In today’s newsletter:

• News - 😬 NYT Accuses OpenAI of Erasing Key Lawsuit Data

• Culture - 🌟 Top AI Leaders and Their Predicted Artificial General Intelligence (AGI) Timelines

😬 NYT Accuses OpenAI of Erasing Key Lawsuit Data

The ongoing copyright battle between The New York Times and OpenAI is heating up, with the newspaper accusing the AI company of erasing critical data during the discovery process. OpenAI was tasked with providing its training data in a controlled “sandbox” environment for review, but The Times claims a technical mishap erased over 150 hours’ worth of organized data. Although OpenAI managed to recover much of it, the file names and folder structures were lost, forcing The Times’ team to start from scratch. OpenAI has called the issue a “glitch,” but The Times isn’t thrilled with the significant resources required to redo their work.

The case underscores the growing tension between publishers and AI companies as the courts grapple with how generative AI tools use copyrighted material. While OpenAI and Microsoft argue that The Times should provide more documentation on its own use of AI, the newspaper is pushing for text messages, Slack chats, and even phone records from OpenAI executives. Meanwhile, OpenAI is working on licensing deals with other media giants like The Atlantic and Condé Nast, hinting at a potential shift in how publishers and AI companies collaborate—or clash—in the future. Whatever the outcome, these lawsuits will set the stage for how AI tools interact with copyrighted content in the U.S. moving forward.

🗞️ In Other News…

• Alibaba's Marco-o1, a new Large Reasoning Model, advances open-ended AI problem-solving by integrating techniques like Chain-of-Thought fine-tuning, Monte Carlo Tree Search, and dynamic reasoning strategies to excel in both structured and nuanced tasks.

• Uber plans to invest over $10 million in Pony.ai's IPO, potentially collaborate on projects outside the U.S., and expand its use of autonomous driving technology, as the self-driving sector faces both transformative potential and ongoing regulatory and technological challenges.

• MIT researchers have developed a more efficient reinforcement learning algorithm called Model-Based Transfer Learning (MBTL), which strategically selects tasks to train AI systems, significantly reducing training costs and improving performance for complex, variable problems like traffic control.

• Google Gemini’s new "memory" feature allows subscribers with Gemini Advanced access to save personal details, such as interests, work, and hobbies, to tailor responses. Users can view, edit, or delete this data via the "Saved Info" page, similar to OpenAI's Memory feature in ChatGPT Plus.

• Apple plans a major AI upgrade for Siri, called “LLM Siri,” to enable more conversational, natural interactions using Apple’s AI models. The assistant will handle advanced tasks, interact with third-party apps, summarize text, and gain screen context awareness, with full deployment expected by spring 2026.

🌟 Top AI Leaders and Their Predicted Artificial General Intelligence (AGI) Timelines

This AGI Timeline chart outlines projections from notable tech leaders about when artificial general intelligence (AGI) might be achieved. (AI systems being capable of performing any intellectual task that a human can do.) Sam Altman of OpenAI predicts AGI by 2025, Elon Musk forecasts 2026, and Demis Hassabis envisions 2030, with other figures like Ray Kurzweil and Geoffrey Hinton predicting advancements by 2029. These varied timelines reflect differing perspectives on when AI systems might surpass human intelligence across most tasks.

Elon Musk, despite his track record of overly ambitious predictions, has stated that AGI could be achieved by 2026, driven by his AI startup xAI. Musk attributes delays to hardware shortages but remains optimistic about progress. Similarly, Sam Altman projects AGI by 2025, emphasizing its potential to transform industries and society. However, experts like Oren Etzioni and others warn that the pace of AI improvements is slowing, and achieving AGI may take longer than expected. As the future of AGI remains uncertain, these bold predictions continue to drive investment and innovation, even as researchers question the hype surrounding its feasibility and timeline.

VC Fundraising Rounds

• Lightning AI, a New York City-based developer of the PyTorch Lightning framework, secured $50 million in funding. (11/21/24)

• Wordware, a San Francisco-based startup aiming to simplify programming AI agents by enabling users to write in plain English, has announced securing $30 million in seed funding. (11/21/24)

• xAI, Elon Musk's artificial intelligence startup, has announced to investors that it raised $5 billion in funding, boosting its valuation to $50 billion—over twice its worth from a few months prior. (11/20/24)

• Federato, an AI-driven underwriting platform for insurance, announced that it has raised a total of $80 million, including its recent Series C funding round. (11/20/24)

• Cyera, a leading company in data security, has secured $300 million in Series D funding, bolstering its swift expansion and solidifying its leadership in the $23.68 billion data security industry. (11/20/24)

Meme of the Day