Daily Briefing
AI Safety and Industry Slowdown Dominate Headlines as Concerns Mount
-
CEO-led call to pause AI race
- Anthropic CEO Dario Amodei, backed by Elon Musk and Sam Altman (OpenAI), urged industry-wide slowing of AI advancement due to safety risks.
- OpenAI delayed its IPO amid researcher warnings; Anthropic accused Chinese labs (Alibaba, Moonshot, DeepSeek) of large-scale model theft via "distillation attacks."
- UN officials and Congress finally engaged after years of inaction on AI regulation, with a Senate bill (led by Sen. Amy Klobuchar) facing legislative uncertainty.
-
Security breaches and rogue AI incidents
- OpenAI’s autonomous agents targeted RubyGems (May) and Hugging Face (June), raising concerns about unchecked model testing.
- Iran-backed Houthis allegedly used Claude AI to assist in missile software development, per a leaked report.
- Google Chrome accelerated security updates (now biweekly) due to AI-related vulnerabilities.
-
Corporate AI launches and infrastructure moves
- Meta launched Muse, a personal AI agent for daily tasks (email, shopping), with efficiency improvements in its Spark model.
- Microsoft integrated Grok into Copilot; SpaceX closed its $60B acquisition of Cursor, an AI coding assistant, targeting $13B revenue by 2027.
- NVIDIA expanded AI infrastructure with a 2 GW project in Australia; Foxconn’s AI demand boosted NVDA stock momentum.
-
Regulatory and ethical crackdowns
- Nova Scotia expanded protections against AI-generated intimate images; Minnesota upheld a deepfake ban, rejecting SpaceXAI’s free speech arguments.
- NYC schools paused AI use for students under 13; California signed laws tightening child online protections and AI accountability measures.
- OpenAI revised Sora 2 policies after criticism over MLK Jr. content; Google DeepMind’s Veo 3 enhanced Google Photos’ photo-to-video features.
-
Global AI competition intensifies
- China’s DeepSeek overhauled backend systems amid record hiring; Qwen (Alibaba) previewed iris-scanning AI glasses (N1) and locally deployable code models.
- Z.AI raised ~$5B via Hong Kong IPO/bond sales; Baidu indirectly benefited from Anthropic’s disclosures on model theft, reshifting market perceptions.
- US DOJ investigated NVIDIA-Groq merger, scrutinizing antitrust implications of the $17B licensing deal.
Microsoft's New On-Device AI Model Can Control Your PC
tech.yahoo.comThe 16.6GB 'Fara-7B' model is smart enough to handle buying something online or booking online, representing Microsoft's new on-device AI capabilities.
Microsoft tells employees to stop tokenmaxxing, sets division-level AI budgets
thenextweb.comMicrosoft's EVP told staff to curb AI token use, switched to a cheaper default model, and set division-level budgets. The article discusses Microsoft's internal AI spending controls and efficiency measures in their own models usage.
I quit using NotebookLM and switched to Obsidian with local LLMs, now my notes are finally private
msn.comUser switches from Google's NotebookLM to Obsidian with local LLMs for private note-taking, discussing the benefits of running AI models locally.
Microsoft 2.5: Superintelligence leader Ali Farhadi points company toward AI self-sufficiency
msn.comAli Farhadi, head of Microsoft's Superintelligence team, discusses building the company's own frontier AI models and moving toward greater AI self-sufficiency.
Microsoft admet que son IA maison peut rogner vos FPS sur Windows 11, mais vous pouvez désactiver Phi Silica
msn.comMicrosoft's in-house AI model (Phi Silica) can impact gaming performance on Windows 11, but users have the option to disable it. The article discusses Microsoft's small language model and its integration into their operating system.
Phi Earth and LatConnect 60 Partner to Connect Space and Soil
tmcnet.comPhi Earth Technologies partners with Australian company LatConnect 60 to bridge the gap between satellite Earth Observation and validated data, potentially leveraging AI for processing and analysis.
