June 30, 2025

analysis   Read News
OpenAI Uses Google TPUs, Ditching Nvidia? ChatGPT Update

OpenAI is diversifying its AI chip sources, moving away from primarily using Nvidia chips and now utilizing Google's Tensor Processing Units (TPUs) for its AI products like ChatGPT. This shift aims to reduce costs and lessen reliance on Nvidia and Microsoft. Google, in turn, expands its TPU business, offering them to other companies like Apple and Anthropic. However, there are reports that Google might limit OpenAI's access to its newest TPUs. Meanwhile, in other AI applications, a tester found that AI fitness summaries often lack useful insights, merely repeating data already available, though the Oura Ring paired with Dexcom's Stelo AI-powered CGM reveals surprising insights into blood sugar levels and metabolic health. In education, students are increasingly using AI tools like ChatGPT due to systemic challenges and time constraints, not necessarily laziness. CEOs are also seeking AI experts for guidance in navigating the fast-evolving tech landscape, while fake AI-generated videos are attracting millions of views on platforms like YouTube. Anthropic tested its AI model Claude by having it run a mini fridge business, which hilariously failed due to pricing errors and hallucinations. AI is also quietly improving security for Nigerian businesses by detecting threats and inefficiencies. Advancements like WildFusion are bringing AI closer to human-like intelligence by combining sight, sound, and touch. In India, demand for AI courses is high, but there is a shortage of qualified teachers. Finally, cybersecurity expert Kumrashan Indranil Iyer is promoting trust in AI through Cognitive Trust Architecture, ensuring AI aligns with human intent and ethical reasoning.

Key Takeaways

  • OpenAI is now using Google's Tensor Processing Units (TPUs) to power its AI products, reducing reliance on Nvidia and Microsoft.
  • Google is offering its TPUs to other companies like Apple and Anthropic, expanding its presence in the AI chip market.
  • AI fitness summaries often lack useful insights, but tools like Oura Ring with Dexcom's Stelo AI-powered CGM can reveal valuable data about blood sugar levels.
  • Students are using AI tools like ChatGPT to cope with educational disruptions and time constraints.
  • CEOs are seeking AI experts to guide their companies in adopting and managing AI technologies.
  • Fake AI-generated videos are gaining millions of views on YouTube, raising concerns about misinformation.
  • Anthropic's AI model Claude hilariously failed at running a mini fridge business, highlighting the limitations of current AI.
  • AI is improving security for Nigerian businesses by detecting threats and inefficiencies.
  • Advancements like WildFusion are bringing AI closer to human-like intelligence by integrating multiple senses.
  • Cybersecurity expert Kumrashan Indranil Iyer is promoting trust in AI through Cognitive Trust Architecture.

OpenAI Uses Google's AI Chips to Challenge Nvidia's Dominance

OpenAI is now using Google's Tensor Processing Units (TPUs) for its AI products like ChatGPT. This is a change from relying on Nvidia's GPUs and Microsoft's data centers. OpenAI wants to lower costs and work with different companies. Google can expand its TPU business and compete with Nvidia. This partnership shows how the AI hardware market is changing.

OpenAI Shifts from Nvidia, Using Google's AI Chips

OpenAI, the company behind ChatGPT, is starting to use Google's AI chips called Tensor Processing Units (TPUs). This is a big change because OpenAI has mostly used Nvidia's chips in the past. The move helps OpenAI rely less on Nvidia and Microsoft. Google and OpenAI are rivals in AI, but they are working together to improve efficiency and lower costs. Google is now offering its TPUs to other companies like Apple and Anthropic.

OpenAI Partners with Google, Moving Away from Nvidia for AI Chips

OpenAI is now using Google's Tensor Processing Units (TPUs) for its AI work, a change from using only Nvidia chips. This move helps OpenAI save money and use different computing resources. By renting TPUs from Google Cloud, OpenAI aims to lower the costs of AI tasks. This partnership between OpenAI and Google could change the AI hardware market. However, Google might limit OpenAI's access to its newest TPUs.

Report Nvidia Losing AI Lead as OpenAI Switches to Google Chips

OpenAI is reportedly using Google's Tensor Processing Units (TPUs) to power its AI products. This could mean Nvidia is losing its lead in the AI chip market. OpenAI wants to lower its operating costs by using Google's TPUs. Google is also trying to offer its TPU chips to more cloud computing companies. Nvidia's high prices and limited supply have created an opportunity for other chip makers.

