10 results on this page
A community test on an AMD R9700 with ROCm reports noticeable quality and long-context retention differences between FP16 and q8_0 KV cache for Qwen3.8-27B. FP16 reportedly produces more careful structured output and maintains performance beyond 120k tokens, challenging the assumption that both formats are equivalent.
VentureBeat has appointed Rob Strechay as its first Lead Analyst and founding analyst of VentureBeat Research. He will focus on enterprise AI infrastructure, platform engineering, DevOps, observability, and AI security as organizations move from experimentation to production.
Google’s Pixel 11 Pro and Pixel 11 Pro XL introduce highly capable voice typing alongside other software improvements. However, weak gaming performance and an ineffective rear LED prevent the phones from standing out as major flagship upgrades.
Apple announced changes to its EU App Store charging rules, prompting a rapid backlash from long-time rival Epic Games. Epic argues that the revisions remain inadequate and continue to impose excessive costs on developers.
Ant Group is shifting Alipay away from optimizing daily active users as its primary objective. It no longer requires every service invocation to return to the Alipay app, aiming to expand cross-platform reach and create more opportunities for service usage and transactions.
Tactile robotics is expanding from sensor hardware into datasets, foundation models, and evaluation benchmarks. Research shows touch improves physical reasoning and robot success, but inconsistent sensor formats create major cross-device transfer challenges.
Waymo has published the architecture of its robotaxi onboard computer, offering its first detailed look inside the vehicle system. The company also introduced a custom 5-nanometer chip and identified seven suppliers supporting the platform.
Uber has begun offering Baidu’s driverless Apollo Go vehicles through its app in Dubai. Riders booking UberX or Uber Comfort may be matched with an autonomous vehicle, while a dedicated Autonomous option increases the chance of requesting one.
Nvidia announced full-scale production of its Spectrum-X Ethernet Photonics co-packaged optical switch for AI factories, using 200G-per-lane links. SK hynix simultaneously published a CPO technology roadmap targeting over 100 Tb/s per node, sub-1 pJ/bit energy consumption, and sub-10 ns chip-to-chip latency.
Kuaishou’s Kling AI raised nearly $3 billion at a post-money valuation of $18 billion, with Kuaishou’s stake falling to 68.33% and a potential Hong Kong IPO targeted by 2031. Despite strong revenue growth, Kling faces slowing momentum, talent departures, and intense competition from ByteDance’s Seedance and open-source video models.