🗞 OpenAI Approaching Human-Level AI Reasoning, Company Claims 🧠

🔎 The Latest on the AI Frontier:

• OpenAI Approaching Human-Level AI Reasoning, Company Claims 🧠

• Sam Altman's Nuclear Startup Aims to Power AI with Small Reactors ⚛️

• OpenAI CEO Calls for Unity Following Trump Shooting Incident 💬

• Microsoft's Copilot AI Tackles Handwritten Notes in OneNote ✍️

• Privacy Concerns Halt Meta's AI Expansion Plans in Europe 🛑

• Samsung and Apple Compete in AI Features Race 📱

• AI Weapons Spark Ethical Debates and Arms Race Concerns 🚀

• AI Liability Insurance Emerges as New Growth Area for Insurers 💼

• Nvidia's Next-Gen GPUs May See Significant Power Consumption Increase 🖥️

🧠 OpenAI claims it's nearing AI with human-level "reasoning" capabilities. Link

• The company has defined 5 levels of artificial general intelligence (AGI), with "reasoners" at level 2 capable of human-level problem-solving.

• OpenAI leaders told staff they are "on the cusp" of achieving level 2, moving beyond current chatbot capabilities.

• Experts remain divided on whether large language models can truly achieve human-like reasoning and understanding.

⚛️ OpenAI founder Sam Altman's nuclear startup aims to power AI with small modular reactors by 2027. Link

• Oklo, chaired by Altman, is working to build its first reactor at Idaho National Laboratory.

• The company's compact "Aurora powerhouse" reactors could provide sustainable, reliable power for data centers and AI companies.

• Oklo plans to offer long-term energy agreements rather than licensing, potentially opening up partnerships with AI firms like OpenAI.

💬 OpenAI CEO Sam Altman responds to Trump shooting, calls for unity and reduced political rhetoric. Link

• Altman expressed relief at Trump's safety and reflected on how close the incident came to changing history.

• He praised Democrats for "having the grace to step up and lead" in response to the attack.

• Altman, a consistent Democratic donor, urged for combating extreme political polarization and finding "slightly more unity" in the wake of the incident

✍️ Microsoft's Copilot AI can now understand and analyze handwritten notes in OneNote. Link

• The feature can summarize handwritten notes, generate to-do lists, and convert handwriting to text.

• Early testing shows impressive accuracy in deciphering and reformatting handwritten content.

• Available to Copilot for Microsoft 365 subscribers and Copilot Pro users, with beta testing already underway.

🛑 Meta's AI expansion in Europe halted due to privacy concerns. Link

• Irish Data Protection Commission (DPC) paused Meta's plans to train AI models on European user data.

• Key issues include lack of explicit consent, overly broad data collection, and transparency problems.

• The pause impacts Meta's AI development, metaverse plans, and competitive edge in large language models.

📱 Samsung's Galaxy AI vs Apple Intelligence: Early comparisons show a tight race. Link

• Samsung takes an early lead with features like Interpreter for real-time translation and Circle to Search.

• Apple excels in photo editing and creative tools like Genmoji and Image Playground.

• Both offer AI-powered writing assistance and image generation, with slight differences in implementation.

• The competition remains close, with Apple's full AI rollout still to come in iOS 18 later this year.

🚀 AI-enabled autonomous weapons are rapidly entering battlefields, sparking a multibillion-dollar arms race and ethical concerns. Link

• The US military has over 800 active AI-related projects and requested $1.8bn for AI in the 2024 budget alone.

• Major tech companies like Google and Amazon are increasingly embracing defense contracts despite employee protests.

• Experts warn of a "double black box" effect, where secretive AI technologies within classified military operations lack transparency and accountability.

💼 AI liability insurance emerges as a potential growth area for insurers, with unique challenges and opportunities. Link

• Deloitte projects $4.7 billion in AI insurance premiums over 10 years, though this represents only about 0.012% of current global non-life insurance premiums.

• Setting appropriate premiums is challenging due to limited historical data on AI-related damages and evolving court rulings on liability issues.

• As AI becomes more ubiquitous, insurance will play a crucial role in managing risks and shaping the future of AI adoption across industries.

🖥️ Nvidia's RTX 50 series GPU TDPs potentially revealed by Seasonic's PSU calculator, suggesting significant power increases. Link

• The RTX 5090 may require up to 500W, a 50W increase from the RTX 4090's 450W TDP.

• Largest generational TDP increase appears to be in the '60 series, jumping from 115W to 170W.

• If accurate, these figures indicate Nvidia is pushing for higher performance at the cost of increased power consumption across the lineup.

Have any feedback? Send us an email