Naveen Rao, the founder of Unconventional AI, is building new computers inspired by the human brain. He argues that current computers are far too inefficient for modern AI tasks, noting that the human brain operates on just 20 watts of power. His company is developing hardware that uses non-linear dynamics instead of traditional math, and they have already created a working prototype in just six months. Rao believes this approach could make AI computing dramatically more efficient.
Meanwhile, Google is expanding its Tensor Processing Unit (TPU) in the United States to reduce reliance on Taiwan Semiconductor Manufacturing Co. However, data center operators are not enthusiastic about the move. They cite NVIDIA's GPU dominance and their long-term partnerships as reasons to stay with the existing ecosystem. Market analysts say the expansion is unlikely to challenge NVIDIA's strong grip on the data center market.
AI is also driving urgent needs in IT asset disposal, according to SK tes. As AI adoption speeds up hardware upgrades, companies are retiring servers and networking gear faster than ever, creating risks for data security and lost value. SK tes, which operates over 40 facilities worldwide, notes that AI data centers now consume huge amounts of memory, and older DDR4 memory is being phased out.
A new report from Cybersecurity Ventures, led by Steve Morgan, finds that many companies want to use AI for security but are not prepared. While 71% of organizations plan to use AI security tools in the next two years, only 22% understand how to use AI to improve security. Companies are also neglecting basic tools like multi-factor authentication. Morgan says companies need a more balanced approach that includes both human and machine defenses.
In healthcare, a new method called atomic fact-checking is helping doctors trust AI recommendations more. This approach breaks down AI treatment suggestions into small claims that can be checked against source guidelines. In a study with 356 clinicians, trust jumped from 26.9% to 66.5% when using atomic fact-checking. The research suggests that making AI outputs easy to verify is key to building trust in healthcare.
AI trickery is creating a growing problem for doctors, according to a new report. Deepfakes are being used to fuel insurance fraud and damage reputations, and the issue could also put patients at risk. The report warns that this problem is getting worse as AI technology improves.
New data from Gallup and the Lumina Foundation shows that more than half of college students use AI for schoolwork. The survey of over 1,000 students found that 52% use AI tools daily or weekly, and 71% believe AI will have a positive impact on their learning. Students who use AI tend to have higher GPAs and are more likely to graduate on time.
A new volume co-edited by Stanford Law professor Nathaniel Persily examines how AI is changing politics. The book brings together over 50 political scientists to discuss topics like democracy, elections, national security, and public opinion. The editors say AI's impact on politics could be much bigger than social media, and they hope the book will spark a society-wide conversation.
Finally, a report from SimSpace shows a gap between confidence in AI security and actual performance. While 78% of security leaders say they are confident in their defenses, real-world readiness scores are as low as 30%. The report found that 73% of organizations already use AI agents in their security operations centers, but only 29% conduct continuous simulation testing. SimSpace says rigorous testing is needed to build trust in AI agents before they go into production.
Key Takeaways
- Naveen Rao's Unconventional AI is building brain-inspired hardware using non-linear dynamics, with a working prototype completed in six months.
- Google is expanding its TPU in the US to reduce reliance on Taiwan Semiconductor, but data center operators remain committed to NVIDIA's GPU ecosystem.
- SK tes reports that AI adoption is accelerating hardware upgrades, creating urgent needs for secure IT asset disposal and data protection.
- Cybersecurity Ventures CEO Steve Morgan says 71% of organizations plan to use AI security tools, but only 22% understand how to use them effectively.
- Atomic fact-checking boosts doctor trust in AI recommendations from 26.9% to 66.5% by breaking down claims for verification.
- AI deepfakes are fueling insurance fraud and reputation damage in healthcare, with risks to patient safety.
- 52% of college students use AI tools daily or weekly, and 71% believe AI will positively impact their learning, according to a Gallup and Lumina Foundation survey.
- A new book co-edited by Stanford's Nathaniel Persily explores AI's potential to reshape politics, with over 50 political scientists contributing.
- SimSpace reports that 78% of security leaders are confident in their AI defenses, but real-world readiness scores are as low as 30%.
- 73% of organizations already use AI agents in security operations centers, but only 29% conduct continuous simulation testing.
Naveen Rao builds new AI computers inspired by the brain
Naveen Rao, founder of Unconventional AI, wants to rebuild computers for the AI age. He says current computers are too inefficient for modern AI tasks. Rao looks to biology for inspiration, noting the human brain uses only 20 watts of power. His company is building new hardware that uses non-linear dynamics instead of traditional math. They have already created a working prototype in just six months. Rao believes this approach could make AI computing far more efficient.
