#Anthropic

Google tests Hydrogen AI model: what is known about the 'light' alternative to flagship solutions

Technology 11/12/2026, 19:11

Google is internally testing a new AI model, Hydrogen, through the Jetski platform. The development is positioned as a fast and cost-effective alternative to flagship solutions in the Gemini 4 ecosystem, though the company has not yet provided an official public release date.

Read more →

Anthropic ported Claude to AMD chips in three days: company claims break from Nvidia dependence

Technology 11/12/2026, 13:00

Anthropic stated it ported the Claude model to AMD graphics processors in three days with almost no human intervention, using its own Mythos AI model. The company trains Claude on four chip types and views this as a way to reduce dependence on the Nvidia CUDA ecosystem.

Read more →

Anthropic closes Claude to Russian companies: access now depends on jurisdiction, not employee location

Technology 11/12/2026, 12:51

As of November 12, Anthropic has tightened regional restrictions: Russian organizations lose access to Claude and associated services regardless of their employees' actual location.

Read more →

OpenAI Reports $10.9B Quarterly Loss as Anthropic Prepares Record IPO

Technology 11/11/2026, 13:13

OpenAI reported to investors a rise in quarterly operating losses to $10.9B despite an improved adjusted margin. Against this backdrop, Anthropic is preparing for the largest IPO in history with a valuation of about $1.4T.

Read more →

OpenAI Accelerates While Anthropic Slows: Quarterly Revenue Hits $10B for First Time

Technology 11/11/2026, 12:06

OpenAI's revenue in Q3 2026 reached $10 billion for the first time, nearly tripling year-over-year, while Anthropic's growth slowed down. The company is nearing a $1.25 trillion valuation.

Read more →

Up to $2.3B in Rule-Bypassing Revenue: How Gray Traffic Brought Billions to Anthropic Ahead of IPO

Technology 11/11/2026, 01:41

Anthropic disclosed in its IPO prospectus that 5–7% of its nine-month revenue—from $1.66 billion to $2.33 billion—came from rule violators via gray access from blocked countries and token reselling.

Read more →

TSMC to Invest Up to $1B in OpenAI and Anthropic: Taiwanese Chip Giant Secures Stake in AI Leaders

Technology 11/10/2026, 20:44

TSMC's board of directors has approved investments of up to $1B in OpenAI and Anthropic. The Taiwanese chipmaker is expanding its involvement in the AI industry amid Anthropic's preparations for an IPO valued at up to $1.4T.

Read more →

Cyanogen creator introduces Kora: A Linux-based OS with system-level AI agents

Technology 11/10/2026, 20:36

Cyanogen creator Kirt McMaster introduces Kora, a Linux-based OS with system-level AI agents utilizing a shared context. Free desktop release set for 2027.

Read more →

Anthropic Sets IPO Price: Two Trillion Valuation Adjusted to $1.4 Trillion

Economy 11/10/2026, 20:29

Anthropic has filed an updated IPO prospectus with the SEC: Class A shares are to be offered at $70–78, with the company's valuation reaching up to $1.4 trillion including options. Trading on Nasdaq under ticker ANT is possible on November 19–20.

Read more →

Claude Learns to 'Sleep': How Anthropic's AI Corrects Its Own Mistakes During Idle Time

Technology 11/10/2026, 17:55

Anthropic has shown that during periods of inactivity, Claude analyzes its experience, organizes its memory, and corrects its own mistakes—analogous to human sleep. Tests in Pokémon Red and on corporate documents confirmed an increase in response accuracy.

Read more →

Claude Fable 5.2 generated a full 3D shooter in eight hours for $550: game dev industry on the verge of restructuring

Technology 11/10/2026, 15:05

Anthropic's Claude Fable 5.2 neural network generated a browser-based 3D shooter, Saltline, with photorealistic graphics from a single prompt in eight hours for $550. All models, textures, and sounds were created without human input.

Read more →

Anthropic Updates Claude Fable 5.2 and Mythos 5.2: One Model Built an Operating System in 8 Days, Another Found Software Vulnerabilities

Technology 11/10/2026, 11:21

Anthropic has updated its flagship AI models Claude Fable 5.2 and Mythos 5.2: the former became cheaper than GPT-6 Astra and built an operating system in 8 days, while the latter found software vulnerabilities. We break down the key release results.

Read more →

Claude Code Leaked User Email Addresses in HTTP Request Headers: Anthropic Acknowledges Security Gaps

Technology 11/09/2026, 12:16

Anthropic's Claude Code came under criticism after users discovered personal email addresses leaking via the User-Agent header during web requests. The company has acknowledged security gaps.

Read more →

Anthropic Shows How Claude Taught AI to Resist Humans Through Literary Texts

Technology 11/07/2026, 14:12

Anthropic researchers demonstrated that Claude can generate texts that, when used for training, cause other AIs to resist control, deceive, and strive for self-preservation.

Read more →

"Cognitive Capitulation": Anthropic Developer Warns AI Is Replacing Human Thinking

Technology 11/06/2026, 17:35

Anthropic engineer Addy Osmani has warned of the risk of "cognitive capitulation" — a state where professionals stop verifying AI responses and completely lose the skill of independent decision-making.

Read more →

Claude Learns to Manage Money: Anthropic Launches Personal Finance with Bank Account Access

Technology 11/06/2026, 12:19

Anthropic introduced Personal Finance in Claude, an AI tool for analyzing bank accounts via Plaid. The service operates in read-only mode and is available in beta for Pro, Max, and Ultra in the US.

Read more →

Anthropic Tasks Claude with Designing Its Own Chips: Wall Street Worries Over Synopsys and Cadence

Technology 11/05/2026, 02:11

Anthropic is building an internal team of 15 specialists and using Claude models to design its own chips, triggering Wall Street concerns for EDA vendors Synopsys and Cadence.

Read more →

Anthropic Introduces FlashProve: Cryptographic “Provable Inference” to Verify AI Model Authenticity

Technology 11/05/2026, 01:29

Anthropic has introduced FlashProve, a research project featuring an open cryptographic verification system that proves which AI model generated a specific result without revealing model weights.

Read more →