NerdNews

Tesla has launched its Full Self-Driving (FSD) driver-assistance system in China, one of 10 countries where the supervised version is now available. The technology will cost Chinese customers a one-time fee of 64,000 yuan ($9,410) and is listed as an add-on option for the Tesla Model 3. Unlike the current version, unsupervised FSD will offer fully automated driving with no need for human intervention.

SpaceX is betting its future on beating Big Tech at AI, with a projected $26.5 trillion market opportunity. However, its Grok AI chatbot has lagged behind competitors like OpenAI and Anthropic. SpaceX plans to deploy up to 1 million satellites as orbital data centers to support its AI ambitions.

Google is set to remake search with agentic AI in 2026, accelerating its AI search evolution. The new search experience will employ agents to answer questions, powered by Gemini 3.5 Flash, and may create custom apps to help users with their queries. This shift is expected to change the way people interact with search engines, with fewer blue links and more AI-generated content.

Google has announced Gemini 3.5 Flash, a more efficient AI model that offers frontier-level intelligence while being cost-effective. The model can output nearly 300 tokens per second and has been rolled out across various Google products. Additionally, Google has introduced Gemini Spark, an AI agent that can run multiple agentic workflows, and Gemini Omni, a multimodal model that can accept any kind of input data and produce various types of output.

Spotify has introduced AI-powered Q&A and briefing generation features to its podcasts. Users can create personalized podcasts based on custom prompts, and the app will generate content. Additionally, a dedicated desktop app called Studio by Spotify Labs allows users to create personalized briefings. The company is also rolling out an AI-powered Q&A feature for Premium mobile users.

OpenAI's ChatGPT is now integrated with Microsoft PowerPoint, allowing users to generate, edit, and update presentation slides using natural language prompts or connected services like Gmail and Outlook. The feature is currently in beta and available to most users, including those on the free access tier and subscribers to ChatGPT Business.

Workflow automation is becoming trendy due to AI and industry standards like BPMN. It enables adaptive process automation, making it possible for organizations to automate workflows and drive operational change. Modern workflow automation uses microservice-based architecture, allowing for more lightweight and scalable enterprise systems.

J. Paul Reed discusses the ironies of automation and AI in incidents, highlighting how advanced systems can make human operators more crucial, while simultaneously degrading the skills needed to intervene. He shares real-world stories of AI-fueled incidents and provides guidance on how to maintain resilience.

A study found that large language models can be shaped by authoritarian states through state media, potentially spreading propaganda. This can happen when AI models are trained on data that includes state-scripted media, which can be more prominent in certain languages. The study tested this hypothesis using China as a test case and found that AI models can give more pro-regime answers when prompted in Chinese. This raises concerns about the potential for AI to spread state propaganda and highlights the need for AI labs to ensure their models are not biased towards authoritarian regimes.

Google's SynthID AI watermarking technology is being adopted by OpenAI, Nvidia, and more to help identify AI-generated content. SynthID is a digital watermark that is deeply integrated with AI-generated content and is robust to different kinds of transformations. The technology has been used to label 100 billion images and videos, and will be integrated with Circle, Search, Lens, and AI Mode to allow users to check for the watermark.

Google introduced a redesigned Android command-line interface (CLI) that enables building apps up to 3x faster by using AI agents. The new CLI provides consistent, scriptable access to the Android toolchain and is compatible with third-party agents. It also includes Android Skills, a collection of files that describe how specific development tasks should be performed, and a built-in knowledge base that agents can query in real time.

Figure AI's humanoid robots have been livestreaming package handling for nearly a week, with 24/7 autonomous operation and a competition against a human intern. The robots use the Helix 02 neural network system for full-body control and long horizon autonomy.

Two AI-based science assistants, Google's Co-Scientist and FutureHouse's Robin, have successfully completed drug-retargeting tasks by generating hypotheses and analyzing biological data. These tools can help scientists stay on top of the vast amount of scientific information and identify potential connections between different fields, potentially leading to new discoveries and treatments.

University graduates are expressing outrage towards corporate executives who praise AI during commencement ceremonies, citing concerns over job replacement and the technology's consequences. The backlash is part of a larger movement against AI, with many young people skeptical of its benefits and wary of its impact on their futures.

Stellantis has partnered with Wayve to bring hands-free driving to its vehicles in 2028. Wayve's self-driving system uses an end-to-end neural network and can run on various chips and sensors, making it an attractive option for cost-sensitive automakers. The technology will be used in Stellantis' North American market first, with plans to expand to other regions.

Elon Musk's xAI is being sued for using polluting generators at its data center. Despite this, the company plans to buy $2.8 billion more turbines, which could worsen air quality. The lawsuit claims xAI is operating dozens of unregulated gas turbines, emitting over 2,000 tons of NOx pollution annually.

Anthropic is expected to post $10.9 billion in revenue for the quarter ending in June, making it the company's first profitable quarter since its founding in 2021. The company expects to post $559 million in operating profit, driven by the growth of its AI products, including its chatbot Claude. This milestone is significant, as rival OpenAI has yet to reach profitability and doesn't expect to until 2029 or 2030.

Nokia and KDDI have successfully tested Intelligent 4D Resource Optimisation Technology, which reduces energy consumption in 5G/6G infrastructure by up to 40% while maintaining high communication quality. The technology optimizes the control of four key radio resources: time, frequency, space, and transmission power.

Waymo has paused its robotaxi service in four cities, including Atlanta, San Antonio, Dallas, and Houston, due to issues with its vehicles driving into flooded areas. The company is working to develop a solution to avoid flooded roads and has shipped an update to its fleet with restrictions in high-risk areas. The pause is a precautionary measure to ensure safety, and Waymo is in communication with the National Highway Traffic Safety Administration (NHTSA) regarding the issue.

Waymo has suspended its robotaxi service on freeways in several cities due to struggles with construction zones. The company is working to improve its software and expects to resume freeway rides soon. This interruption comes as Waymo aims to expand its service to new cities and offer up to one million paid rides per week by the end of 2026.