Smart Car Speed Record Broken: First Pure On-Device Large Model Mass-Produced in Just 10 Months
I watched a pure on-device LLM hit mass production in just 10 months, shattering speed records and reshaping the 2025 automotive AI landscape.
Browse AI news across models, agents, media, industry, and compute policy.
I watched a pure on-device LLM hit mass production in just 10 months, shattering speed records and reshaping the 2025 automotive AI landscape.
I read the release notes. The hype is real, but I worry about latency costs for routine tasks.
I see capital betting heavily on AI labs via chips and credits. Yet, ByteDance's new 200B model outperforms DeepSeek's 671B version, challenging scale assumptions.
I track how MediaTek’s MDDC 2025 pushes local agent AI adoption while Google Cloud Next defines global interoperability standards via Ironwood TPUs.
I watched Meta launch Llama 4, challenging DeepSeek's coding parity with fewer parameters. The release sparks debate over its evaluation integrity.
I read the 2nd China Embodied Intelligence Conference summary covering 2025-2026 extensions. In practice, lab demos don't fix on-call pain. This adds to our timeline tracking.
I read OPPO and HKUST(GZ)'s OThink-MR1. It targets multimodal generalization reasoning, a key industry focus for 2025–2026.
BAAI unveils a cross-embodiment brain-cerebellum framework for embodied AI. I read the release notes; it pushes swarm intelligence via open-source collaboration.
I read the East China Normal & Donghua University review on 2025-26 agent tech. My read: Frameworks clarify hype, but governance remains unaddressed.
I read about Google's new Gemini 2.5, which boosts reasoning and search integration. As a dev, I wonder if this actually speeds up coding workflows or just adds more complexity to the toolchain.