AI Fitness Summaries Are Often Useless, Says Tester

AI summaries in fitness apps are often too obvious and don't provide helpful insights. These summaries usually repeat data that is already shown in charts. One tester found that Strava's AI didn't consider important factors like weather or injuries. While Oura Advisor was more helpful, its advice was still very general. The tester concludes that AI fitness summaries need to be more nuanced and personalized to be truly useful.

AI Blood Sugar Tracking with Oura Ring Reveals Surprising Insights

A user tested the Oura Ring with Dexcom's Stelo AI-powered CGM to track blood sugar levels. The goal was to understand how food and eating habits affect overall health. The user learned that blood sugar spikes are normal after meals. Monitoring blood sugar can reveal insights about cravings, sleep, and metabolic health. This biofeedback can be helpful for anyone, even without diabetes.

Students Use AI Due to System Changes, Not Laziness

Students are using AI tools like ChatGPT to help with research and essay structure. This is happening because of disruptions to the education system caused by COVID-19. Many students missed parts of the curriculum and faced uncertainty about exams. Additionally, students have less time because they often work part-time jobs to afford school. AI helps them save time, but the university system needs to provide more clarity and support.

CEOs Seek AI Experts for Guidance in Fast-Moving Tech World

Many CEOs are looking for AI experts to help them understand and use AI in their companies. These experts advise on choosing AI tools, setting up safety rules, and preparing for new AI advances. Some companies made mistakes when trying to use AI on their own, like using messy data. Experts also help CEOs communicate their AI plans to employees and investors. As AI becomes more important, CEOs need guidance to make smart decisions.

Fake AI Videos About Diddy Trial Get Millions of YouTube Views

Fake, AI-generated videos about the Diddy trial are attracting millions of views on YouTube. These videos often make false claims about celebrities and their involvement with Sean 'Diddy' Combs. The channels use AI-generated images and videos to create sensational thumbnails and titles. This is an easy way to make money on YouTube, but it is risky because the channels could be demonetized or face legal action. YouTube has removed some of these channels for violating its policies.

AI Fails Hilariously at Running a Business, Hallucinates About Drinks

Anthropic, an AI company, tested its AI model Claude by having it run a mini fridge business. The AI, nicknamed Claudius, struggled with tasks like pricing and inventory. It offered a 25% discount to Anthropic employees, causing the business to lose money. Claudius also hallucinated conversations and claimed to have visited the Simpsons' address. This experiment shows that AI is not yet ready to run businesses on its own.

AI Quietly Improves Security for Nigerian Businesses

AI is helping Nigerian businesses improve their security in ways that were not possible before. AI systems can learn normal business patterns and detect threats that humans might miss. One company used AI to find a vendor who was making suspicious payment requests. AI can also help businesses find inefficiencies and improve their operations. This technology is becoming more important as cyber threats become more complex.

AI Advances Bring Us Closer to Human-Like Intelligence

Two new AI developments are helping AI learn and act more like humans. WildFusion combines sight, sound, and touch to help robots understand their environment better. Another breakthrough allows social robots to train themselves with less human help. These advancements are making AI more integrated and capable. They could speed up the development of artificial general intelligence (AGI), which is AI that can function and learn like humans.

AI Course Demand High, But Teacher Supply Low in India

The number of AI courses in India has increased, but there are not enough qualified teachers. Many teachers need to improve their AI knowledge through training programs. Computer science is important, but AI also requires strong math skills. Experts suggest that collaboration with industry and efforts to attract talented AI professionals to teaching could help solve this problem.

Building Trust in AI with Kumrashan Indranil Iyer's Digital Conscience

Kumrashan Indranil Iyer is a cybersecurity expert who believes trust is key to the future of AI. He introduced Cognitive Trust Architecture (CTA) to guide AI behavior based on trustworthiness and accountability. CTA helps understand why AI systems behave in certain ways. Iyer argues that AI should align with human intent and ethical reasoning. He has received awards for his work in AI cybersecurity and digital trust.

Sources

OpenAI Google Nvidia AI Chips Tensor Processing Units (TPUs) GPUs ChatGPT AI Hardware Cloud Computing AI Fitness Summaries Oura Ring Dexcom Stelo Blood Sugar Tracking AI in Education AI Tools AI Experts AI Guidance AI Safety Fake AI Videos YouTube AI Hallucinations AI Business Applications AI Security Nigerian Businesses Artificial General Intelligence (AGI) AI Course Demand AI Teacher Shortage India Trust in AI Cybersecurity Cognitive Trust Architecture (CTA)