SK tes says AI is driving urgent need for secure IT asset disposal
SK tes highlights the growing need for secure and sustainable IT asset disposition as AI adoption speeds up hardware upgrades. Companies are retiring servers and networking gear faster than ever, creating risks for data security and lost value. SK tes offers global decommissioning services with over 40 facilities worldwide. The company notes that AI data centers now consume huge amounts of memory, and older DDR4 memory is being phased out. SK tes helps organizations manage complexity, protect data, and recover value from old equipment.
Doctors face growing problem with AI deepfakes
AI trickery is creating a growing problem for doctors, according to a new report. Deepfakes are being used to fuel insurance fraud and damage reputations. The issue could also put patients at risk. The article warns that this problem is getting worse as AI technology improves.
Businesses want AI security but are not ready for it
A new report from Cybersecurity Ventures finds that many companies want to use AI for security but are not prepared. The report says 71% of organizations plan to use AI security tools in the next two years. However, only 22% understand how to use AI to improve security. Companies are also neglecting basic tools like multi-factor authentication. Steve Morgan, CEO of Cybersecurity Ventures, says companies need a more balanced approach that includes both human and machine defenses.
Atomic fact-checking boosts trust in AI for doctors
A new method called atomic fact-checking helps doctors trust AI recommendations more. This approach breaks down AI treatment suggestions into small claims that can be checked against source guidelines. In a study with 356 clinicians, trust jumped from 26.9% to 66.5% when using atomic fact-checking. Traditional methods only showed small improvements. The research suggests that making AI outputs easy to verify is key to building trust in healthcare.
New book explores how AI will reshape politics
A new volume co-edited by Stanford Law professor Nathaniel Persily examines how AI is changing politics. The book brings together over 50 political scientists to discuss topics like democracy, elections, national security, and public opinion. It is a follow-up to an earlier book on social media and democracy. The editors say AI's impact on politics could be much bigger than social media. They hope the book will spark a society-wide conversation about AI and politics.
AI shapes learning for Wisconsin class of 2026 graduates
New data from Gallup and the Lumina Foundation shows that more than half of college students use AI for schoolwork. The survey of over 1,000 students found that 52% use AI tools daily or weekly. Students use AI for writing papers, creating presentations, and solving math problems. The report also found that 71% of students believe AI will have a positive impact on their learning. Students who use AI tend to have higher GPAs and are more likely to graduate on time.
Google TPU expansion gets cool reception from US data centers
Google is expanding its Tensor Processing Unit (TPU) in the United States to reduce reliance on Taiwan Semiconductor Manufacturing Co. However, data center operators are not excited about the move. They say NVIDIA's GPU dominance and long-term partnerships keep them committed to the existing ecosystem. Operators are hesitant to switch to new technology because of their investment in current infrastructure. Market analysts say the expansion is unlikely to challenge NVIDIA's strong grip on the data center market.
Security leaders overconfident in AI defenses, SimSpace report finds
A new report from SimSpace shows a gap between confidence in AI security and actual performance. While 78% of security leaders say they are confident in their defenses, real-world readiness scores are as low as 30%. The report found that 73% of organizations already use AI agents in their security operations centers. However, only 29% conduct continuous simulation testing. The report warns that teams are deploying AI without fully testing it in realistic conditions. SimSpace says rigorous testing is needed to build trust in AI agents before they go into production.
Sources
- Naveen Rao on Rebuilding Computers for the AI Age
- SK tes Highlights the Rising Need for Secure, Sustainable, and Scalable IT Asset Disposition in the AI Era
- Doctors' growing AI deepfakes problem
- Businesses eager but unprepared for AI to transform their security strategies
- Atomic Fact-Checking Boosts AI Clinical Trust
- How Will AI Reshape Politics? New Volume Co-Edited by Stanford Law’s Nathaniel Persily Explores the Stakes
- How AI shaped learning for Wisconsin's class of 2026 graduates
- Google's TPU Expansion Faces Cool Reception from U.S. Data Centers
- SimSpace Report: Nearly 80% of Security Leaders Report High Confidence in Their AI Defenses, but Real World Defensive Security Readiness Scores Are As Low As 30%
Comments
Please log in to post a comment.