🗞 OpenAI's Ex-Chief Scientist Secures Billion-Dollar Funding for AI Safety Venture 🚀

🔎 The Latest on the AI Frontier:

• OpenAI's Ex-Chief Scientist Secures Billion-Dollar Funding for AI Safety Venture 🚀

• Elon Musk's xAI Unveils Massive GPU-Powered Supercomputer 🖥️

• OpenAI CEO Plans Ambitious Global AI Infrastructure Investment 🌐

• Meta Releases New Benchmark to Assess AI Cybersecurity Risks 🛡️

• Microsoft Rebrands and Enhances Copilot for Office Integration 📊

• Google Embeds Gemini AI Directly into Chrome's Address Bar 🔍

• Google Introduces 'Gems' to Simplify AI Prompt Engineering 💎

• Amazon's Just Walk Out Tech Expands to More Stadiums and Campuses 🏟️

• Paytm CEO Sparks Debate on ChatGPT vs. Perplexity AI Capabilities 🗣️

• AI Breakthrough Could Solve Antibiotic Resistance Crisis 🧬

• Nvidia Shares Tumble Amid Escalating Antitrust Investigation 📉

• NVIDIA's Next-Gen GPUs Rumored to Consume Significantly More Power ⚡

💰 OpenAI's former chief scientist raises $1 billion for new AI safety startup. Link

• Safe Superintelligence (SSI), co-founded by Ilya Sutskever, has secured $1 billion in funding to develop safe AI systems that surpass human capabilities.

• The startup, valued at $5 billion, plans to use the funds for computing power and to hire top talent, focusing on building a small, trusted team of researchers and engineers in Palo Alto and Tel Aviv.

• This significant investment highlights continued investor interest in exceptional talent focused on foundational AI research, despite a general waning of interest in funding such potentially unprofitable ventures.

🖥️ Elon Musk's xAI launches massive AI supercomputer with 100,000 Nvidia GPUs. Link

• The "Colossus" supercomputer, located in Memphis, Tennessee, was assembled in just 122 days and is now online.

• It uses 100,000 Nvidia H100 GPUs, likely costing around $3 billion, with plans to expand to 200,000 GPUs in the coming months.

• The facility aims to accelerate AI training for xAI's projects, including the chatbot Grok, though its claim as the "most powerful AI training system in the world" is debatable given similar large-scale efforts by other tech giants.

🌐 OpenAI's Sam Altman plans massive AI infrastructure investment, starting in the U.S. Link

• The initiative aims to build global AI infrastructure, potentially costing tens of billions of dollars, with initial focus on U.S. states.

• Projects include setting up data centers, expanding energy capacity, and increasing semiconductor manufacturing, with potential investors from Canada, Korea, Japan, UAE, and private companies like Microsoft.

• The plan could face U.S. national security scrutiny, particularly regarding foreign investments, but OpenAI argues it would boost America's geopolitical advantage in AI competition with China.

🛡️ Meta releases CyberSecEval 3 to assess AI cybersecurity risks and capabilities. Link

• New benchmark suite evaluates 8 different risks across third-party and developer/user categories for LLMs.

• Llama 3 405B model demonstrated ability to automate "moderately persuasive multi-turn spear-phishing attacks" similar to GPT-4 Turbo.

• Top strategies include deploying AI guardrails, enhancing human oversight, strengthening phishing defenses, continuous security training, and multi-layered approach.

🔍 Microsoft is gearing up for the "next phase" of Copilot innovation. Link

• Say goodbye to "Copilot in Word" and hello to "Microsoft 365 Copilot in Word." Microsoft is set to rebrand its AI assistant offerings to emphasize integration with its core products.

• New Features on the Horizon: the event promises to unveil exciting new features aimed at enticing businesses to opt into their $30 per user monthly plan—proving the value of AI integration in Office apps.

• With potential updates to the consumer-focused Copilot Pro, Microsoft continues to refine its AI offerings, making them more appealing to both businesses and individual users.

🌐 Google integrates Gemini AI directly into Chrome’s address bar. Link

• Seamless AI integration allows Chrome users to access Gemini by typing “@gemini” in the address bar, making AI a standard feature for millions of users globally.

• This development enhances AI accessibility, potentially reshaping user interactions and accelerating the adoption of AI tools in everyday online activities.

• Google's move is part of its broader AI-first strategy, which could have significant implications for data privacy, digital marketing, and how businesses leverage AI technology.

🧪 Google's 'Gems' feature offers a user-friendly introduction to AI prompt engineering. Link

• Gems are pre-built, customizable chat sessions in Google's Gemini AI that help users with tasks like brainstorming strategies, improving study habits, and refining writing.

• Available to Gemini Advanced subscribers ($19.99/month), Gems provide a template for prompt engineering that users can modify, offering insights into crafting effective AI interactions.

• While valuable for beginners, Gems have limitations such as lack of memory across sessions and inability to connect later conversation turns with earlier ones, indicating areas for potential improvement in AI model development.

🏈 Amazon's Just Walk Out tech expands to more NFL stadiums and universities this fall. Link

• Six new stores at Lumen Field (Seattle), seven at Commanders Field (Washington D.C.), and one at M&T Bank Stadium (Baltimore).

• Over 30 university stores now powered by Just Walk Out technology worldwide.

• Lumen Field saw 60% increase in customer throughput and doubled transactions per game with first Just Walk Out store in 2022 season.

🤖 Paytm CEO compares ChatGPT to "talking Taylor Swift to a grandfather" in AI debate. Link

• Vijay Shekhar Sharma weighs in on ChatGPT vs Perplexity AI comparison, highlighting limitations of ChatGPT for recent events.

• Perplexity AI CEO Aravind Srinivas sparked debate by showcasing more precise, up-to-date responses from his platform.

• Sharma notes ChatGPT's data mining limits, while praising Perplexity as "perpetually intelligent" for current information.

🧬 AI revolutionizes drug discovery, potentially solving antibiotic resistance for $20 billion. Link

• Scientists use AI to screen millions of molecules, identifying novel antibiotics like Halicin, which kills drug-resistant bacteria without developing resistance after 30 days of exposure.

• AI-powered models can now generate entirely new drug candidates, with one system creating 30 billion theoretical molecules and successfully synthesizing 58 novel compounds, 6 of which proved lethal against MRSA.

• While AI accelerates early-stage drug discovery, experts caution that clinical trials remain essential, as AI-discovered molecules still fail in later stages at similar rates to traditionally discovered drugs.

🏦 Nvidia shares slump amid reports of escalating US antitrust investigation. Link

• Nvidia's stock fell 2.4% in after-hours trading following a 10% drop during regular trading that wiped $279 billion off its market value, marking the largest one-day drop for a US company.

• The US Department of Justice reportedly sent subpoenas to Nvidia and other tech companies, signaling an escalation of its antitrust probe into whether Nvidia has made it difficult for clients to switch chip suppliers.

• The sell-off was part of a broader market decline sparked by weak US manufacturing data, with the S&P 500 falling over 2% and the Nasdaq Composite dropping nearly 3.3%.

🖥️ NVIDIA's next-gen GPUs reportedly to consume significantly more power than current models. Link

• The GeForce RTX 5090 is rumored to have a 600W TDP, a 33% increase over the RTX 4090's 450W rating.

• The RTX 5080 is expected to be rated at 400W, 25% higher than the RTX 4080's 320W TDP.

• While TDP ratings don't necessarily reflect real-world power consumption during gaming, these increases suggest NVIDIA's next-gen GPUs may require more robust cooling and power supply solutions.

Have any feedback? Send us an email