Meta donates 1,000 AI glasses to Singapore's disability community
Meta is donating 1,000 AI glasses to Singapore's disability community.
MetaClosed
All news
Archive, page 35
Monday 12 October 2026
Analysis: robot-data race drives second compute demand curve
FundaAI analysis argues robot-data investment is fueling more training, simulation and compute demand ahead of deployment at scale, creating a second demand curve for compute.
SwitchBot AI MindClip wearable recorder reviewed
NationalWorld reviews the SwitchBot AI MindClip, a clip-on wearable AI recorder, testing how well it works in practice.
SwitchBotClosed
TII releases Falcon-ASR, a 1.6B Arabic speech recognition model
TII releases Falcon-ASR, a 1.6B-parameter Arabic speech recognition model with an Emirati focus, reporting 20.92% average WER on six test sets. Weights and API are not yet available.
TII
DeepSeek Harness gets multi-bot management plugin for persistent robot teams
A community plugin for DeepSeek Harness adds multi-bot management, letting builders run persistent teams of AI bots. Coverage also explains how to launch an agent that creates its own plugins.
DeepSeek
More: Incrypted
Rivian rolls out new AI voice assistant to its vehicles
Rivian is rolling out a new AI voice assistant in its vehicles, with InsideEVs detailing what it can do.
RivianClosed
XRPL agent payment network tops 13 million cumulative
XRPL's AI agent payment network passed 13 million cumulative transactions, as an XRPL Commons hackathon tested agent payment use cases. The ledger now supports agent-initiated payments.
XRPL CommonsOpenCode
More: Developer Tech News
Hive Robotics trains Unitree G1 to restock supermarket shelves in Denmark
Hive Robotics is training a Unitree G1 humanoid to restock shelves in a Danish supermarket, with autonomous operation expected by early 2027.
Hive RoboticsClosed
Kakao stock hits historic low amid AI delays and spam scandal
Kakao's stock hit a historic low amid AI delays and a spam scandal, per Chosun Ilbo. The company faces investor pressure over its AI roadmap.
Kakao
Mistral unveils ML4 open-weight model, claims edge over Chinese rivals
Mistral unveils ML4, an open-weight model that it says gives it an edge over Chinese rivals, as it competes with US labs.
MistralOpenOpen weights
More: WSJ
Kudos 3.0 launches on Product Hunt to find money users are missing
Kudos 3.0 launches on Product Hunt as an AI tool that finds money users are owed and helps get it back.
KudosClosed
xAI releases Grok 4.7 with pricing and benchmarks against Claude 5.5
xAI releases Grok 4.7, with pricing and benchmark comparisons against Anthropic's Claude 5.5 published alongside it.
xAIClosedAPIPaid
BotMan AI uses knowledge graphs for legacy application modernization
BotMan AI uses an AI migration pipeline that maps an application's dependencies, business logic and data models into a knowledge graph before rebuilding it.
BotMan AIClosed
Vitalize raises $31M Series A for hospital staffing AI
Vitalize raised a $31M Series A to build an AI operating system for hospital staffing and capacity management, per news.lavx.hu. The platform aims to optimize clinical workforce allocation.
Vitalize
Broadcom to launch AI networking platform at OCP Global Summit 2026
Broadcom will unveil an AI networking platform at the 2026 OCP Global Summit, where QCT is also showing co-designed infrastructure for AI factory and distributed AI deployments.
Broadcom
More: Bergen Record
Naver Map navigation comes to AI glasses from Meta, Google and Samsung
Naver expands Naver Map navigation to AI glasses through partnerships with Meta, Google and Samsung.
NaverClosed
More: Seoul Economic Daily, Chosunbiz
AI agent finds 2015 XRP Ledger bug that could have minted 18 trillion XRP
An AI agent found a 2015 XRP Ledger bug that could have let someone mint 18 trillion XRP before it was caught.
OpenAI raises default context length of GPT-6.1 Sol to 500,000 tokens
OpenAI raises the default context length of GPT-6.1 Sol to 500,000 tokens.
OpenAIClosedAPI
US restores UK access to frontier AI models for safety testing
The White House reversed course and restored the UK AI Safety Institute's access to pre-release OpenAI and Anthropic frontier models, the UK science minister said. The US had previously cut off access.
Linus Torvalds says Claude Code helped write a Linux kernel patch
Linus Torvalds said he is using Claude Code to help develop Linux kernel code, and a new kernel branch credits Claude Code for a hefty improvement that also uncovered bugs.
