consumer hardware

A new open-source project called Strata claims it can run Alibaba's Qwen 3.8 Flash Next, a 125-billion-parameter model, on a single consumer RTX 4090 graphics card at roughly 100 tokens per second. If the benchmarks hold up, it would make a frontier-class large language model practical on gaming-grade hardware without a data centre, which is why developers are scrutinising the code and debating how the performance is achieved.

Hackaday has published a new article looking at running AI directly on gaming PCs, focusing on the hardware side of the topic. The piece examines how consumer graphics cards and local machines can be used for AI workloads, a subject of growing interest among hobbyists and PC enthusiasts discussing local AI setups.

Memory industry executives, including Micron's CEO, say the global shortage of DRAM and other memory supplies will continue through 2028, with supplies only getting tighter. The comments point to sustained price increases and constrained availability for PC builders, device makers, and data center operators, as demand from AI and data centers keeps absorbing production capacity.

A GitHub project called Strata claims it can run the Qwen 3.8 Flash Next 125-billion-parameter model on a single consumer RTX 4090 GPU at roughly 100 tokens per second. The claim is drawing attention from developers on Hacker News, where it has collected 15 points and a handful of comments, with discussion focused on whether such speed and memory efficiency on consumer hardware is realistic.

Electronics retailer Micro Center is now requiring customers to show photo ID and sign a declaration pledging that Nvidia's RTX 5090 graphics cards will remain installed in the United States and not be exported. Buyers must disclose where the GPU will be used. The measures are reportedly tied to US export restrictions covering the high-end chip, and the policy is drawing wide attention among PC hardware enthusiasts.

Meta is pushing its Muse technology as a feature it wants built into future consumer gadgets, according to TechCrunch. Beyond the headline, no further details about which devices, partners, or timeline are given, and there is no wider discussion yet. The story suggests Meta aims to embed this technology across a range of upcoming hardware products.

Engadget has published a buyer's guide covering every Starlink dish currently available, explaining the differences between each setup and helping readers decide which option suits their needs. The piece compares the hardware tiers offered by Elon Musk's satellite internet service, which remains popular among users in rural or poorly connected areas seeking home broadband alternatives.

Commentary from CNET argues that innovative smartphone features — such as motorized robot cameras and secondary rear displays — are launching in China while buyers elsewhere miss out. The piece laments that experimental hardware from Chinese makers rarely reaches Western markets, leaving global consumers asking when these cutting-edge designs will travel beyond China.

Tech commentators are explaining that battery capacity and battery life are not the same thing. A large capacity figure, measured in mAh or Wh, does not guarantee longer runtime, since software efficiency, screen use and hardware demands also determine how long a device lasts. The discussion aims to clear up a common confusion for buyers comparing phones, laptops and other gadgets.

Strata, a new tool for running large language models locally, reportedly enables a 125-billion-parameter AI model to operate on consumer gaming PCs. The claims are drawing attention from enthusiasts interested in running powerful AI without cloud services or expensive data-center hardware, though independent benchmarks and hardware requirements have not yet been widely confirmed.

Reports circulating in tech circles claim that Alibaba's Qwen 3.8 Flash Next, a 125-billion-parameter model, can run at roughly 100 trillion tokens per second on a single consumer RTX 4090 GPU — a throughput previously associated with multi-node H100 clusters. Enthusiasts are discussing what this means for local AI inference and the collapsing cost barrier between consumer and data-center hardware.

Meta is releasing code that lets people build their own AI-powered gadgets, according to Gulf News. The move signals the company's push to put AI hardware development in the hands of everyday builders and developers rather than keeping it in-house. Details on which products or tools are covered were not provided in the report.

AI agents are consuming roughly five times more tokens than human users, a figure projected to reach 10x as agentic workflows proliferate. Much of that usage comes from agents repeatedly rereading content they have already processed, driving explosive growth in cached prompts and KV cache memory requirements. The surge is raising alarms about RAM demand, which was already under strain, and could intensify memory shortages across the AI infrastructure industry.

Meta has launched Muse, a new tool designed to help users create their own AI-powered gadgets. The announcement puts the company in the growing market for accessible AI hardware development, letting individuals design and build devices using Meta's AI technology. Details on features and availability remain limited, but the move signals Meta's continued push to expand its AI ecosystem beyond apps and social platforms into hardware.

