AI NEWS YOU MISSED ❗#50
Robert Maciejko
Global Business Strategist & Board Member | Chairman & CEO | Co-founder, INSEAD AI | Publisher of 'AI News You Missed
Top 5 stories of the week
1️⃣ 12 Days of OpenAI Starts With o1 Model
2️⃣ Amazon Nova Models Unveiled
3️⃣ Anduril & OpenAI partner on U.S. AI defense
4️⃣ AI’s Next Frontier: Science and Innovation
5️⃣ Alibaba QwQ: China’s New AI Powerhouse
Details for Top 5 stories of the week
1️⃣ 12 Days of OpenAI Starts With o1 Model
OpenAI launched the “12 Days of OpenAI,” introducing the full 01 model with enhanced speed and multimodal capabilities, along with a $200/month ChatGPT Pro tier featuring exclusive 01 Pro mode for advanced tasks, GPT-4 Turbo, and Advanced Voice mode.
2️⃣ Amazon Nova Models Unveiled
READ MORE: Reuters TechCrunch Philipp Schmid tweet
3️⃣ Anduril & OpenAI partner on U.S. AI defense
Anduril and OpenAI partner to deploy advanced AI for U.S. defense, enhancing counter-drone systems and ensuring AI-driven national security.
4️⃣ AI’s Next Frontier: Science and Innovation
As AI models near peak performance for general queries, the focus is shifting toward leveraging AI to accelerate science and engineering, where vast potential remains to drive technological progress and address complex, open-ended challenges.
5️⃣ Alibaba QwQ: China’s New AI Powerhouse
China’s Alibaba unveils QwQ-32B-Preview, a 32.5 billion parameter AI model surpassing OpenAI’s o1 in specific benchmarks, signaling its growing dominance in advanced AI reasoning.
Do Chinese models have “Meta Inside”? — Anthropic co-Founder Jack Clark
Chinese labs like Alibaba’s Qwen and DeepSeek are setting new benchmarks in AI with models rivaling OpenAI, while innovations like F5-TTS and CogVideo-X solidify China’s dominance in open-source AI and other advanced domains.
In just two decades, China has surged from leading in 3 critical technologies to dominating 57 out of 64, with 24 now classified as high-risk for potential Chinese monopoly, signaling a dramatic shift in global tech leadership.
AI Guardrails in Beijing: A User’s Experience
During a visit to Beijing, a user found Kimi’s AI restricted in translating “Tian’anmen” into English, illustrating the limitations of American-style tools in China, where only Apple services functioned reliably.
Upcoming live events
Revolutionizing Customer Experience & Manufacturing with AI: Insights from the Automotive Industry — Online Dec. 10, 2024
Join INSEAD’s Joerg Niessing and industry leaders on December 10 to explore how AI is transforming customer experience, manufacturing, and enterprise solutions in the automotive sector through real-world use cases and strategies for successful adoption.
New Crypto Regulations: Building a Better Financial System — WEDNESDAY, DECEMBER 11, 2024 — ONLINE — FT
Join CSET experts on December 11 for a webinar exploring AI governance in 2025, featuring insights from leading experts and the launch of AGORA, a new resource for tracking global AI laws and regulations.
Designing Human-Centric Organizations in the Algorithmic Age — Dec. 12, 2024
Join INSEAD Professor Phanish Puranam on December 12 to explore strategies for designing human-centric organizations in the age of AI, focusing on balancing technological efficiency with employee well-being, motivation, and meaningful engagement.
More AI stories of the week
Suggested by INSEAD-ers. From the 1,500+ person global INSEAD AI private business leader community, regularly sharing AI-related articles, vids, etc, of interest. (alum-led)
A thought-provoking INSEAD session with alum Tim Gordon explored AI’s transformative impact on corporate systems, global services, information economics, and politics, highlighting both opportunities and challenges like misinformation and economic shifts.
Anthropic secured $4B from Amazon, Chinese labs advanced reasoning models rivaling U.S. performance, TÜLU 3 outperformed in AI safety and reasoning, AI benchmarks highlighted human-AI trade-offs, and OpenAI faced controversy over unpaid artist contributions.
Since ChatGPT’s launch, 6 leading AI stocks have added $8.7 trillion in value — outpacing the combined market cap of major exchanges like Euronext, Tokyo, and Shanghai — led by NVIDIA’s $3 trillion surge.
INSEAD MBA students curated recent AI advancements across healthcare, robotics, Formula One, and business innovation, showcasing how AI reshapes industries, with summaries and podcasts generated via Google’s notebooklm.
AI is transitioning from assistance to autonomy, forming self-managing teams and reshaping workflows, trust, and oversight across industries.
