Meta AI Agent Takes Over WhatsApp Business, Anthropic Powers Salesforce Sales Skills | AI Daily Brief
Audio in Mandarin Chinese · English transcript below
⚡ Meta lets Agents run WhatsApp Business, Anthropic arms Claude with Salesforce skills, Perplexity locks Crusoe compute for full LLM lifecycle—AI Agents are evolving from chatbots to operators who execute business workflows.
The AI industry is shifting from model competition toward deep integration of infrastructure and commercial interfaces. Meta's embedding of AI Agents into WhatsApp's business ecosystem, Perplexity's multi-year compute deal with Crusoe to support high-concurrency inference, and Google's Gemini 3.8 Live upgrade for real-time voice interaction collectively demonstrate that unlocking LLM value requires simultaneous restructuring of compute and application interfaces. Notably, the convergence of Steve Bannon and Bernie Sanders on AI existential risk signals an unprecedented urgency for governance even as the industry sprints forward.
Today's Top 3 Headlines
- AI
🤖 Meta Launches WhatsApp Business Tools MCP, AI Agent Manages Messaging Without Switching 4 Tools
Meta launches WhatsApp Business Tools MCP, letting AI coding Agents connect to WhatsApp Business Platform to manage messaging directly. For developers, this shifts messaging setup from multi-tool toggling to Agent automation.
Source ↗ - AI Industry News
🤖 Anthropic + Salesforce launch Claude Agent with 37 built-in sales skills
Anthropic and Salesforce launch "Salesforce in Claude" beta plugin, connecting customer accounts, opportunities and sales workflows into Claude with 37 sales skills covering morning briefings, call prep, deal evaluation and post-call CRM auto-updates.
Source ↗ - AI Industry News
🤖 Crusoe inks multi-year deal with Perplexity, GB300 cluster delivers 9.9x first-token speedup
Crusoe inks multi-year deal with Perplexity to train models on NVIDIA GB300 NVL72 clusters, with inference first-token latency cut by 9.9x. Gives Perplexity dedicated compute backbone for low-latency, high-throughput inference and sharper AI Q&A experience.
Source ↗
+4 more headlines
- 🤖 Google launches Gemini 3.8 Live dual models, voice inference hits 97.7%
- 🤖 Qwen3.8-27B inference tokens slashed 40%, efficiency leap
- 🤖 Ferrovalle picks INFORM to deploy AI yard system at Mexico City's 550K TEU terminal
- 🤖 Bannon, Sanders Rarely Agree: AI Poses Existential Risk to Humanity
Podcast highlight
Today's main plot: the settings page vanished, and what's hiding behind it is who's actually collecting the money.