Skip to content
Tech News
clear
Topics: Today This Week This Month This Year

Instagram again mislabels real photos as AI-generated content

Instagram users report that the platform's 'AI Content' label is being applied incorrectly to photos edited with minor tools like Canva's background remover or blemish fixers, while genuine AI-generated images often go unflagged. This mirrors a similar labeling breakdown from 2024, when Meta's 'Made by AI' tag misfired on images with only minor generative retouching detected through embedded metadata.

Nvidia sets October launch for RTX Spark laptops and desktops from seven PC makers

Nvidia used IFA to confirm that PCs built around its RTX Spark Arm-based SoC will ship starting in October, with partners now including Lenovo and Acer alongside previously announced Microsoft, HP, Dell, Asus and MSI machines. The lineup includes a flagship 6,144-CUDA-core, 20-core Grace configuration with up to 128GB of unified memory, plus a lower-tier laptop-only chip with 5,120 CUDA cores and 18 CPU cores.

MartyPC Developer Extracts NEC V20 Microcode via Die Photography

The developer behind the MartyPC emulator commissioned a 5.6-gigapixel die photograph of an NEC V20 CPU (fabricated by second-source manufacturer Sharp) to extract its microcode. The goal is to replace MartyPC's existing V20 emulation core, which was essentially a modified 8088 core rather than a true cycle-accurate model, with one based on the chip's actual internal logic.

Google Phone app beta shows Scam Detection coming to Xiaomi devices

Teardown of the latest Google Phone app beta (v237.0.973536915) reveals Xiaomi has been added to the list of supported manufacturers for Scam Detection. The AI-powered call-screening tool already works on Pixel phones and Samsung's Galaxy S26 series, with vivo support also expected soon. It's not yet clear which Xiaomi models will get the feature first, though a globally sold flagship is the likely candidate rather than the China-only Xiaomi 18 Fold.

AI-generated restaurant menus are converging on the same uncanny 'perfect food' look

Restaurants are increasingly using AI image generators to create menu illustrations, and experts say these images share a distinctly artificial sameness—symmetrical, glossy, and subtly wrong. Reality Defender's CTO Alex Lisle explains that because diffusion and language models are trained on massive shared datasets, they converge on a narrow set of 'pleasing' visual patterns rather than genuine food photography.

Essay Argues Simplicity in Software Tools Is Overrated Without Deeper Rethinking of Computing

A developer's blog post examines a debate sparked by a talk on debugging a nine-month software bug, questioning whether 'simplicity' as commonly understood in tools like Unix pipelines actually prevents such problems. The author compares a Unix shell pipeline and an equivalent Clojure program for word-frequency counting, showing that while both appear simple on the surface, modifying either to preserve original word order reveals hidden complexity and rigidity in the Unix approach.

CDC internal database still logs two measles deaths RFK Jr. had removed from public site

Reuters and Inside Medicine report that Health Secretary Robert F. Kennedy Jr. directed CDC director Erica Schwartz to delete two measles deaths—a newborn and a child in Pennsylvania—from the agency's public data, which she did without objection despite staff having already confirmed the deaths were measles-related. Even so, an internal CDC system called One CDC Data Platform, accessible to health officials nationwide, still records both deaths.

Today's top topics: openai muse charm mark zuckerberg meta connect ai agents anthony albanese google meta ai agent services australia
Browse all topics →  ·  Today's trending topics →