Friday, May 1, 2026
Latest

Archive — Page 6

69 articles total
OpenAI launches GPT-5.4 with native computer use and spreadsheet plugins

OpenAI launches GPT-5.4 with native computer use and spreadsheet plugins

Two variants target professional work with 1M-token context and integrated financial tools for Excel and Google Sheets.

Databricks Builds RAG Agent Handling Six Enterprise Search Types

Databricks Builds RAG Agent Handling Six Enterprise Search Types

KARL agent trained on reinforcement learning handles multiple search behaviors that typically require separate optimization.

Google DeepMind Releases Nano Banana 2 for Fast Image Generation

Google DeepMind Releases Nano Banana 2 for Fast Image Generation

New model combines advanced capabilities with Flash-speed inference, targeting production workflows.

Google Adds Music Generation to Gemini App With Lyria 3

Google Adds Music Generation to Gemini App With Lyria 3

Users can now create 30-second tracks from text or image prompts using Google's latest music synthesis model.

Google Opal Shows Enterprise Teams How to Build Safe AI Agents

Google Opal Shows Enterprise Teams How to Build Safe AI Agents

Google Labs quietly released a visual agent builder that answers the year-long debate about agent autonomy levels.

Atlassian Embeds AI Agents Into Jira, Treating Them as First-Class Team Members

Atlassian Embeds AI Agents Into Jira, Treating Them as First-Class Team Members

The latest update allows teams to assign, manage, and track AI agents alongside human workers in a unified workflow interface.

Enterprise Security Firm Wraps OpenClaw Agent in Safety Controls

Enterprise Security Firm Wraps OpenClaw Agent in Safety Controls

Runlayer launches secure deployment layer for open-source AI agents, addressing widespread adoption concerns among large organizations.

Google Gemini 3.1 Pro Adds Adjustable Reasoning Tiers to Match Complex Task Demands

Google Gemini 3.1 Pro Adds Adjustable Reasoning Tiers to Match Complex Task Demands

Google's latest flagship model introduces three levels of computational thinking, positioning itself as a practical alternative to specialized reasoning models.