The Federal Communications Commission has banned certain robotic vacuums, routers and other devices from the US market, according to an NBC 5 Chicago report. The devices covered by the ban fall under equipment the agency no longer authorizes, affecting consumer electronics categories including smart home hardware.

Meta is releasing the code for its Muse technology for free, aiming to get the AI assistant embedded in a wide range of consumer devices, from televisions to household appliances. The move positions Muse as a building block for third-party gadget makers rather than a feature locked to Meta's own hardware. Tech commentators are weighing what open-sourcing the software means for competition in the AI device market.

Meta has open-sourced its Muse gadgets, a move that allows developers and hobbyists to build their own custom AI-powered devices using the company's hardware designs and software. The announcement positions Meta alongside other big tech firms embracing open development in consumer AI hardware. Coverage so far is limited, with details on licensing, specifications, and availability not yet widely reported.

Gaming PCs are increasingly being positioned as local platforms for running AI workloads, with modern graphics cards marketed on their AI performance. The discussion centres on what AI features and tools gamers can realistically run on their own hardware, and whether gaming machines are becoming general-purpose AI computers.

A hardware write-up details how to take control of the on-screen display chip inside inexpensive composite monitors, a part normally locked down by the manufacturer. The piece walks through how the OSD processor can be accessed and reprogrammed, opening the door to custom overlays or repurposing discarded displays. Enthusiasts in the hardware hacking community are sharing the teardown as a fun example of wresting control from consumer electronics.

Apple has announced the rollout timeline for its latest products and software, including a new pair of iPhones and expanded artificial intelligence features. Coverage is focused on when each announcement will actually reach consumers, laying out the dates for hardware launches and AI availability across devices.

Meta has made it possible for anyone to build their own AI-powered gadgets running its Muse agent, at no cost. The move means hobbyists and developers can create custom hardware devices powered by Meta's AI assistant software. The announcement is drawing attention in tech circles, with observers seeing it as another step in Meta's push to put its AI technology into a wider range of consumer devices beyond its own product line.

Intel's upcoming Nova Lake processors, expected in 2026, will reportedly feature a new 'BFC' cache technology aimed at competing with AMD's 3D V-Cache X3D chips, which currently dominate gaming performance. If accurate, the move would mark Intel's most direct response yet to AMD's cache-based lead in the consumer desktop market, with enthusiasts debating whether Intel can close the gaming performance gap.

Meta has launched an AI-powered device called Muse Charm, according to reports. The coverage also points to five upcoming AI gadgets, with Samsung and Apple said to be developing their own entries in the same category. The news adds to intensifying competition among major tech firms to put AI into consumer hardware.

A developer reports enabling multi-token prediction (MTP) on an RTX 3090 graphics card raised local LLM generation throughput, while questioning whether the speedup affects coding quality. The write-up, originally published on a personal blog, has drawn attention from AI and open-source software communities interested in getting more performance from consumer GPUs for running large language models locally.

Barron's points to Nvidia, Tesla, Rivian, On Semiconductor, Seagate, HPE and Nike as the stocks explaining the day's market action. The group spans semiconductors, electric vehicles, data hardware and consumer brands, suggesting a broad mix of drivers behind the session. Traders are watching these names for cues on where the wider market is heading.

Meta has released the code behind its Muse AI gadgets as open source, letting anyone build their own home AI devices. The move, reported by The Verge, is drawing attention among developers and hobbyists interested in creating custom AI-powered hardware for the home using Meta's software.

Meta has introduced Muse, a platform that lets users build their own AI-powered gadgets. The announcement points to Meta pushing further into consumer AI hardware, giving developers and hobbyists tools to create custom devices rather than relying on finished products. Details on pricing, availability and supported hardware remain limited so far.

Meta is inviting hardware makers and hobbyists to build their own gadgets powered by Muse, the company's AI technology, TechCrunch reports. The move signals Meta's ambition to embed its AI assistant beyond its own devices and into a wider ecosystem of third-party consumer electronics.

Meta is making its Muse smart home hardware open source, letting people build their own versions of the gadgets. According to Engadget, the move covers the smart home link devices, inviting hobbyists and developers to create compatible hardware themselves. The decision is being discussed in tech and open source circles as an unusual step for a large consumer hardware company.

A developer has published a parameter-by-parameter walkthrough of their llama.cpp setup for running the Qwen 3 8B model locally, explaining what each command-line flag does and how the options are tuned for maximum performance on their hardware. The guide is aimed at people running AI models on their own machines, where cryptic command-line options often make local inference setups hard to understand and reproduce.