Tomorrow
← All briefings

The briefing · 20 Aug 2026 · 120 tracked changes

Enterprise Agents and Upgraded Hardware Fuel the Next Wave of Automation

A wave of agentic software updates, deep-tech hardware transitions, and advanced multimodal features dominated the tech landscape today.

Coding and Automation Shift Toward Autonomous Agents

Major software platforms are rapidly integrating autonomous agents to handle complex workflows. Microsoft's Copilot Studio introduced new AI agents and Model Context Protocol (MCP) support, while GitHub Copilot added custom instructions. Meanwhile, standalone coding platforms like Claude Code and Lovable added capabilities like terminal integration and natural language coding to streamline software development.

Why it matters Software is transitioning from tools that merely assist to active agents that can execute entire workflows autonomously.

copilot-studiogithub-copilotclaude-codelovable

Built from 4 tracked changes

Corporate Giants Deploy Specialized Workforce Assistants

Enterprise software giants are rolling out specialized assistants to target specific business operations. Oracle added Human Capital Management (HCM) and supply chain features to its Fusion AI Agents, while Zoom's AI Companion introduced live translation and screen sharing. Data platforms like Databricks Mosaic AI also added agent evaluation tools to help companies deploy and monitor these new corporate assistants.

Why it matters AI is moving deeply into back-office corporate operations, managing everything from employee relations to supply logistics.

oracle-fusion-aizoom-ai-companionmosaic-ai

Built from 3 tracked changes

Next-Gen Hardware Optimizes for Heavy Workloads

Hardware companies are upgrading their silicon architectures to run larger, more complex models more efficiently. Google's TPU Ironwood, now in its 7th-generation, added OpenXLA support, while Tenstorrent's Blackhole chip integrated RISC-V architecture. Additionally, SambaNova's SN40L added vision support and integrated Meta's massive Llama 3.1 models directly into its stack.

Why it matters Faster, specialized hardware is crucial to lowering the costs and energy required to run advanced AI.

google-tputenstorrentsambanova

Built from 5 tracked changes

Multimodal Models Bridge Video, Voice, and Design

AI applications are expanding their sensory abilities, combining language processing with visual and interactive tools. Google DeepMind's Nano Banana image tool introduced multimodal reasoning and conversational editing, while Figma AI introduced new image tools, layers, and interactive capabilities. Additionally, China's Shengshu AI added text-to-speech to its Vidu video generator, bridging the gap between sound and video.

Why it matters Creative tools are moving away from isolated tasks, allowing users to edit images and video using natural, conversational instructions.

nano-bananafigma-aividu

Built from 3 tracked changes

See the raw change log