Listen To The Show
Transcript
Welcome to The Prompt by Kuro House, your daily AI update. Today, we’ve got some fascinating developments from Apple, Europe, and robotics. So, let’s dive right in and see what’s shaping the future of artificial intelligence.
Apple is reportedly working on a major AI upgrade for Siri that could lean on Google’s technology. According to Bloomberg’s Emma Roth, Apple’s new feature, internally called “World Knowledge Answers,” aims to deliver AI-generated summaries using web results. This means Siri won’t just answer questions but will also show text, photos, videos, and points of interest, making it a real contender against AI search tools like OpenAI’s. Interestingly, Apple and Google have reached a formal agreement for Apple to test a Google-designed AI model for Siri’s summaries, while Apple evaluates other AI models for planning functions. Apple plans to launch this upgraded Siri alongside iOS 26.4 as early as next March, so keep an eye out.
Over in Europe, Switzerland has launched its own open-source AI model named Apertus. As reported by The Verge’s Elissa Welle, Apertus was trained exclusively on public data from websites that allowed scraping, respecting copyright and AI ethics. The model supports over 1,800 languages and comes in two sizes, with 8 billion and 70 billion parameters, making it comparable to Meta’s Llama 3. This move sets a new standard for trustworthy and legally compliant AI models in Europe, providing an alternative to proprietary models like ChatGPT and Claude. All aspects of Apertus, from source code to model weights, are fully open and available on HuggingFace.
French AI startup Mistral is on the verge of a massive $14 billion valuation, Bloomberg reports. The company, founded by former DeepMind and Meta researchers, just closed a €2 billion investment round, making it one of Europe’s most valuable tech startups. Mistral develops open-source language models and an AI chatbot called Le Chat, tailored for European users. This surge in valuation reflects a broader trend, with European AI startups securing 55% more investment in early 2025 and 12 new unicorns already this year. It’s clear that Europe is rapidly becoming a powerhouse in AI innovation and investment.
In a legal showdown, Scale AI is suing a former employee and rival company Mercor over alleged theft of confidential customer data. TechCrunch reports that Scale claims the ex-employee stole more than 100 confidential documents and tried to pitch Mercor to one of Scale’s biggest clients, dubbed “Customer A.” Mercor denies using any of Scale’s trade secrets but admits investigating some files found in a personal Google Drive of the employee. This lawsuit highlights the fierce competition in the AI data training space, especially as Meta recently invested $14.3 billion for a stake in Scale AI. It’s a reminder that behind AI’s technical advances, there’s intense corporate rivalry and legal battles shaping the industry.
Finally, Boston Dynamics’ humanoid robot Atlas is showing remarkable progress by mastering walking and grasping with a single AI model. Wired’s Will Knight explains that this generalist model controls both arms and legs, allowing Atlas to perform complex tasks like rebalancing while reaching or instinctively picking up dropped items. This approach mimics the breakthroughs seen in large language models, hinting at a future where robots can learn a wide range of skills without retraining. Experts say this could be a big step toward robots that work seamlessly in messy, real-world environments. It’s exciting to see robotics moving closer to the kind of generalized learning that revolutionized AI with ChatGPT.
That’s all for today’s AI highlights on The Prompt. From Apple’s Siri to European open models, legal battles, and humanoid robots, the pace of change is incredible. Thanks for tuning in, and we’ll catch you tomorrow with more AI news you need to know.
