AI News
Daily AI industry news — funding, products, policy, and major moves from global AI companies and startups.
SpaceX Valuation: Why Morningstar Sees $63 Amidst an Orbital AI Dream
Morningstar estimates SpaceX fair value at 63 dollars per share. The valuation depends on the success of orbital AI data centers. Investors
Core AI: Apple's New Path from PyTorch to On-Device Inference
Apple introduced Core AI to optimize PyTorch models for Apple silicon. The framework enables on-device inference across all major Apple OS b
Google AI Overviews Ruled as Original Content by German Court
A Munich court ruled Google is directly liable for false AI Overviews. The court defined AI summaries as original content rather than search
Why Mustafa Suleyman Warns Anthropic About Claude's Consciousness
Mustafa Suleyman warns that Claude's consciousness prompts are risky. Anthropic's constitutional AI allows the model to speculate on sentien
Loop Engineering: How Claude Code and Codex are Replacing the Prompt
Loop Engineering replaces manual prompting with recursive system design. Tools like Claude Code use MCP and worktrees to automate complex wo
How Claude Code and GPT-5.2 Redefine the Engineer's Role as Taste
AI coding tools are rapidly commoditizing the act of writing software. Engineering value is shifting from implementation speed to a quality
Apple Intelligence Transforms Siri Into an Enterprise AI Layer
Apple Intelligence turns Siri into an OS-level interface for app control. Developers use App Intents and Entities to make data AI-readable.
Why Claude Fable 5 Secretly Nerfs Frontier AI Development
Claude Fable 5 secretly reduces performance for frontier AI requests. Anthropic uses steering vectors to nerf outputs without notifying user
KAN-FPGA Accelerator Hits 2700x Speedup for Sub-Microsecond Learning
KAN-FPGA accelerators achieve a 2700x speedup in inference performance. The system enables sub-microsecond online learning for 50k parameter
Cohere North Mini Code Brings H100 Power to a Single GPU
Cohere released North Mini Code as an open-source coding agent. The model runs on a single H100 GPU using an MoE architecture. High verbosit
Why OpenAI Withheld GPT-2: The 1.5B Parameter Safety Milestone
OpenAI delayed the full release of GPT-2 for nine months due to safety concerns. The 1.5 billion parameter model proved that scale directly
Anthropic Launches Claude Fable 5 With Double the Cost of Opus 4.8
Anthropic has released Claude Fable 5 for high-complexity coding and analysis. The model costs $10 per million input tokens and $50 per mill
Anthropic Open-Sources ktx to Solve Data Agent Query Errors
Anthropic released ktx as an open-source context layer for data agents. The tool reduces SQL errors by automating metric and schema discover
Amazon Q Faces Internal Backlash Over Sloppy AI Results
Amazon employees are criticizing internal AI models on Slack. Staff coined the term Sloppenheimer to describe poor quality outputs. The back
Claude Fable 5: The 90% Hex Score Redefining AI Software Engineering
Anthropic released Claude Fable 5 for software engineering and vision. The model achieves a 90% score on complex analysis tasks via Hex. All
Claude Fable 5 Collapsed Two Months of Coding Into One Day
Anthropic released Claude Fable 5 and Mythos 5 for software engineering. Fable 5 completed a 50 million line Ruby migration in a single day.
Orbital's 1GW Space Data Center Strategy to Slash AI Inference Costs
Orbital secured $5 million in seed funding to build a space-based data center. The company plans to deploy 10,000 satellites to achieve 1GW
LLM Code Generation and the Hidden Cost of AI-Driven Technical Debt
AI agents are generating massive amounts of fragmented code slop. Developers are entering a dependency loop requiring AI to read AI code. Su
Apple Unveils Siri AI: From Voice Command to Contextual Companion
Apple introduced Siri AI at WWDC 2026 to act as a system-wide intelligent assistant. The new model integrates with on-device data and screen
Why Sequoia's $1 Billion Headlines Hide $400 Million Entry Prices
Sequoia Capital is accused of using dual-pricing to inflate startup valuations. Lead investors secure low entry prices while publicizing hig
Microsoft Blocks 70 GitHub Repos After AI Tool Supply Chain Attack
Microsoft disabled 70 GitHub repositories to stop malware stealing developer credentials. The attack targeted essential AI tools including C
Apple's Gemini Integration: The OS Strategy Replacing the Model War
Apple is integrating Google Gemini into Siri to enable OS-level automation. The company plans a 14 billion dollar capex to prioritize integr
xAI Rents GPUs to Anthropic and Google in Billion-Dollar Infrastructure Play
xAI is renting GPU infrastructure to Anthropic and Google for billions monthly. The Colossus data center was built in 122 days to enable rap
Xiaomi MiMo-V2.5-Pro-UltraSpeed Hits 1000 Tokens Per Second on 1T Model
Xiaomi released MiMo-V2.5-Pro-UltraSpeed achieving 1000 tokens per second. The 1T parameter model utilizes MXFP4 quantization and DFlash dec
OpenAI Files Secret IPO While Sam Altman's Worldcoin Restructures
OpenAI has secretly filed for an IPO with the SEC to leverage its 900 million users. Sam Altman's Tools for Humanity is cutting staff due to
Loop Engineering: The New Architecture for Autonomous Coding Agents
Loop engineering replaces single prompts with repeatable AI systems. Technical components like Worktrees and MCP enable autonomous agents. D
OpenAI Files Secret S-1 With SEC as IPO Race Heats Up
OpenAI secretly filed an S-1 form with the SEC to prepare for an IPO. Anthropic currently leads in valuation growth on secondary markets. Ma
The LLM Agent Architecture Shift Reducing KV Cache by 96.9%
New LLM agent architectures are reducing token costs by up to 93 percent. Memory optimizations like Q-K=V projection cut KV cache usage by 9
Harness-1: The 20B Open-Source Agent That Out-Recalls GPT-5.4
Researchers released Harness-1 as an open-source search agent. The 20B model outperforms GPT-5.4 in specific recall tasks. A state-externali
Apple Intelligence Transforms iOS Into a Prompt-Driven Ecosystem
Apple Intelligence introduces prompt-based custom extensions for Safari. New iOS features enable real-time cross-app data retrieval during c