Anthropic’s Claude Haiku 5.5: Cheaper, Faster and Beating GPT-6 Luna on Agent Tasks

Anthropic’s Claude Haiku 5.5: Cheaper, Faster and Beating GPT-6 Luna on Agent Tasks

Anthropic launched Claude Haiku 5.5 on October 7, 2026, positioning it as the cheapest, fastest and most capable small model the company has ever shipped. The economics explain the fanfare: for prompts of up to 100,000 tokens — roughly 90% of the requests the Haiku line receives — input costs drop to $0.10 per million … Read more

Google Playground Turns a Text Prompt Into a Playable Game — No Code Required

Google Playground Turns a Text Prompt Into a Playable Game — No Code Required

Google has launched Playground, an experimental platform that assembles complete video games — multiplayer modes, leaderboards and visual assets included — from plain-language text instructions, with no coding required. The service went live on October 7, 2026 at playground.google for users in the United States aged 18 and over, and Google One AI Premium subscribers … Read more

SpaceX’s Grok Bot Puts Claude Opus 5.5, Midjourney and Suno on the Support Desk

SpaceX’s Grok Bot Puts Claude Opus 5.5, Midjourney and Suno on the Support Desk

SpaceX has launched Grok Bot, an AI customer-support platform that routes requests through Claude Opus 5.5 for reasoning, Midjourney for image generation and Suno for audio output. Announced by Elon Musk on X on 6 October 2026, the system is available for Starlink subscribers, SpaceX customers and enterprise users of xAI services. The pitch is … Read more

Nous Research Raises $90M at $1.5B Valuation to Sell Open-Source Hermes Agents

Nous Research Raises $90M at $1.5B Valuation to Sell Open-Source Hermes Agents

Nous Research, the lab behind the open-source Hermes agent, has closed a $90 million Series B at a $1.5 billion valuation — a final figure that landed above the roughly $75 million discussed while negotiations were still in progress. Announced on October 7, 2026, the round was led by Robot Ventures, with Nvidia, Union Square … Read more

OpenAI Publishes 722 AI-Generated Math Proofs, Machine-Checked With Lean

OpenAI Publishes 722 AI-Generated Math Proofs, Machine-Checked With Lean

OpenAI has released the largest publicly available corpus of AI-generated mathematical research to date: 722 manuscripts organised into 372 families of results, produced by an internal frontier model that the company has not yet made public. The GitHub repository, live since early October 2026, pairs many of the papers with formalisations written in Lean, the … Read more

Meta Deploys LLM Detection to Catch Ads That Funnel Users to CSAM

Meta Deploys LLM Detection to Catch Ads That Funnel Users to CSAM

Meta has deployed a new layer of artificial intelligence aimed at a stubborn loophole in its advertising machine: campaigns that look entirely harmless but exist only to funnel users toward child sexual abuse material hosted off its platforms. Instead of hunting explicit imagery — a job its hash-matching systems have handled since 2011 — the … Read more

Google Opens SynthID Detector to Everyone as Apple Signs On to Its AI Watermark Standard

Google Opens SynthID Detector to Everyone as Apple Signs On to Its AI Watermark Standard

Google has opened SynthID Detector to the public worldwide, turning a tool that spent more than a year gated behind a waitlist for journalists, media professionals and researchers into open verification infrastructure. From today, anyone can visit synthid.com, upload an image, a video or an audio clip, and learn within seconds whether it carries an … Read more

EmbeddingGemma 2 Puts Five-Modal AI Search Directly on Your Smartphone

EmbeddingGemma 2 Puts Five-Modal AI Search Directly on Your Smartphone

Google DeepMind has launched EmbeddingGemma 2, a 740-million-parameter multimodal embedding model engineered to run entirely on the device rather than in the cloud. The model maps text, code, images, video and audio into a single 768-dimensional vector space, so an application can compare a voice memo against a video clip or a written query against … Read more

Mirror Particle Builds a World Model of Human Behavior to Outpredict LLMs

Mirror Particle Builds a World Model of Human Behavior to Outpredict LLMs

Large language models can draft a persuasive email, summarise a dense research paper and hold a coherent conversation. What they cannot do reliably, according to Mirror Particle, is predict what one specific person will actually do next Tuesday. The San Francisco startup is betting that a purpose-built world model of human behaviour can close that … Read more

Hark Pro Puts Brett Adcock’s AI Agent on the Desktop, With Hardware Due in 2027

Hark Pro Puts Brett Adcock’s AI Agent on the Desktop, With Hardware Due in 2027

Brett Adcock made his name in hardware: first as co-founder of electric air-taxi developer Archer Aviation, then as founder of Figure AI, the humanoid robotics company backed by Microsoft, BMW and OpenAI. With Hark, his newest venture, he is reversing the order of operations — software first, machines later. The debut product, Hark Pro, ships … Read more