Yoshua Bengio is a computer scientist at the University of Montreal in Canada. In 2019, he won an A. M. Turing Award ' considered the most prestigious honour in computer science ' for pioneering the 'deep learning' techniques that are now making artificial intelligence (AI) ubiquitous. Last month, he also became the first person to top one million citations on Google Scholar. Bengio has since turned his focus to exploring the risks posed by AI. He chairs an international panel of advisors in this field, which includes representatives from 30 countries, the European Union, the OECD and the United Nations. The group issued the International scientific report on the safety of advanced AI earlier this year. The string of papers that I co-authored on language modelling and attention ' that started in the late '90s ' on how we could introduce attention mechanisms in neural nets to make them more 'system 2', meaning more deliberate, and not just intuition machines. Yes, but I would mention also another paper, which doesn't get nearly as much attention ' and that's the work on curriculum learning2, in which a machine is trained by feeding it data in a particular order rather than at random. It has become the standard way of doing things. The inspiration for me was learning in animals....

Vishal Gupta, engineering manager, machine learning at Reddit, joins the Me, Myself, and AI podcast with host Sam Ransbotham to explain how the social media community uses artificial intelligence to improve user experience and ad relevance. Much of the advertising work relies on increasingly sophisticated recommender systems that have evolved from simple collaborative filtering to deep learning and large language model'based systems capable of multimodal understanding. Vishal and Sam also explore the philosophical and ethical aspects of AI-driven platforms. Vishal emphasizes the importance of balance ' between exploration and exploitation in recommendations, between advertiser goals and user experience, and between human- and machine-generated content. He argues that despite the rise of AI-generated material, authentic human conversation remains vital and even more valuable as models depend on it for training. Vishal Gupta is a seasoned engineering leader who leads multiple artificial intelligence and machine learning teams at Reddit in the ads domain. He has a decade of experience working on cutting-edge machine learning techniques at companies like DeepMind, Google, and Twitter. Gupta is passionate about applied AI research that significantly contributes to a company's top and bottom lines....

Initially launched as an iOS app in September, Sora quickly rose to the top of the App Store charts, amassing over 1 million downloads in a week. With its arrival on the Google Play Store, Sora is expected to attract a larger user base, likely resulting in a surge in downloads. The Android version retains all the features of its iOS counterpart, including the 'Cameos' feature, which allows users to generate videos of themselves performing various activities using their own likeness. The videos can be shared in a feed reminiscent of TikTok, allowing users to discover and engage with content from others. This appears to be a strategic move by OpenAI to strengthen its position in the competitive landscape of short-form video sharing. The AI giant aims to rival major players like Meta, which has recently launched its own AI video feed called Vibes, as well as existing platforms such as TikTok and Instagram. However, the app has faced criticism for its handling of deepfakes. After its initial launch, users began uploading disrespectful videos of historical figures, including Martin Luther King Jr. As a result, Sora paused the generation of content depicting Dr. King last month and strengthened its guardrails....

A decade after Samsung's Family Hub Refrigerator symbolized the smart home revolution, 'smartness' alone no longer sells. Although the smart home products industry is projected to hit $537 billion by 2030, roughly three-quarters of projects still fail to deliver value. Part of the problem is that many brands rely on vague claims of intelligence''advanced algorithms,' 'deep learning''without showing how products meet customer needs. The authors' research, published in the Journal of Consumer Research, shows that this approach can hurt sales. Surveying 1,000+ customers across 31 categories, they identified three buyer personas: comfort-seekers (46%), purpose-seekers (18%), and efficiency-seekers (36%). Defined by how customers extract value, these groups differ sharply. Comfort-seekers prize ease and fun; purpose-seekers want mission alignment and data; efficiency-seekers focus on ROI and time savings. Success depends on tailoring design and messaging to each persona's needs....