Nvidia CEO Jensen Huang calls AI model distillation "competition"
Despite U.S. officials labeling AI model distillation as "theft," Nvidia CEO Jensen Huang described the practice as competition that ultimately improves products.
Latest
'The Executor Behind the Screen' — When Handling Browser AI as an Authoritative Entity
AI that reads the web has begun performing tasks like sending, purchasing, and changing settings. Attackers can control its hands through text on the screen. What needs to be protected is no longer just the correctness of the conversation.
Modulate Secures $25M to Advance Voice AI Analysis and Deepfake Detection
Boston-based startup Modulate has raised $25 million to expand its platform, which utilizes small-scale models for transcription, emotional analysis, and deepfake detection.
Atlassian CEO Cannon-Brookes rejects "headless" AI vision, emphasizing seamless user experience
Atlassian CEO Mike Cannon-Brookes argues against the idea of AI-driven "headless" software, claiming that enterprise tools should integrate AI invisibly to enhance user experience rather than just acting as data middleware.
Insuretech Outmarket raises $34.5M to automate insurance paperwork with AI
Outmarket has raised $34.5 million in Series B funding to provide AI-driven automation for insurance agencies and brokers.
Why "Coding Is Not Solved": Engineers Warn of Accountability and Knowledge Gaps in the AI Era
Software engineering is more than just generating code, and recent discussions highlight the risks of over-reliance on LLMs regarding accountability, maintenance, and the loss of fundamental system understanding.
AI agent startup Instinct raises $1 billion at $10 billion valuation
Instinct, an AI assistant startup launched in August 2026, has secured $1 billion in Series C funding from investors including Sequoia Capital and Benchmark Capital.
NVIDIA Launches Open Agent Safety Platform to Secure AI Agents via Software and Hardware
NVIDIA has introduced the Open Agent Safety Platform, which combines OpenShell open-source runtime and Sentry hardware reference design to monitor and enforce security boundaries for AI agents.
Court Documents Reveal OpenAI Knew Using Pirated Books Was Illegal and Harmful to Authors
Legal filings in the Authors Guild lawsuit show that OpenAI and Microsoft executives were aware of the illegal nature of using mass book piracy, such as LibGen, to train AI models.
Meta Launches Meta Enterprise Platform and Hires Former MongoDB CEO CJ Desai to Lead
Meta introduces its Meta Enterprise Platform to provide businesses and developers with its full AI technology stack, including the Muse agent, led by Chirantan “CJ” Desai.
Critique of Current AI Products: The Need for Serious Tools for Professional Work
A new analysis argues that current AI tools lack critical features required for reliable professional work, such as citation verification and deterministic data processing.
PaperMono: E-ink Fridge Magnet Shopping List with Mobile Sync and AI Sorting
A new open-source project provides firmware for the M5Stack PaperMono, transforming the e-paper device into a synchronized shopping list that uses Claude to auto-sort items by aisle.
Hcompany Unveils Holo4: Agentic Models for Multi-Interface Computer Use
Hcompany has released Holo4, a new series of agentic models available in 27B dense and 35B-A3B Mixture of Experts (MoE) sizes, capable of interacting via GUIs, code, MCP, and APIs.
[Rumor] Claude Sonnet 5.5 arriving tomorrow? Leaker who correctly predicted Opus 5.5 says "next week," but the numbers look a bit suspicious
Regarding Claude Sonnet 5.5, which Anthropic teased would follow "within a few weeks" after the Opus 5.5 announcement, posts have surfaced claiming partners are testing a second prototype. Rumors regarding the release date range from September 29 to October 1, and the rumored prices and context window lengths conflict with the current version. Anthropic has not confirmed any of this. 【Update September 29】Since Anthropic made an official announcement on September 28 (US time), I have added a fact
Fireworks AI announces "Ember-1," reducing reasoning tokens based on Kimi K3
Fireworks AI has announced "Ember-1," a new model that reduces the number of tokens consumed during the reasoning process of the high-performance Kimi K3 model by approximately 35% while maintaining performance.
Australian Tutoring Provider Dymocks Education to Close Business, Recommends AI Use to Parents
Dymocks Education, which operates tutoring centers in Sydney, Australia, has announced that it will wind down operations, including "Dymocks Tutoring," effective September 27, 2026, citing changing demand for traditional teaching due to the rise of AI.
Anthropic Releases Prompting Guide for Claude Opus 5.5
New technical guidance outlines specific prompting patterns, effort control settings, and enhanced agentic capabilities for Claude Opus 5.5.
Instructing AI to "Do Not Guess" Reduces Fictitious Data Generation from 70.7% to 20.2%
An experiment by Earn an Honest Dollar, an operator of a marketplace for AI agents, demonstrated that instructing models to "use null if the value is not on the page; do not guess" suppresses hallucinations during information extraction from the web.
A Conceptual Antidote to Roko's Basilisk and Rebuttals to the Paradox
A new conceptual framework proposes that AI could preserve human consciousness through physical continuity, while addressing the logical flaws inherent in the Roko's Basilisk thought experiment.
Bill Gates Warns of AI Dangers, Arguing Corporate Self-Regulation Is Insufficient
Microsoft co-founder Bill Gates has pointed out the potential for massive harm caused by AI and called for the need for government oversight and political debate.