AI leaders struggle to push limits of new models

+ Should we be worried about AI’s feelings?

November 13, 2024

Save 1 hour every day with an AI executive assistant: Fyxer AI automates daily email and meeting tasks, saving users up to 60 minutes a day.

  • Prioritizes important emails, ensuring high-value messages are addressed quickly.

  • Automatically drafts email responses in your tone of voice. It's so accurate that 63% of emails are sent without edits.

  • Listens to your video calls and sends a meeting summary with clear next steps.

Latest headlines 👇

OpenAI, Google, and Anthropic face challenges scaling AI as new models, like OpenAI's Orion, fall short of expectations. Diminishing returns, data scarcity, and rising costs are driving a shift toward alternative methods like reasoning-based techniques. Experts say the race for artificial general intelligence is slowing, with companies exploring smaller, incremental updates and practical AI tools like agents. Read more

As companies strive to develop more intelligent and human-like AI systems, scientists and philosophers are increasingly debating the possibility of AI attaining consciousness or emotions. A recent report by Eleos AI, featuring experts like NYU’s David Chalmers, highlights the uncertainty surrounding AI sentience and calls for serious consideration of AI welfare by businesses and policymakers. While some researchers believe true artificial consciousness is unlikely, others warn of significant ethical risks if AI were to develop the capacity to feel. This debate underscores the need for greater oversight and thoughtful regulation as AI technologies continue to advance. Read more

Baidu has launched AI-integrated smart glasses, powered by its Ernie language model, as it vies with global rivals like Meta to dominate the AI hardware market. The glasses, set to launch next year in China, offer features like calorie tracking, music playback, and video recording. While lagging in large language model development, Chinese tech companies aim to leverage their advanced electronics sector to produce competitive AI-driven consumer hardware. This move signals intensifying competition among Chinese internet giants in the AI-integrated devices space. Read more

Elon Musk’s xAI has rapidly built Colossus, a record-breaking supercomputer powered by 100,000 GPUs, challenging rivals like OpenAI and Meta. Constructed in just 122 days at a Memphis data center, this feat upended industry norms, sparking curiosity and concern among competitors. Musk bypassed traditional methods, using gas turbines for power and cutting regulatory corners to accelerate progress.

Rivals, including OpenAI and Amazon, are now scrambling to scale their own AI infrastructure, with new facilities under construction in Texas. The race highlights escalating stakes in the quest for AI dominance. Read more

Former OpenAI CTO Mira Murati is building a new venture, recruiting key OpenAI alumni. Mianna Chen, a research program manager instrumental in launching GPT-4o and Advanced Voice Mode, is reportedly joining Murati alongside ex-head of post-training Barret Zoph and senior researcher Luke Metz. While the startup's focus remains unclear, Murati’s team has sparked interest from VCs. OpenAI, meanwhile, faces challenges as gains from pre-training slow, though it plans to release its full o1 reasoning model by year’s end. Read more

OpenAI co-founder and president Greg Brockman is back after a three-month sabbatical, ahead of schedule. His return follows the recent departures of key executives, including CTO Mira Murati and Chief Research Officer Bob McGrew. Brockman is now working with CEO Sam Altman to shape a new role addressing critical technical challenges for the company. Read more