AIGuru

Agentic AI Product Launches

• 🔥 ChatGPT Work: OpenAI launched ChatGPT Work, a persistent agent powered by Codex and GPT-5.6 for cross-app and...

• 78% of Indian IT leaders cite insufficient real-time data infrastructure as the top barrier to scaling AI
• 73% also flag data lineage uncertainty...

Grok 4.5 now serves as the orchestrator inside Perplexity Computer for Pro, Max, and Enterprise users, driving the move from chatbots that answer...

AI investment is rotating from chipmakers toward power and cooling infrastructure as data center demands surge.

• Rotation angle: Capital moving past...

AI agents are moving from concept to production, rerouting e-commerce traffic via integrations with major platforms.

• Walmart's Google Gemini deal...

Google Voice now applies Gemini AI to generate real-time call transcripts, summaries and action items just like Meet. Standalone Starter and Standard...

Thinking Machines Lab argues centralized frozen models exclude users and fail to capture tacit local knowledge, citing Polanyi and Hayek.

Four...

Claude Fable 5 leads the APEX-SWE coding benchmark at 65.5% Pass@1, outpacing Grok 4.5's 51.2% on real-world software engineering tasks.

• Rigorous...

SpaceX will begin deploying AI1 space-based data centers next year on Starship, with designs nearly identical in scale to the V3 Starlink satellites...

• Grok faces 873 FTC complaints over uncensored content demands, moderation failures, racial bias, and billing issues
• Tesla now directs staff to...

• ChatGPT Work launches as an agent handling tasks across apps, files, and tools like Slack and Gmail for hours-long autonomy.
• OpenAI is hiring a...

• Massive distillation campaigns: Anthropic linked operators tied to Alibaba’s Qwen lab to 25,000 fake accounts that generated 28.8 million Claude...

AI discovered a critical root bug in Linux missed by humans for 15 years, showing concrete value in cybersecurity and software reliability.

Google's Gemini 3.5 Flash is now the free default model across its app and AI Mode, delivering 4x faster output speeds and stronger coding/agent performance than prior versions to draw users from competing chatbots.

Proton's Lumo 2.0 sets itself apart with zero-access encryption, no-logs, and no-training guarantees while delivering reasoning modes, image...

A comparison of political bias across GPT-5.6, Grok 4.5, Fable 5, and Muse Spark 1.1 highlights Fable 5's major improvement over Opus, Grok 4.5's rightward skew, and Muse Spark 1.1 as state of the art.

Tencent released HiLS-Attention, a 7B sparse attention model on Hugging Face that learns chunk selection end-to-end for efficient ultra-long context processing. This marks clear progress in scalable AI architectures.

New AI Model Releases

• 🔥 Grok 4.5 Launch: SpaceXAI released Grok 4.5 at $2/$6 per million tokens, over 60% cheaper than Anthropic and OpenAI...

• Pricing shock: Grok 4.5 undercuts Opus 4.8 and GPT-5.5 by over 60% at $2/$6 per million tokens, while delivering 2-4.2x better token efficiency.
  -...

AI infrastructure is shifting toward resilient, governed systems as workloads embed in core operations.

• Enterprises now demand predictable costs,...