All Updates
Page 1861 of 1898
February 16, 2026
Bendigo Bank Cuts Dev Costs and Time
Bendigo Bank has reduced costs and time in software development efforts. Digital onboarding experience serves as their first productivity 'trophy'.
OpenClaw 2026.2.15 Release Ships
OpenClaw version 2026.2.15 is now available. It introduces sub-sub-agents and Discord Components v2. Additional security hardening is included.
OpenClaw 2026.2.15 Ships with Enhancements
OpenClaw version 2026.2.15 has been released. Key additions include sub-sub-agents and Discord Components v2. It also features another round of security hardening.
OpenClaw 2026.2.15 Ships with Agents and Security
OpenClaw has released version 2026.2.15. Key additions include sub-sub-agents and Discord Components v2. It also features another round of security hardening.
Starmer Targets AI Chatbots Like Grok
UK Prime Minister Starmer vows to battle AI chatbots as with Grok. New government plans ensure no online platform gets a free pass on children's internet safety.
GitHub AI Agent Ends Dev Chores, 10x Efficiency
GitHub launches AI agent eliminating 3-hour miscellaneous tasks for developers. Boosts efficiency by 10 times. Marks shift to intelligent collaboration platform.
GitHub AI Agent Boosts Efficiency 10x
GitHub launches an AI agent that eliminates 3 hours of miscellaneous developer tasks, increasing efficiency by 10 times. This represents a pivotal shift for GitHub from a code repository to an intelligent collaboration platform.
Info Bound on World Models in Optimal Policies
New ArXiv paper quantifies information optimal policies encode about environments. Proves mutual information of exactly n log m bits in Controlled Markov Processes. Bound holds for finite-horizon, discounted, and average reward maximization.
GT-HarmBench: Game Theory AI Safety Benchmark
GT-HarmBench introduces 2,009 high-stakes multi-agent scenarios using game theory like Prisoner's Dilemma to benchmark AI safety risks. Frontier models select socially beneficial actions only 62% of the time, often leading to harm. The benchmark, code, and analysis are available on GitHub.
EST Boosts Temporal KG Forecasting
Entity State Tuning (EST) introduces persistent entity states to TKG forecasters, overcoming stateless methods' long-term dependency issues. It uses a closed-loop design with topology-aware perception and dual-track evolution. Achieves state-of-the-art results across benchmarks with code on GitHub.
BrowseComp-V³ Benchmark for Multimodal Agents
BrowseComp-V³ is a new benchmark with 300 challenging questions for evaluating multimodal browsing agents on deep multi-hop reasoning across text and visuals. It features subgoal-driven process evaluation and publicly searchable evidence for reproducibility. Experiments reveal state-of-the-art models achieve only 36% accuracy, highlighting integration bottlenecks.
Adaptive Framework for Utility-Weighted AI Benchmarking
This paper introduces a theoretical framework that reimagines AI benchmarking as a multilayer, adaptive network connecting evaluation metrics, model components, and stakeholder priorities through weighted interactions. It embeds human tradeoffs using conjoint-derived utilities and a human-in-the-loop update rule, allowing benchmarks to evolve dynamically while maintaining stability. The approach generalizes traditional leaderboards and promotes context-aware, human-aligned evaluations.
OpenClaw Beta: Nested Agents & Security Boost
OpenClaw 2026.2.15-beta.1 introduces Discord Components v2 for rich interactions, nested sub-agents with depth controls, and platform-specific enhancements like per-channel ack reactions. It adds webhook toggles for cron jobs and deduplicates channel resolutions. Multiple security fixes harden sandboxing, redact tokens, and restrict dangerous configs.
OpenClaw Founder Joins OpenAI Team
Peter Steinberger, creator of viral AI agent OpenClaw, has joined OpenAI. OpenClaw will transition to an open, independent foundation.
US May Reverse China Tech Bans
The US appears open to reversing some technology bans on Chinese companies. This policy shift is highlighted in Asia In Brief alongside India's two-hour deepfake takedown demands and Singapore's AI adoption.
US Eyes Reversal of China Tech Bans
The US may reverse some technology export bans on Chinese firms. Related briefs include India's mandate for two-hour deepfake takedowns and Singapore's embrace of AI. Other notes cover a cuddly Japanese robot wolf.
US eyes reversing China tech bans
The US appears ready to ease some restrictions on Chinese technology companies. This policy shift is highlighted in Asia In Brief. Additional notes include India's two-hour deepfake takedown mandate and Singapore's AI adoption.
OpenClaw Beta: Discord v2 & Subagents
OpenClaw 2026.2.15-beta.1 adds Discord Components v2 for rich interactions, nested sub-agents, plugin hooks, and channel-specific overrides. Security fixes include SHA-256 hashing, token redaction, and Docker sandbox hardening. Additional improvements cover webhooks, gateways, and skills security.
OpenClaw 2026.2.15-beta.1 Beta Release
OpenClaw beta introduces Discord Components v2 for rich interactions, nested sub-agents, and plugin hooks. Additional enhancements cover per-channel reactions, webhook toggles, and channel deduplication. Multiple security fixes address hashing, token redaction, sandbox restrictions, and more.
Ant Open-Sources 1T Ling-2.5 Instant Model
Ant Group open-sourced Ling-2.5-1T, a trillion-parameter instant model with 63B active parameters, trained on 29T tokens. It upgrades architecture for 1M token context, boosts token efficiency, and enhances preference alignment and agent interactions. The model outperforms predecessors and rivals large instant models in reasoning and instruction following.