Google and AMD Partner on Next-Gen AI Chip Design

Andrej Karpathy, a prominent AI expert, has shifted his focus from 'vibe coding' to 'agentic engineering', a harder discipline that involves writing detailed specs, reviewing agent-generated code, and designing automated evaluation loops. This change reflects the evolving landscape of AI-assisted development, with a growing emphasis on professional oversight and rigorous testing.

Dario Amodei, CEO of Anthropic, has addressed the growing backlash against AI, describing it as 'fundamentally a crisis of trust'. He emphasizes that AI is not a monolithic entity, but rather a collection of tools and techniques that can be used for various applications.

Google and AMD are partnering on a next-generation Tensor Processing Unit (TPU) design, which is expected to integrate on-package CPU cores for reinforcement learning and agentic workloads. This partnership could lead to significant advancements in AI computing and performance.

Lumistar Carry, an AI-powered quad camera basketball training partner, is now live on Kickstarter. The device provides instant feedback and analysis, helping users refine their technique and gain a competitive edge.

Researchers are seeking a better way to measure whether AI use supplements or replaces human interaction. AI agents can now mimic human web browsing using Core Data Platform (CDP), and are being explored for various applications.

The Palm Beach Police Department is using AI to improve efficiency and effectiveness in law enforcement, with AI-powered tools enhancing communication and response times.

Key Takeaways

* Andrej Karpathy shifts focus from 'vibe coding' to 'agentic engineering', emphasizing professional oversight and rigorous testing in AI-assisted development. * Dario Amodei, CEO of Anthropic, addresses AI backlash as a 'crisis of trust' and highlights the need for nuanced discussion. * Google and AMD partner on next-gen AI chip design, integrating on-package CPU cores for reinforcement learning and agentic workloads. * Lumistar Carry, an AI-powered basketball training partner, is live on Kickstarter. * Researchers seek better metrics to measure AI's impact on human interaction. * AI agents can mimic human web browsing using Core Data Platform (CDP). * AI struggles with original scientific research. * Palm Beach Police Department leverages AI for enhanced efficiency in law enforcement. * Anthropic is working on developing AI responsibly. * AI-powered tools are being used for various applications, including language translation and search functionality.

Andrej Karpathy Shifts Focus from Vibe Coding to Agentic Engineering

Andrej Karpathy, a prominent AI expert, has announced his shift from 'vibe coding' to 'agentic engineering'. In February 2025, Karpathy coined the term 'vibe coding' to describe casual, low-scrutiny AI-assisted programming. However, by April 2026, he declared vibe coding passe and introduced agentic engineering as a harder discipline. This change reflects the evolving landscape of AI-assisted development, with a growing emphasis on professional oversight and rigorous testing. Karpathy's new approach involves writing detailed specs, reviewing agent-generated code, and designing automated evaluation loops.

Lumistar Carry: AI-Powered Basketball Training Partner

Lumistar Carry is an AI-powered quad camera basketball training partner designed for serious hoopers and coaches. It is now live on Kickstarter, offering a new way for players to improve their skills with the help of artificial intelligence. The device provides instant feedback and analysis, helping users to refine their technique and gain a competitive edge.

Anthropic CEO Addresses AI Backlash and Trust

Dario Amodei, CEO of Anthropic, has spoken out against the growing backlash against AI, describing it as 'fundamentally a crisis of trust'. He emphasizes that AI is not a monolithic entity, but rather a collection of tools and techniques that can be used for various applications. Amodei highlights the need for a more nuanced discussion about AI and its limitations, as well as the importance of developing AI responsibly.

Measuring AI's Impact on Human Interaction

Researchers are seeking a better way to measure whether AI use supplements or replaces human interaction. Current metrics may not accurately capture the impact of AI on human behavior, and a new approach is needed to understand the effects of AI on society. The researchers are exploring new methods to assess the level of engagement with AI systems and human interaction.

AI Agents Mimic Human Web Browsing with CDP

AI agents can now mimic human web browsing using Core Data Platform (CDP). This technology allows AI agents to learn human-like browsing patterns, potentially overcoming website defenses. The 'sense, act, verify' loop and 'meatbag ladder' approach enable AI agents to navigate the web more effectively.

AI Struggles with Original Scientific Research

AI agents are struggling to perform original scientific research, according to recent findings. Despite advancements in AI capabilities, the technology still has limitations when it comes to conducting innovative research. The challenges of AI in scientific research are being addressed by ongoing studies and developments.

Palm Beach Police Leverage AI for Enhanced Efficiency

The Palm Beach Police Department is using AI to improve efficiency and effectiveness in law enforcement. AI-powered tools, such as language translation and search functionality, are being used to enhance communication and response times. The department's Real Time Crime Center is also utilizing AI to identify suspects and track crimes.