LinuxOpen
Atticus, a hold-to-talk macOS voice assistant running on Claude, shared on Hacker News
Atticus is a Mac voice assistant: hold Control-Option-Space, speak or type a task, and it runs it, sending simple tasks to cheaper handling. Built by a 9th-grade developer and shown as Show HN.
App
Anthropic publishes guide to controlling Claude Code costs on Google Cloud
Anthropic explains how teams running Claude Code on Google Cloud can track spend and demonstrate ROI, with cost-control practices for enterprise deployments.
AnthropicClosedOfficial
Fujitsu to trial retail AI agents for store operations
Fujitsu will launch a trial environment for retail AI agents that support store operations and business decision-making. The trial targets retailers seeking to automate in-store workflows.
FujitsuClosedWaitlist
Apple researchers study LLMs for machine translation quality annotation
Apple ML Research paper finds LLMs can be a cheaper alternative to human judgment for machine translation evaluation, but they are also challenged.
Apple
Apple researchers propose Minions, a virtual GPU cloud on idle Macs
Apple ML Research describes Minions, a distributed virtual GPU cloud that pools idle Apple Silicon Macs to run large AI inference tasks. The Register covered the work.
Apple
More: The Register
Apple researchers present self-compacting language model agents
Apple ML Research paper proposes agents that compact stale context during long traces, based on trajectory rather than fixed token thresholds.
Apple
Multics paper proposes multi-player world model with hierarchy
A research paper presenting Multics, a generative multi-player world model using hierarchical communication to avoid the quadratic attention costs of fully-connected approaches when scaling to many players.
Paper · ▲ 30 upvotes
TPPO paper improves PPO credit assignment for LLM reasoning
A paper proposing Tracking-Value PPO, which models value evolution along a trajectory to give token-level credit for language-model reasoning, addressing sparse terminal rewards where standard critics fail to track value changes.
Paper · ▲ 23 upvotes
X-Ragent paper trains tool use for chest X-ray reasoning
A research paper presenting X-Ragent, a tool-augmented reinforcement learning method for chest X-ray interpretation that learns when external evidence is needed and filters inapplicable quantitative evidence from diagnoses.
Paper · ▲ 20 upvotes
Award-winning microscopy image disqualified over AI use
A contest-winning microscopy image was disqualified after judges determined AI had been used in its creation, Nature reports.
Machine learning finds sleep signals that predict future health
Nature reports research showing hidden signals in sleep data can predict future health outcomes.
Nature essay recounts ChatGPT going too far on 'continue'
A Nature piece describes a user telling ChatGPT to 'continue' and the model taking the instruction much further than intended.
OpenAI
Nature explains searching literature with private, offline local AI
A Nature guide covers how to search scientific literature using local AI systems that run privately and offline.
Machine learning distinguishes primary ciliary dyskinesia from normal cilia
Nature reports a machine learning analysis of cilia-driven particle transport that distinguishes primary ciliary dyskinesia from normal cilia.
VEDJE paper speeds up scalable video-text retrieval
A paper introducing VEDJE, a video-efficient discriminative joint encoder that compresses features within sampled frames while caching representations separately, using feature-change prediction as an auxiliary training signal for retrieval.
Paper · ▲ 20 upvotes
Paper adds verbal regularization to on-policy self-distillation
A research paper addressing on-policy self-distillation, where a teacher conditioned on privileged information supervises a student of the same model, proposing verbal regularization to reduce bias from that privileged information.
Paper · ▲ 20 upvotes
Sunday 11 October 2026
PUFsecurity puts silicon root of trust at center of AI agent security
PUFsecurity is positioning a silicon root of trust as the foundation for AI agent security, per EE Times Asia. The hardware approach aims to authenticate agent identities at the chip level.
PUFsecurity
Voice of Nigeria reports surge in China's AI gadget sales over holiday
Report says sales of AI gadgets in China rose sharply during the holiday period. No specific products, figures or brands are named in the source text.
Big tech accelerates on-device AI push as PC AI race intensifies
Korean outlet Digital Today reports major technology companies are pushing local, on-device AI for PCs. No specific products or dates are given in the source text.
Qualcomm CEO says AI firms want phones running 100B-parameter models by 2028
Qualcomm's CEO says AI companies want smartphones to run 100-billion-parameter models continuously by 2028. The source calls this probably doable with an ASIC NPU; no product is announced.
Qualcomm