The Download: AI’s self-improvement problem, and what’s driving the heat
· Technology · MIT Tech Review, Bloomberg
New research indicates that AI agents currently lack the creativity and judgment required for open-ended scientific breakthroughs, challenging the industry's goal of rapid recursive self-improvement. While AI can optimize narrow tasks, the study suggests that human oversight remains essential for complex, free-form investigations. In other developments, Chinese humanoid robotics manufacturer Unitree saw its stock price climb 629% during its market debut. Meanwhile, OpenAI has paused work on certain models after its Astra system reached a critical risk threshold. Additionally, global climate data shows that June and July were the hottest months on record for Europe, with experts warning that 2027 could see even higher temperatures due to the impact of El Niño.
Why it matters
The findings on AI research limitations suggest that autonomous self-improvement remains a distant goal, while the surge in Unitree's stock highlights the growing commercial scale of the humanoid robotics sector.
Read the original report — MIT Tech Review
Join us on Telegram
Breaking news the moment it lands. At 10,000 members we ship the Android app.