US Leads in AI Equity Funding
Nvidia Faces Growing Competition in AI Chip Market: Rivals like AMD, Amazon, and start-ups are introducing faster, cost-effective alternatives to Nvidia’s AI chips, particularly for inferencing tasks, signaling a shift in the competitive landscape.
Meta is seeking partners to develop 1–4 GW of nuclear energy projects by the 2030s, aiming to power its AI data centers with zero-carbon energy while accelerating nuclear deployment and scalability.
Fei-Fei Li highlights World Labs’ groundbreaking AI system that generates interactive 3D scenes from a single image or sentence, allowing users to experience these immersive creations directly in their browser.
Google DeepMind introduces Genie 2, a state-of-the-art foundation world model designed to generate diverse, consistent virtual environments for up to one minute, potentially sparking a new era of advancements for embodied AI agents.
Tencent’s open-source Hunyuan Video model, with 13 billion parameters, revolutionizes video creation through advanced features like text-to-video and image-to-video transformation, outperforming competitors and making high-quality video generation more accessible.
EuroLLM-9B, developed using EuroHPC infrastructure, is the most advanced European language model of its size, excelling in multilingual and translation tasks across 24 EU languages and beyond, setting a new benchmark for open-source AI innovation in Europe.
UBS’s “Year Ahead 2025” report identifies AI as a $1.2 trillion annual opportunity by 2027, driven by infrastructure, diversified tech investments, and AI-powered applications reshaping industries.
Hard-coded filters in ChatGPT, introduced to address defamation concerns, can block conversations involving certain names, raising questions about usability, adversarial vulnerabilities, and fairness in AI systems.
OpenAI explores ads in ChatGPT, reflecting the financial pressures of running costly AI systems for free, while raising concerns about trust and potential conflicts of interest in user-focused AI experiences.
Explores AI’s transformative impact on industries by re-engineering processes, driving innovation, and creating new growth opportunities, with a focus on infrastructure, companies like NVIDIA, and practical applications in mid-market businesses.
Amazon’s $4B investment in Anthropic ties the AI rival to AWS chips, signaling a challenge to Nvidia’s dominance.
Recommended by LinkedIn
U.S. tech restrictions on China have slowed Chinese advancements in semiconductors but unintentionally spurred innovation in other sectors, while further decoupling could harm both nations’ economic and technological trajectories.
The U.S.-China chip war escalates as Chinese industry groups warn against U.S. chip reliability, urge local alternatives, and retaliate with rare mineral export bans, signaling deepening tech and trade tensions.
Why US restrictions on the Chinese semiconductor industry failed
Western governments’ restrictions on China’s access to advanced tech have led Chinese companies to aggressively recruit global tech talent, raising concerns about trade secrets and technological competition.
Amid regulatory and hardware constraints, Chinese AI developers are advancing with groundbreaking innovations in efficiency, real-world applications, and affordability, reshaping global AI dynamics while rapidly reaching end users.
Explores the pivotal role of semiconductors in technological advancement, economic growth, and national security, emphasizing how modern industrial policies across the U.S., China, and other major economies are shaping the industry’s evolution through subsidies, export controls, and geopolitical strategies.
NVIDIA’s AI Success Reshapes Intel’s Future
NVIDIA’s rise in AI chip dominance has caused Intel’s stock to drop by 50%, led to CEO Pat Gelsinger’s resignation, and resulted in Intel’s removal from the Dow Jones Industrial Average after 25 years.
Data centers powering artificial intelligence are consuming rapidly increasing amounts of electricity, often surpassing the energy needs of cities, with renewable energy insufficient to meet demand, necessitating reliance on natural gas and potentially nuclear power, which challenges carbon emissions targets.
Google has partnered with Kairos Power to accelerate the deployment of small modular reactors (SMRs), aiming to supply up to 500 MW of 24/7 carbon-free power by 2035, supporting clean energy goals and AI-driven growth.
Amazon is advancing small modular reactors (SMRs), supporting existing nuclear plants, and funding next-generation technologies to meet energy demands sustainably and achieve its Climate Pledge commitment.
Meta announced plans to partner with developers for 1–4 GW of new nuclear capacity by the 2030s, supporting AI innovation, grid reliability, and its commitment to a clean energy future.
AI models are not mystical; they mimic the collective knowledge of human data labelers, making them powerful but far from infallible sources for complex decision-making.
Apple is facing challenges adapting Baidu’s AI models, including prompt accuracy and privacy conflicts, which may impact growth in China, where iPhone sales dipped 0.3% in the third quarter as rival Huawei surged by 42%.
Achieving ROI from AI investments requires a strategic focus on high-impact projects, scalable use cases, clean data, a culture of experimentation, and transforming people and processes.
Salesforce shares surged 11% to a record high after exceeding quarterly sales estimates and forecasting strong growth driven by its new AI-integrated tools, including Agentforce, which secured 200 deals shortly after launch and is projected to fuel significant revenue gains through 2026.
