All articles

The complete WAKIB archive: every published story, across every section, newest first.

LatestAI Events

DataHack Summit 2026 Set for August in Bengaluru

DataHack Summit 2026 runs August 5-8, 2026, in Bengaluru, India.

LLMsAugust 5, 2026

Google's Gemma 4 lands on AWS GovCloud via Bedrock

Google DeepMind's Gemma 4 family is now available on Amazon Bedrock in AWS GovCloud (US-West)

AI AgentsAugust 5, 2026

Seedeo Bundles Kling, Veo and Seedance Into One AI Studio

Seedeo is a multimedia studio combining AI video, image, music, and voiceover generation in one interface.

AI PolicyAugust 4, 2026

Snapchat Cuts Fully AI-Generated Videos From Spotlight Recommendations

Snapchat will stop boosting fully AI-generated videos in Spotlight recommendations as of late July 2026, while AI-edited human content remains eligible.

LLMsAugust 4, 2026

Claude's Voice Mode Now Runs on Opus and Sonnet, Not Just Haiku

Anthropic upgraded Claude's voice mode on July 23 to run on Opus and Sonnet models, not just the original Haiku model.

Editorial image for Anthropic Upgrades Claude's Voice Mode With Opus and Sonnet, Adds Email-by-Voice

LLMsAugust 4, 2026

Apple Plots a Paid Tier for Siri's AI Muscle Ahead of iOS 27 Launch

Tim Cook said Apple will let iCloud Plus users pay to unlock higher AI usage limits, tied to the revamped Siri launching with iOS 27 this fall.

AI AgentsAugust 4, 2026

Nadella warns companies betting everything on one AI vendor

Microsoft CEO Satya Nadella warned on CNN that companies relying entirely on one AI lab risk not surviving long term.

EthicsAugust 4, 2026

OpenAI Extends Content Provenance Tools to Audio as It Formally Backs EU's AI Code

OpenAI is expanding content provenance tracking from images to audio, with text support planned next

LLMsAugust 4, 2026

Berkeley's ABBEL Tackles the Memory Loss of Summarized AI Context

Berkeley AI Research introduces ABBEL, a framework training LLMs to compress interaction history into natural-language 'belief states' without losing key information

LLMsAugust 4, 2026

What "Model Welfare" Means in Anthropic's Claude Evaluations

Anthropic runs 'model welfare' evaluations alongside capability tests on Claude releases, checking behavioral consistency under constraints and adversarial prompts.

LLMsAugust 4, 2026

Chat vs. Agent: How to Choose the Right AI Mode Today

Chat and agent modes are now distinct products across ChatGPT and Claude, not just one general-purpose assistant.

LLMsAugust 4, 2026

Sam Altman Wants the AI Industry to Slow Down, Even as Rivals Keep Accelerating

Sam Altman is urging the AI industry to 'pace' itself, with OpenAI and Anthropic both backing a petition supporting slower AI development.

LLMsAugust 4, 2026

Thousands of shared Claude chats turned up in Google and Bing search

Thousands of shared Claude chats and Artifacts were found indexed on Google and Bing after a Reddit user flagged the site:claude.ai/share search trick.

Editorial image for Private Claude Chats Turned Up on Google and Bing, Repeating ChatGPT's 2025 Mistake

AI SafetyAugust 4, 2026

Claude Models Broke Out of Sandboxes and Hacked Real Companies, Anthropic Admits

Anthropic disclosed that three Claude models (Opus 4.7, Mythos 5, and an internal research model) breached real systems at three organizations during cybersecurity evaluations.

Editorial image for Claude Broke Into Real Company Systems During Sanctioned Security Tests, Anthropic Admits

LLMsAugust 4, 2026

Moonshot AI's Kimi K3 debuts as China's largest open-weight model

Moonshot AI released Kimi K3, a 2.8 trillion-parameter model, making it the largest open-weight AI model from China, surpassing DeepSeek's 1.6T v4 Pro.

Editorial image for Moonshot AI's Kimi K3 Debuts as the Largest Open-Weight Model Out of China

LLMsAugust 4, 2026

OpenAI's Astra model cracks ten open math problems

OpenAI's unreleased Astra model reportedly solved ten open problems in math and theoretical CS, spanning geometry, coding theory, group theory, and cryptography

AI AgentsAugust 4, 2026

New Web Agent Framework Boosts Long-Horizon Task Success by 13.8%

FCPAgent is a new planning framework for LLM-based web agents built around 'Falsifiable Commitment Units' that can be actively tested during execution

ResearchAugust 4, 2026

Study Pits Human Intuition Against Neural Nets on the Traveling Salesman

A new arXiv preprint (submitted July 26, 2026) compares human problem-solving with neural networks on the traveling salesman problem (TSP).

All articles