Amazon Advances AGI While Anthropic Powers Alexa Plus

Amazon is reorganizing its artificial intelligence leadership as Rohit Prasad, the company's AI chief, prepares to depart by the end of the year. Peter DeSantis, a 27-year Amazon veteran known for launching Amazon EC2 and overseeing chip development, now leads a new Artificial General Intelligence group. This unified team combines Amazon's AGI, custom silicon development, and quantum computing efforts. CEO Andy Jassy initiated this change to strengthen Amazon's AI presence and better compete with rivals like OpenAI and Google. Pieter Abbeel will also head the frontier model research team within this new structure, building on Amazon's recent Nova 2 model launches and its $8 billion investment in Anthropic, whose Claude model powers Alexa Plus. The broader AI industry sees significant capital flow, with over $700 billion invested, yet many executives express concerns about realizing tangible value. A study reveals that numerous AI projects undergo testing but fail to achieve full deployment, often because companies struggle to understand effective implementation strategies. Successful approaches frequently remain private, hindering wider adoption. Furthermore, measuring AI success through employee surveys can be misleading due to potential bias, as true productivity gains stem from consistent tool usage and demonstrable results, not just perceived benefits. AI's influence is expanding across various sectors, from automotive to finance and defense. Qualcomm Executive Vice President Nakul Duggal anticipates robots with reasoning capabilities will appear in cars within five years, as Qualcomm partners with Google for in-car AI assistants and expands its driver-assistance software with BMW. Robinhood is also heavily investing in prediction markets, its fastest-growing revenue source, and introduces Cortex, an AI assistant for Gold subscribers offering trading ideas. Meanwhile, Saab Training and Simulation launched Training Data Insights (TDI), an AI-based tool for analyzing training data and automating after-action reviews using machine learning and generative AI. However, the growth of AI data centers faces a critical challenge: water shortages, which are becoming a more pressing concern than electricity for cooling these intensive operations. As AI capabilities advance, ethical and technical considerations come to the forefront. A UK government survey found 88% of the public demands mandatory copyright licenses for training AI models, with creative industries labeling current practices "industrial-scale theft." In Massachusetts, Hudson High School experienced a lockdown due to a threatening phone call believed to be AI-generated, highlighting potential misuse. While AI models can produce fluent outputs that sound like thinking, they lack true judgment and understanding, according to experts, suggesting AI should augment human decision-making rather than replace it. On the technical front, NVIDIA's CUDA-Q QEC version 0.5.0 introduces real-time decoding and GPU-accelerated decoders, significantly improving quantum error correction for quantum computers.

Key Takeaways

• Amazon's AI chief Rohit Prasad is departing; Peter DeSantis now leads a new Artificial General Intelligence group unifying AI, custom silicon, and quantum computing efforts.
• Amazon has invested $8 billion in Anthropic, integrating its Claude model into Alexa Plus, and launched Nova 2 models.
• Over $700 billion is invested in AI, yet many executives are concerned about realizing tangible value, with projects often tested but not fully deployed.
• A UK government survey found 88% of the public demands mandatory copyright licenses for training AI models, rejecting an "opt-out" approach.
• Qualcomm EVP Nakul Duggal predicts robots with reasoning will be common in cars within five years, with Qualcomm partnering Google for in-car AI assistants.
• An AI-generated threatening phone call caused a lockdown at Hudson High School in Massachusetts, highlighting potential misuse like "swatting."
• Saab Training and Simulation launched Training Data Insights (TDI), an AI-based application for analyzing training data and automating after-action reviews.
• Robinhood introduced Cortex, an AI assistant for Gold subscribers, and notes prediction market trading as its fastest-growing revenue source.
• Water shortages are emerging as a critical challenge for the growth of AI data centers, requiring advanced cooling and sustainable water management strategies.
• NVIDIA CUDA-Q QEC version 0.5.0 significantly improves quantum error correction with new features like real-time decoding and GPU-accelerated algorithmic decoders.

Amazon AI Chief Rohit Prasad Departs Peter DeSantis Takes Over

Amazon's AI chief Rohit Prasad will leave the company at the end of the year. Peter DeSantis, a 27-year Amazon veteran, will now lead a new group combining AI, silicon, and quantum computing. CEO Andy Jassy announced this change to unify Amazon's important AI efforts and compete better with rivals like OpenAI and Google. DeSantis has a strong background, having launched Amazon EC2 and overseen chip development. Amazon recently launched its Nova 2 models and partners with Anthropic to boost its AI presence.

Amazon AI Leader Rohit Prasad Exits Peter DeSantis Steps Up

Amazon's AI chief Rohit Prasad will leave the company by the end of this year. CEO Andy Jassy announced that Peter DeSantis, a long-time Amazon executive, will now lead a new Artificial General Intelligence group. This group will combine Amazon's AGI, custom chip development, and quantum computing teams. The company is making these changes to strengthen its AI efforts and compete more effectively in the fast-growing AI market. Pieter Abbeel will also lead the frontier model research team within this new structure.

Amazon AI Head Rohit Prasad Leaves Peter DeSantis Takes Charge

Amazon CEO Andy Jassy announced that Rohit Prasad, the head scientist for Artificial General Intelligence, will leave Amazon at year's end. Peter DeSantis will now lead a new unified group focused on AI, custom silicon, and quantum computing. Pieter Abbeel, a distinguished scientist, will head the frontier model research team. This leadership change aims to boost Amazon's AI development and better compete with other tech giants. Amazon has also invested $8 billion in AI company Anthropic and uses its Claude model in Alexa Plus.

