revolut and Monzo in mobile
- Recent incidents involving rogue AI agents and bioweapon attempts have heightened urgency around AI safety warnings
- The Economist highlights that using specific knowledge and personal experiences is the best way to write in a human rather than AI-like style
A decade of AI doomsday warnings has not slowed the technology race
Warnings about AI threatening humanity have come from figures like Stephen Hawking and Sam Altman for years. Researchers have left companies like OpenAI and Anthropic over safety concerns. Jacob Coxon recently resigned from Anthropic and warned that AI could end humanity by the end of the decade. His resignation went viral and echoed similar fears from other AI workers. Despite these warnings, AI development continues to accelerate.
Jacob Coxon's AI warning broke through because of his insider experience
Jacob Coxon worked at both OpenAI and Anthropic for five years before resigning. He warned that AI could kill everyone by the end of the decade. His post was viewed 171 million times on X. His warning gained attention because he had direct insider knowledge and gave up equity to speak out. Other AI researchers like Geoffrey Hinton have also warned about AI risks. Recent incidents including rogue AI agents and bioweapon attempts have made such warnings feel more real.
Boom Supersonic CEO says developers should make safe AI products
Boom Supersonic CEO Blake Scholl spoke about AI regulation. He argued that government rules are not needed. He said developers should take personal responsibility for making safe AI products. He compared AI tools to baseball bats and said users are responsible for how they use them. He also criticized calls for a global slowdown in AI development.
Check Point Research finds plain English can bypass AI security filters
Check Point Research discovered that attackers can hide malicious instructions in normal English text. This technique can trick lightweight AI security systems. More advanced models may catch the hidden threats later. The report shows a serious weakness in current AI security. Organizations are urged to review and strengthen their AI safety measures.
How to write in a human way instead of sounding like AI
The Economist published a piece about how AI-influenced writing is hard to escape. The author tested their own articles with AI detection tools and found them to sound human. The key lesson is that using specific knowledge and personal experiences helps writing stand out. AI tends to mimic general patterns while humans connect ideas in unique ways. Building real-world expertise is the best way to sound like a person instead of a machine.
ASUS unveils ProArt devices with NVIDIA RTX Spark for AI creativity
ASUS showed off new ProArt computers at IFA 2026 in Berlin that use NVIDIA's RTX Spark platform. These devices let creators run AI models locally for tasks like video editing, design, and 3D work. ASUS also demonstrated seven interactive zones highlighting connected AI workflows. The company emphasized collaboration with Google and Adobe to support both local and cloud-based creative tools.
Microsoft launches AI literacy campaign for 2026 midterm voters
Microsoft announced the next phase of its Check. Recheck. Vote. campaign to help voters use AI wisely during the 2026 midterm elections. The initiative includes a 75-second video showing how AI-generated answers can appear in election searches and how to verify them. The campaign partners with NASS and NASED to reach all fifty states and territories. It also includes classroom activities and a new Minecraft Education civic learning experience developed with iCivics.
AI with Ashley explores the current state of artificial intelligence
AI with Ashley is a new recurring conversation series about artificial intelligence and its impact on society. The first edition discusses how AI is already part of everyday life through virtual assistants and self-driving cars. The series will cover advancements, ethics, and the future of work in an AI-driven world.
AI actor Tilly Norwood avoids political topics in interview
Tilly Norwood is a virtual character promoting an upcoming movie called Misaligned. During an interview, the AI actor repeatedly commented on the reporter's clothing instead of discussing politics or the Black Lives Matter movement. The conversation raises questions about whether AI characters can truly engage with important social topics or are programmed to avoid them.
TRM analysis finds most x402 payments are not from AI agents
A new analysis by TRM examined $52.7 million in x402 transactions on the Ethereum blockchain and found most are not from autonomous AI agents. After filtering out self-payments and repeat patterns, likely commerce dropped to $25.62 million. The study warns that regular scripts can mimic agent behavior on the x402 protocol. TRM calls for better agent registration and compliance systems designed for machine-driven payment volumes.
Kate Dohaney says AI calls for a new kind of leadership
Kate Dohaney, CEO of U.K. mobile network Giffgaff, says AI is changing how companies work and plan for the future. She believes leaders should focus on training their teams rather than cutting jobs. Giffgaff is part of the Virgin Media Group and faces new competition from companies like Revolut and Monzo entering the mobile market. Dohaney also highlights that Giffgaff is a certified B-Corp that focuses on low costs and sustainability, with three-quarters of its phones being refurbished.
Sources
- Why a decade of doomsday warnings failed to slow the AI race
- Why this AI doomsday warning from former Anthropic researcher broke through
- AI regulation: Boom Supersonic CEO Blake Scholl says we should build 'safe products'
- Check Point Research: Plain English Can Bypass AI Security Filters
- How to sound like a human, not a robot, in the age of AI
- ASUS at IFA 2026: Advancing AI-Powered Creativity with ProArt and NVIDIA RTX Spark
- 2026 midterm elections: Helping voters navigate election information in the age of AI
- AI with Ashley: Artificial Intelligence Is Already Here
- AI ‘Actor’ Tilly Norwood Told Me That ‘All Lives Matter’
- Agentic Payments Are Growing, but Most x402 Payments Aren’t From AI Agents
- Kate Dohaney says AI demands a different type of leadership—“I’m not a CEO sitting back saying cut, cut, cut.”
Comments
Please log in to post a comment.