Toggle light / dark theme

What is consciousness, and could AI have it?

In the Voltaire Lecture 2025, Professor Anil Seth will set out an approach to understanding consciousness which, rather than trying to solve the mystery head-on, tries to dissolve it by building explanatory bridges from physics and biology to experience and function. In this view, conscious experiences of the world around us, and of being a ‘self’ within that world, can be understood in terms of perceptual predictions that are deeply rooted in a fundamental biological imperative – the desire to stay alive.

At this event, Professor Seth will explore how widely distributed beyond human beings consciousness may be, with a particular focus on AI. He will consider whether consciousness might depend not just on ‘information processing’, but on properties unique to living, biological organisms, before ending with an exploration of the ethical implications of an artificial intelligence that is either actually conscious – or can convincingly pretend to be.

How Distillation Makes AI Models Smaller and Cheaper

Considering that the distillation requires access to the innards of the teacher model, it’s not possible for a third party to sneakily distill data from a closed-source model like OpenAI’s o1, as DeepSeek was thought to have done. That said, a student model could still learn quite a bit from a teacher model just through prompting the teacher with certain questions and using the answers to train its own models — an almost Socratic approach to distillation.

Meanwhile, other researchers continue to find new applications. In January, the NovaSky lab at the University of California, Berkeley, showed that distillation works well for training chain-of-thought reasoning models, which use multistep “thinking” to better answer complicated questions. The lab says its fully open-source Sky-T1 model cost less than $450 to train, and it achieved similar results to a much larger open-source model. “We were genuinely surprised by how well distillation worked in this setting,” said Dacheng Li, a Berkeley doctoral student and co-student lead of the NovaSky team. “Distillation is a fundamental technique in AI.”

​Tsinghua University holds Tsinghua AI Agent Hospital Inauguration and 2025 Tsinghua Medicine Townhall Meeting-Tsinghua University

On the morning of April 26, Tsinghua University held an inauguration ceremony for Tsinghua AI Agent Hospital and the 2025 Tsinghua Medicine Townhall Meeting at the Main Building Reception Hall. Tsinghua President Li Luming and Vice President Wang Hongwei attended the event.

President Li Luming reviewed the progress of Tsinghua University’s medical programs over the past year, emphasizing the University’s strong commitment to the development of medical disciplines. He highlighted Tsinghua’s strength in fundamental research in Artificial intelligence, which has already led to a series of high-level innovations at the intersection of AI and medicine. The establishment of the Tsinghua AI Agent Hospital represents a new initiative by Tsinghua to leverage its strengths in science and engineering to empower the advancement of medicine.

President Li encouraged Tsinghua Medicine to remain committed to fostering virtue and talent, cultivating a new generation of medical innovators with both a strong medical foundation and AI literacy. He also called for deeper integration across disciplines, particularly between engineering and medicine, as well as closer ties between clinical practice and technology. Finally, he urged Tsinghua Medicine to align its work with cutting-edge global trends and national strategic needs, driving medical advancement and contributing to the protection of public health.

Tesla FSD Competitors Admit DEFEAT: “Elon Was Right”

Questions to inspire discussion.

Safety and Performance.

🛡️ Q: How does Tesla’s full self-driving system compare to human driving in terms of safety? A: According to Elon Musk, Tesla’s end-to-end neural networks trained on massive video datasets have been proven to be dramatically safer than average human driving.

⚡ Q: What recent hardware upgrade has improved Tesla’s full self-driving capabilities? A: Tesla’s AI4 hardware has been upgraded to 150–200 watts, enabling more complex neural networks and faster decision-making, achieving 36 frames per second processing.

Scalability and Efficiency.

📈 Q: Why is Tesla’s vision-only approach considered more scalable than competitors’ methods? A: Tesla’s vision-only approach is more scalable than competitors’ use of multiple sensors, sensor fusion, and high-definition maps, as stated by BU’s Robin Lee.

SpaceX Finally Reveals Starship Flight 10 Launch Window

🚀 Q: When are the next Starship test flights scheduled? A: Flight 10 is targeting August 4th, while Flight 11 is set for September 1st, 2025, marking the final Block 2 Starship tests.

🛰️ Q: What new AI initiative is SpaceX undertaking? A: SpaceX is hiring AI software engineers to integrate artificial intelligence into engineering workflows supporting Falcon, Starship, and satellite operations.

Common feature between forest fires and neural networks reveals universal framework

Researchers from the University of Tokyo in collaboration with Aisin Corporation have demonstrated that universal scaling laws, which describe how the properties of a system change with size and scale, apply to deep neural networks that exhibit absorbing phase transition behavior, a phenomenon typically observed in physical systems. The discovery not only provides a framework describing deep neural networks but also helps predict their trainability or generalizability. The findings were published in the journal Physical Review Research.

In recent years, it seems no matter where we look, we come across in one form or another. The current version of the technology is powered by : numerous layers of digital “neurons” with weighted connections between them. The network learns by modifying the weights between the “neurons” until it produces the correct output. However, a describing how the signal propagates between the layers of neurons in the system has eluded scientists so far.

“Our research was motivated by two drivers,” says Keiichi Tamai, the first author. “Partially by industrial needs as brute-force tuning of these massive models takes a toll on the environment. But there was a second, deeper pursuit: the scientific understanding of the physics of intelligence itself.”

Netflix’s ‘The Eternaut’ Pioneers Generative AI for 10x Faster VFX, Sparking Hollywood Job Debates

Netflix’s “The Eternaut,” an Argentine sci-fi series, pioneers generative AI for a building collapse scene, enabling 10x faster VFX and cost savings. Co-CEO Ted Sarandos sees it empowering creators, not replacing them. Mixed reactions highlight job fears, signaling AI’s growing role in Hollywood amid ethical debate.