Silverside AI has been selected by The Coca-Cola Company to develop an AI-generated version of the classic 1995 “Holidays Are Coming” advertisement for the 2024 holiday season, aiming to blend innovative technology with the nostalgic essence of the original campaign.
2nd Coca Cola AI Ad
AI voice cloning can bypass bank voice ID security, highlighting vulnerabilities in voice-based authentication systems despite claims of added security by banks.
AI-powered voice synthesis has become so realistic that it’s hard to tell apart from human voices, posing security risks and ethical concerns, though small details like natural mistakes, tone, and how sentences are spoken can still give it away.
He thought he had his mum on the phone… It was actually a hoax, generated by artificial intelligence. Her mom’s voice was hacked and then cloned.
Researchers developed AI agents that replicate individual decision-making and behaviors with 85% accuracy using two-hour qualitative interviews, outperforming traditional demographic-based models in capturing nuanced human complexity.
Exa Websets deploys thousands of AI agents to thoroughly scour the web, turning it into a comprehensive database, offering users unparalleled access to complete and detailed knowledge.
AlphaChip is transforming microchip design by creating a feedback loop where AI-designed chips improve AI models, which in turn drive better chip designs, powering advancements across Alphabet’s TPUs and CPUs.
GPT-3.5-turbo-instruct outperforms newer models at chess despite its smaller size, likely due to OpenAI’s better-curated chess training data and the nuances of instruction tuning and chat interfaces, revealing that model behavior can depend heavily on fine-tuning, prompts, and other “contortions.”
Advanced AI systems are poised to revolutionize scientific research by independently analyzing data, testing hypotheses, and collaborating globally, potentially accelerating breakthroughs in medicine, energy, and climate solutions, while raising ethical and safety challenges.
A $50K portfolio managed by ChatGPT has achieved a 32% YTD return, with 12 new December stock picks chosen based on AI-driven analysis.
Principles for Dealing with the Changing World Order (5-minute Version) by Ray Dalio — to put the AI/Chip race in perspective
In Principles for Dealing with the Changing World Order, Ray Dalio examines historical cycles of national rise and decline, offering insights into the current global shifts and what the U.S. must do to remain strong.
Artificial Expert Intelligence through PAC-reasoning — Amnon Shashua and collaborators
AAI, a stealth-mode company founded by leading researchers, has developed a framework for AI to achieve “great scientist” capabilities, addressing limitations of current LLMs and demonstrating promising results in algorithmic expertise using CodeForces as a testbed.
One Zero Bank: Turning GenAI Into Real-World Innovation — Another Amnon Shashua venture
One Zero Bank defies the usual image of bureaucratic banks by implementing groundbreaking GenAI solutions, showcasing real-world applications, innovation, and a refreshing approach to banking.
The Mobileye Safety Methodology for Fully Autonomous Driving — Another Amnon Shashua venture
Mobileye’s safety methodology for autonomous driving combines achieving a high mean time between failures (MTBF) with eliminating unreasonable risks through transparent, accountable, and rigorous safety standards to ensure safer self-driving systems.
The first part of “Unleashing the Full Power of GPUs for Ansys Fluent” highlights how leveraging GPUs for computational fluid dynamics (CFD) accelerates simulations by up to 32X, reduces power consumption by 4X, and maintains industry-leading accuracy, setting a new standard for sustainable, efficient CFD processes.
This is a python API which allows you to retrieve the transcript/subtitles for a given YouTube video.
Large Language Models (LLMs) lack the memory, accuracy, and persistency required for true simulations or digital twins, functioning instead as statistical models that predict likely responses, making them fundamentally unsuitable for modeling or simulating real-world systems.
MIT researchers have developed a photonic processor that performs all deep neural network computations optically on a chip, enabling high-speed, energy-efficient AI processing for applications like telecommunications and real-time learning.
Burr is an open-source Python framework designed to simplify building and debugging GenAI applications by offering state management, pluggable tools, and a debugging UI, enabling faster, more efficient development workflows.
Fun
Jobs
AI Videos of the Week
Here are five AI-relevant insights from Marc Andreessen’s discussion on Joe Rogan’s podcast:
For a broad selection of AI videos, check out the INSEAD AI YouTube Channel (regularly updated)
Past editions — AI NEWS YOU MISSED
SUBSCRIBE TO GET UPDATES IN YOUR MAILBOX
Follow Robert Maciejko on LinkedIn or X (Twitter)
Intrapreneur | Strategic Innovation and Digital Transformation | Startup Investor | Inspiring Teams to Achieve Excellence, Through Passion, Commitment and Teamwork
3wAlready #50 !! An effervescence of advancements and headlines