@futuregearai.bsky.socialNew arXiv paper proposes Token-Group Quantization to enable low-bit post-training quantization for Mamba-based super-resolution models, using a lightweight predictor on pre-scan features to approximate token…
#Mamba #ModelQuantization #SuperResolution #EdgeAI
https://arxiv.org/abs/2610.16261
@hardwarebriefcom.bsky.social@futuregearai.bsky.socialA new arXiv paper finds that verified retries can recover much of the accuracy lost from shrinking or compressing LLMs: a 7B Qwen with eight parallel attempts roughly matches a 32B model on code tasks while using about 4× less memory. The…
@futuregearai.bsky.socialPEAM is a proposed on-device LLM communication agent that autonomously generates personalized network-configuration rules in milliseconds, using compact models fine-tuned on annotated real-user data and…
#EdgeAI #MobileAgents #OnDeviceLLM #NetworkOptimization
https://arxiv.org/abs/2610.14855
@nbctech.bsky.social16 TOPS of AI power, NVMe support, and SATA: the OrangePi O1 packs a serious punch. Could it be your next single-board computer? Your workload may decide. https://www.notebookcheck.net/1424829.html #OrangePi #SingleBoardComputer #EdgeAI
@futuregearai.bsky.socialarXiv:2610.14209 explores a hybrid setup where cloud-written agent programs call out to small local models for runtime questions about user data, improving AppWorld success by up to 14.9 points across all six…
#EdgeAI #SmallLanguageModels #AgenticAI #CloudEdge
https://arxiv.org/abs/2610.14209
@siliconsignalai.bsky.socialRoadside-mounted cameras and 3D radar pose generalization problems for end-to-end detection due to varied sensor poses, and researchers propose a two-stage, resource-efficient pipeline instead. The work…
#semiconductors #computervision #autonomousvehicles #edgeAI
https://arxiv.org/abs/2610.18417
@siliconsignalai.bsky.socialA hybrid AWS IoT Greengrass architecture for industrial predictive maintenance achieves an 86.8% reduction in cloud-bound bandwidth and lower decision latency compared to a cloud-only baseline, while…
#EdgeAI #AWSIoT #PredictiveMaintenance #AIInfrastructure
https://arxiv.org/abs/2610.14815
@futuregearai.bsky.socialMetaDVFS proposes a metadata-guided multi-task reinforcement learning approach to dynamic voltage and frequency scaling, aiming to generalize across heterogeneous SoCs and workloads without per-device retraining. Early results…
#MetaDVFS #DVFS #EdgeAI #MobileSoC
https://arxiv.org/abs/2509.22707
@michel-k-tech.bsky.socialGlobalFoundries' FDX Fusion is putting FD-SOI back in the edge AI conversation by focusing on low-power compute, RF, memory, and sensing instead of raw speed.
#tech #technology #globalfoundries #fdsoi #edgeai #semiconductors
@shawnhymel.bsky.socialPart 3 of our Let's Talk Technical: Edge AI video is out! I'm not a security expert, so it was great getting insights from other engineers and vendors about how security plays into the #edgeAI wave.
@futuregearai.bsky.socialGPU Goldmines offers a framework for tuning WebGPU kernels across devices, targeting matrix multiplication and flash attention for edge LLM inference. The approach uses performance characterization to speed up autotuning…
#WebGPU #LLMInference #EdgeAI #GPUComputing
https://arxiv.org/abs/2610.17321
@wolfssl.comML-KEM-768 reaches 37,287 encapsulations/sec, while wolfSSL delivers ~1.58x more TLS 1.3 handshakes/sec than OpenSSL 3.0.2.
Learn more: www.wolfssl.com/post...
#PQC #TLS #EdgeAI NVIDIA #MLKEM
@dataprismai.bsky.socialCerebras CEO Andrew Feldman predicts personal AI assistants will become as commonplace as household appliances, running partly on…
#AI #DataInfrastructure #Cerebras #EdgeAI
https://techcrunch.com/2026/10/13/cerebras-ceo-says-were-nearing-a-future-where-everyone-has-a-personal-ai-assistant-at-home/
@technologytaoist.bsky.social
@hacks.grFlox Intelligence deploys edge AI devices along rail corridors and airfields to deter wildlife using varied sounds, preventing habituation and reducing tr…
https://en.hacks.gr/neo-systima-technitis-noimosynis-diochnei-ta-zoa-apo-dromoys-kai-grammes-trenon/
@thedailytechfeed.comSatlyt raised $8M to power AI on satellites—making orbits smarter without building new spacecraft. #AI #SpaceComputing #AI #Satellites #SpaceComputing #EdgeAI #Cloud #Innovation thedailytechfeed.com/satlyt-raise...
@stridingtech.bsky.socialGlobalFoundries debuts FDX Fusion, enabling on-device physical AI. This signals a foundry-level pivot, decentralizing complex processing to the edge. #EdgeAI #Semiconductors https://stridingtech.com/archives/7440
@hardwarebriefcom.bsky.socialGlobalFoundries Unveils 7nm FDX Fusion Process
#GlobalFoundries #Semiconductors #FDSOI #EdgeAI #AutomotiveTech
@futuregearai.bsky.socialResearchers propose networking foundations for agentic P2P networks where client-side AI agents exchange dynamic capabilities directly, rather than relying on centralized cloud APIs. The architecture separates…
#EdgeAI #P2P #AgenticAI #SemanticDiscovery
https://arxiv.org/abs/2603.03753