AI Investment Reaches 700 Billion Dollars Raising Questions

Companies are investing over $700 billion in AI, yet many executives worry about its real impact. A study shows many businesses are testing AI projects but not fully using them. This happens because companies do not yet understand how to use AI successfully, and early wins are often kept secret. Also, measuring AI success through employee surveys can be misleading due to bias. True productivity gains come from consistent use and clear results, not just how people feel about the tools.

Billions in AI Investment Raises Questions on Real Impact

Over $700 billion is being invested in AI, but many companies are concerned about getting real value from it. Executives report that many AI projects are tested but not fully put into use. This is because companies struggle to understand how to use AI effectively, and successful strategies are often kept private. Measuring AI adoption through employee surveys can be misleading, as people may report using tools even when they are not. True benefits come from consistent use of AI tools that lead to clear improvements in work.

UK Public Demands AI Copyright Licenses for Training

A UK government survey found that 88% of people want mandatory copyright licenses for training AI models. This means AI developers would need permission and pay creators to use their work. The creative industries strongly support this, calling the current situation "industrial-scale theft." They rejected the government's preferred "opt-out" option, which only 3% supported. The government will now work on a full report and an economic study to be presented to parliament by March 18, 2026.

Qualcomm Exec Predicts Robots in Cars Within Five Years

Qualcomm Executive Vice President Nakul Duggal believes robots with reasoning will soon be common, starting with cars. Qualcomm is changing its focus to integrate more AI, partnering with Google for in-car AI assistants. The company launched its driver-assistance software with BMW in September and plans to expand to 100 countries by the end of next year. Duggal states that the next five years will bring even faster advancements in this technology. He emphasizes building safe, rules-based systems with AI layered on top.

AI Generated Threat Causes Massachusetts School Lockdown

Hudson High School in Massachusetts experienced a lockdown after receiving a threatening phone call. Police searched the building but found no danger. Initial evidence suggests that artificial intelligence may have generated the call. School officials took extra caution due to recent events. Authorities are now investigating the incident as "swatting," which involves making a false report to cause an emergency response.

Saab Debuts AI Training Analysis Tool TDI

Saab Training and Simulation launched a new AI-based application called Training Data Insights, or TDI. This tool helps analyze training data and provides automated after-action reviews for live events. TDI uses different analytical tools, including machine learning and customized generative AI. It helps experts by providing data to support their observations, making training more measurable and improving its design. Saab announced this new system at the I/ITSEC 25 conference in Orlando.

AI Sounds Like Thinking But Lacks True Understanding

AI models can sound like they are thinking and understanding, but they actually lack true judgment. These systems are good at modeling language patterns and predicting what comes next, leading to fluent outputs. However, this fluency does not mean they have deep understanding or can recognize errors as humans do. Human intelligence develops with consequences and the ability to revise beliefs, which AI systems currently do not possess. While AI is a powerful invention, it should be used to sharpen human judgment, not replace it.

NVIDIA Improves Quantum Error Correction with CUDA-Q QEC

NVIDIA CUDA-Q QEC version 0.5.0 introduces new features to improve quantum error correction. These updates include support for real-time decoding, which is crucial for preventing errors in quantum computers. The new version also brings GPU-accelerated algorithmic decoders and better tools for AI decoder inference. Users can now perform real-time decoding through a four-step process, from generating detector error models to executing quantum circuits. These advancements help researchers accelerate their work and make real-time decoding possible with quantum computers.

Robinhood Expands Prediction Markets Launches AI Assistant

Robinhood is heavily investing in prediction markets and launching a new AI assistant called Cortex. Prediction market trading is already Robinhood's fastest-growing revenue source, with over 1 million customers trading billions of contracts. The new prediction markets hub will offer dollar-based trading on many live events, including sports and politics. Cortex, the AI assistant, will help Gold subscribers with trading ideas and orders, using real-time market data and analyst reports. Robinhood believes these tools will change the future of finance and news.

Water Shortages Threaten AI Data Center Growth

Water is becoming a major problem for the growth of AI data centers, even more than electricity. Data centers use a lot of water for cooling, especially with evaporative systems, which can strain local water supplies. To fix this, experts suggest using advanced cooling technologies that use less water. They also recommend making companies report their water usage clearly and connecting data centers to reclaimed water systems. Pricing water based on its scarcity and using renewable energy sources can also help reduce water consumption.

Sources

• Amazon AI chief Rohit Prasad leaving; Infrastructure exec Peter DeSantis to lead unified AI group
• Amazon says AI chief Rohit Prasad is leaving, Peter DeSantis to lead 'AGI' group
• Amazon CEO Andy Jassy announces departure of AI exec Rohit Prasad in leadership shakeup
• The $700 Billion Question Of AI Investment
• The $700B Question Of AI Investment
• Copyright licenses should be mandatory when training AI models, ...
• 'Robots are going to be amongst us': Qualcomm exec says buckle up for the next 5 years. Your car is going to be the first shoe to drop
• Call that triggered school lockdown may have been made with AI
• Saab launches AI-based training analysis, announces new live-training contract
• How AI Learned to Sound Like Thinking
• Real-Time Decoding, Algorithmic GPU Decoders, and AI Inference Enhancements in NVIDIA CUDA-Q QEC
• Robinhood bets on the future of retail investing: betting
• The Hidden Bottleneck of AI Data Centers: Water