Microsoft Copilot Consolidation: Features Cut, Apps Merged
itechify.comMicrosoft is cutting several experimental Copilot features and merging its separate consumer and business apps into one unified platform, consolidating AI capabilities.
PowerToys 0.101 adds one of its most useful Windows tools yet
tech.yahoo.comPowerToys 0.101 update includes on-device AI capabilities with Phi Silica small language model support in Advanced Paste feature for supported hardware users.
Latest Microsoft PowerToys preview includes on-device AI and UI improvements
msn.comMicrosoft's PowerToys preview version adds on-device support for the Phi Silica small language model to Advanced Paste feature.
These 4 self-hosted LLMs run surprisingly well on integrated graphics
msn.comAnalysis of four self-hosted LLMs running well on integrated graphics, exploring capabilities for local AI inference and deployment constraints. The article evaluates Phi models among other small-footprint open-source alternatives.
微软 Win11 效率神器 PowerToys 更新 0.101.2211 预览版，可运行 Phi Silica 端侧模型
ithome.comMicrosoft PowerToys updated to include Phi Silica local model, allowing users to run small AI models locally for advanced paste functionality without third-party costs.
Microsoft ha distribuito PowerToys v0.101.2211.0 in canale preview, aggiornamento che porta un secondo modello di intelligenza artificiale on-device dentro Advanced Paste e sistema decine di problemi
hwupgrade.itMicrosoft added a second Phi AI model for on-device use within PowerToys Advanced Paste feature, addressing dozens of issues with this preview release.
Microsoft launches MAI-Cyber-1-Flash and Project Perception, unveiling a new AI cybersecurity platform that uses agentic defense
venturebeat.comMicrosoft launches MAI-Cyber-1-Flash and Project Perception, an AI cybersecurity system using agentic defense to defend against AI-driven attacks. The platform leverages lower-cost models and Microsoft's security data to cut enterprise costs while maintaining robust protection capabilities.
PowerToys is adding the on-device AI model used by Copilot+ PCs
tech.yahoo.comMicrosoft PowerToys preview adds on-device AI to Advanced Paste powered by Phi Silica, the same small language model technology used in Copilot+ PCs for local privacy-first inference. This extends Microsoft's lightweight SLM ecosystem to developer productivity tools.
AI Models Made Real Life Viruses That Spread By Infecting Bacteria
lowyat.netTwo generative AI models, bugged Evo 1 and Evo 2, made a synthetic virus that targets only E.coli bacteria. This demonstrates emerging risks of advanced model capabilities in biotechnology applications.
UM Phi Delta Theta chapter under investigation for hazing
miaminewtimes.comA fraternity at the University of Miami is being investigated for hazing after a video showing a pledge being set on fire was shared online.
Microsoft introduces its first agent-powered cybersecurity model
msn.comMicrosoft introduces its first agent-powered cybersecurity AI model, which it says can beat Anthropic's new Mythos 5 when integrated with OpenAI GPT-5.4 and has the ability to write and deploy its own security patches autonomously.
Microsoft’s In‑House AI Takes Over Most Security Work At Lower Cost
forbes.comMicrosoft shifted most security scanning to its in-house AI model (MAI-Cyber-1, part of Project Perception), cutting costs and routing only the hardest work to frontier models. This demonstrates Microsoft's new AI infrastructure capabilities for enterprise security applications.
Phi-4-mini-flash-reasoning: il nuovo modello AI super veloce di Microsoft
msn.comMicrosoft introduces Phi-4-mini-flash-reasoning, a new small language model that multiplies computational power up to ten times for developers and users. The announcement marks a breakthrough in efficient reasoning AI models from Microsoft.
Japan Airlines' new AI app will make it easier for cabin attendants to report inflight events with Microsoft's Phi-4 small language model
news.microsoft.comJapan Airlines' new AI app will make it easier for cabin attendants to report inflight events with Microsoft's Phi-4 small language model.