An analysis argues that the cost of running machine learning models is dropping by orders of magnitude annually, driven by GPU efficiency gains that double roughly every two years—a pace not seen since early Moore's Law. The piece distinguishes proprietary models like GPT-6 Astra from open-weight models such as GLM-5.3-flash, noting that hosted and locally-run versions improve at different rates, with per-token pricing for frontier models not falling as consistently as costs for smaller models.
jyn.dev
· 2026-09-23
Qualcomm announced the Snapdragon 8 Elite Gen 6 and 8 Elite Extreme Gen 6 mobile processors at its Snapdragon Summit, built on a 2nm process with dual 5GHz Oryon Prime cores. The chips bring notable CPU, GPU and NPU gains over the prior generation, alongside power efficiency improvements, and will ship in upcoming flagship phones from brands including Honor, Motorola, Oppo, Vivo and Xiaomi. Qualcomm emphasized AI agent capabilities, such as automated video sorting and location-based notifications, as the lineup's central feature.
techspot.com
· 2026-09-23
Qualcomm announced its next flagship mobile chip, the Snapdragon 8 Elite Gen 6, built on a 2nm process with an eight-core Oryon CPU, a 35% faster Adreno GPU, and an NPU that's 14% faster with 20% better performance-per-watt. The chip supports 8K video at 30fps, up to 320-megapixel photo capture, and connectivity including Wi-Fi 8, Bluetooth 6 and a new 5G modem, with an unspecified Snapdragon 8 Elite Extreme Gen 6 variant offering additional features on the same architecture.
engadget.com
· 2026-09-22
A new Ornn Data paper argues that NVIDIA's older Ampere-generation GPUs, particularly the A100, remain economically useful far longer than assumed because open-weight models can be run cheaply on them. The analysis found that self-hosted open-weight models like gpt-oss-120b produce output more cheaply on A100 chips than on newer H100 hardware, and that A100 rental prices have stayed unusually stable over a five-year contract period compared to Hopper and Blackwell chips.
data.ornn.com
· 2026-09-22
AMD's newly announced RX 9050 GPU exists in two very different versions sharing the same name: an 8GB retail card sold standalone in some markets, and a 4GB variant that can only be bought bundled inside prebuilt desktops. Testing found the 4GB model, discovered inside a $750 CyberPowerPC system from Best Buy, has half the memory bandwidth and significantly worse performance than the retail card most buyers would research online.
techspot.com
· 2026-09-22
Loud laptop fans, especially in gaming laptops, are typically a normal response to heavy CPU and GPU workloads that generate heat needing ventilation. The report explains that fan speed increases with processing demand to prevent overheating, and outlines when loud noise is a genuine concern versus routine operation, along with tips to quiet things down.
engadget.com
· 2026-09-22
AX Gaming, an Inno3D subsidiary, has unveiled a seasonal limited-edition GeForce RTX 5070 12G graphics card decorated with anime artwork, maple leaf motifs, and ARGB lighting. The standout addition is a scent-emitting component the company calls 'e-sports aromatherapy,' claiming users can detect an autumn-like fragrance even when the card is powered off.
tomshardware.com
· 2026-09-21
Android Authority traced worsening game performance on the Pixel 11 series—stuttering and freezing in titles like Genshin Impact—to Google quietly making ANGLE the default OpenGL ES implementation on the new phones. ANGLE translates OpenGL ES calls into Vulkan commands, and testing showed it activating in popular games on the Pixel 11 Pro XL where it wasn't used on the Pixel 10 Pro XL.
androidauthority.com
· 2026-09-21
Hyte is offering its X50 PC case for $99.99, a $50 discount available on the company's website through September 21st, 2026. The case features a curved glass panel spanning its side and top, supports motherboards from ITX to E-ATX and GPUs up to 430mm long, and comes in colors like lavender-toned 'taro milk' and 'matcha' alongside black and white. A mesh-paneled variant, the X50 Air, is also discounted to $79.99 from its usual $119.99.
theverge.com
· 2026-09-19
A new report titled 'Pacing the Frontier' argues that despite widespread agreement among AI leaders that development may need to slow down, there is no clear technical or policy framework for how to actually do it. Coauthor Raymond Douglas of the University of Toronto says the field needs to treat slowdown as an unsolved research problem rather than a vague aspiration, especially as fears grow over AI systems being used to build even more powerful successors.
wired.com
· 2026-09-18
Bend is a newly introduced programming language that compiles to native code, runs on multiple CPU cores or GPUs without manual threading, and includes a built-in proof checker similar to Lean and Rocq. Developers can write rules in a LAWS.bend file, and Bend will refuse to accept any AI-generated code change that violates those rules, checking each change in about a second. Its creators demonstrate this by showing an AI agent attempt to break a game's win condition, which Bend blocks until the AI produces a compliant fix.
bend-lang.com
· 2026-09-17
A developer released an experimental utility called NvpwrControl on GitHub that lets owners of NVIDIA RTX 50-series gaming laptops manually raise their GPU's power limit beyond factory settings. According to VideoCardz, the tool can push the RTX 5090 Laptop GPU's ceiling up to 225W, roughly 40W above its stock maximum with Dynamic Boost.
tomshardware.com
· 2026-09-16
GoKawiil