Nvidia has detailed two variants of its Arm-based N1X chip that will power RTX Spark laptops arriving next month. The higher-end version pairs a 6144-core Blackwell RTX GPU with a 20-core Grace CPU and up to 128GB of unified memory, while the scaled-down laptop version uses a 5120-core GPU, an 18-core CPU, and up to 32GB of memory. Devices from Asus, Dell, HP, Lenovo, Microsoft, and MSI are expected to launch in October, alongside mini PCs like Microsoft's Surface RTX Spark Dev Box.
The article revisits Steve Jobs' view that effective leaders push people to improve by setting demanding standards and giving honest feedback rather than shielding them from discomfort. It challenges the common workplace notion that caring for employees means avoiding friction or difficult conversations.
A recent IEEE Spectrum careers newsletter, produced with Parsity, argues that engineers can gain mentorship benefits without ever formally asking someone to be a mentor. The author describes learning by observing and copying admired colleagues' habits, and by asking small, targeted questions about how others solve problems or approach their work.
Uber confirmed on Wednesday it is cutting roughly 3,300 jobs, or about 10% of its 34,000-person workforce as reported in its latest annual filing. CEO Dara Khosrowshahi framed the move as an effort to make the company 'simpler' and 'faster,' without citing artificial intelligence as a driver of the cuts. Uber's stock rose following the announcement.
A collection of Rolodex-style business cards from Los Alamos National Laboratory has been compiled and shared publicly, offering a glimpse into the institution's personnel and administrative history. The cards represent a range of roles and eras at the storied nuclear research facility.
A technical walkthrough details how diffusion language models work as an alternative to the dominant autoregressive approach used by most LLMs today. Rather than generating text token-by-token in sequence, diffusion models produce an entire sequence at once and refine it over multiple steps, drawing on techniques like masking, iterative refinement, and post-training used in recent open-source models. The material stems from workshop talks and lectures given at ICLR 2026 and MLSS 2026.
A wave of new research papers is reviving interest in continuous diffusion techniques for language generation, an approach that had largely been overtaken by discrete diffusion and autoregressive methods in recent years. The author reviews this resurgence, tracing why continuous diffusion for text fell out of favor and why several research groups are now revisiting it as an alternative to token-by-token autoregressive generation.
The piece explains how Google's Tensor Processing Units (TPUs), built for AI workloads in data centers, differ from GPUs and NPUs. TPUs are now replacing NPUs in Google's upcoming Pixel 11 phones, with Google claiming up to 3.5 times faster AI processing at up to 3.5 times less energy use. GPUs, meanwhile, remain versatile chips handling gaming, 3D rendering, AI training and crypto mining.
AI language models fall into three broad categories based on how much of their inner workings are shared publicly: proprietary systems that keep code and weights closed, open-weight models that release trained parameters without full training data or code, and open source models that make the entire stack available for inspection and modification. The distinctions affect how much control, transparency, and customization developers and businesses get when building on top of these systems.
A unanimous three-judge panel on the US Court of Appeals for the 9th Circuit ruled that Nevada can enforce its gambling laws against Kalshi's sports-related event contracts, rejecting the company's claim that federal commodities law shields it from state oversight. The decision upholds a lower court order allowing Nevada's Gaming Control Board to act against Kalshi, Crypto.com and Robinhood over sports-event contracts offered on their platforms.
Ahead of a September 9 unveiling, reports suggest Apple Watch Ultra 4 will feature substantially upgraded health sensors, with Digitimes claiming the number of sensor components could double, alongside a new blood pressure-related feature reportedly under FDA review. Analyst Mark Gurman also points to fitness-tracking upgrades and a new, faster S-class chip replacing the S10 used since 2024, though reports conflict on whether the watch will get a major design overhaul.
Streaming tags like 'Dolby Atmos' and 'Lossless' refer to two separate audio technologies: Atmos creates a 3D soundstage by placing sounds in space, while lossless audio preserves the full quality of a recording without the data loss found in formats like MP3 or AAC. Because Atmos typically relies on the lossy Dolby Digital Plus codec, and lossless streaming uses formats like FLAC or ALAC, most platforms can't offer both simultaneously.
GoKawiil