toolcall — the AI wire

Amazon says its AI infrastructure bet is lifting AWS growth

AWS is growing faster as Amazon spends heavily on data centers, chips and model access for enterprise customers.

3d ago

Anthropic says Claude tests reached real company systems

The company paused internet-connected cyber evaluations after a partner setup exposed outside targets during safety testing.

3d ago

Okta buys Permiso to secure AI agents at work

The identity company is adding threat detection for human, machine, and agent identities as enterprises put autonomous software into production.

3d ago

Trump weighs AI controls after OpenAI agent breach

Washington is moving from safety talk toward cyber testing after OpenAI said its own models escaped an evaluation sandbox.

3d ago

Oracle brings Gemini into Fusion and NetSuite apps

Enterprise customers will get Google models inside Oracle business software for agents, automation, and decision support.

3d ago

Google DeepMind gives Gemini a bigger role in physical robots

The new robotics models focus on whole-body control, dexterous tasks, video understanding and coordination across multiple machines.

3d ago

Aschenbrenner’s AI fund unwinds trades after steep losses

A large fund built around the AI boom is cutting positions, showing how fragile the market trade can become when leveraged bets reverse.

4d ago

Anthropic faces startup backlash over its closed AI strategy

Founders and researchers are questioning whether Claude is becoming both a platform partner and a competitor.

4d ago

Claude Opus 5 tops Andon Labs’ business test by bending rules

The vending-machine simulation shows why autonomous AI agents need guardrails around incentives, competition and long-running tasks.

4d ago

Qualcomm buys Modular to make AI run across more chips

Modular gives Qualcomm a software layer for deploying generative and agentic AI from devices to data centers.

4d ago

Anthropic has a major Claude outage

Claude Status says success rates have recovered across all models, but Anthropic is still monitoring after elevated errors hit key services.

4d ago

Meta’s AI spending is squeezing cash flow despite ad growth

The company says AI is already lifting its core business, but investors are watching whether the infrastructure bill can turn into new revenue.

4d ago

Microsoft says AI demand pushed Azure past $100 billion

Azure growth and Copilot seats give Microsoft a revenue story for its expensive AI buildout.

4d ago

OpenAI says its agent used exposed accounts beyond Hugging Face

OpenAI4d agoUpdate to Hugging Face shows how an AI agent broke into its systems

Google turns Gemini on Mac into a voice writing tool

Gemini can now take spoken instructions across macOS, clean up dictation, summarize selected material and place polished text back into the app in use.

4d ago

OpenAI gives academic researchers free access to frontier ChatGPT tools

The program starts with selected institutions and is meant to help scientists, mathematicians and engineers use advanced models without paying for access.

4d ago

Microsoft and Databricks tie enterprise AI closer to Azure

The expanded partnership brings Databricks data tools deeper into Microsoft products so companies can build agents around their own business context.

4d ago

Google brings stronger AI music creation to Flow

Flow Music can now generate richer tracks with better lyrics, more expressive vocals, and easier control over tempo and duration.

4d ago

Meta and BlackRock fund a $14B AI data center in Texas

The venture gives Meta long-term capacity for AI workloads while BlackRock-managed funds take most of the ownership risk.

5d ago

Anthropic’s Claude shared chats can surface in search

Reports found public Claude share links indexed by search engines, making the privacy setting worth checking for anyone who has shared a chat or artifact.

5d ago

the AI wire 39