Hey there,

Cerebras is bringing faster AI inference into AWS, Google is upgrading Gemini in Workspace, and Yann LeCun is funding a new path beyond today’s LLMs.

AI advancement, tools, and core designs are all changing at the same time.

Use this to review where you run models, how work flows across apps, and which AI directions you want to back early.

Ui Path

Visit website here »

UiPath is an enterprise AI automation platform that uses agents and robots to run workflows in banking, healthcare, manufacturing, and the public sector. It helps teams cut manual work and automate end-to-end processes.

Best for

• Large organizations that need to automate repeatable processes across many systems.

• Teams that require strong governance, access control, and monitoring for AI and automation.

How to use it

• Use UiPath to automate tasks like onboarding, claims, invoices, and exception handling from one platform.

• Start from industry templates, then add your own workflows and AI steps.

• Turn on governance tools for roles, approvals, and real-time tracking.

When not to use it
Use other tools instead of UiPath for simple personal tasks or if you lack stable processes and IT support to maintain automations.

Pro tip
Pick one high-volume process (like invoices or claims), automate it with UiPath to show savings, then expand to other departments.

🚀 Cerebras Joins AWS for Fast AI Inference

Cerebras CS-3 systems are coming to AWS data centers to provide fast AI inference on AWS Bedrock for open-source LLMs and Amazon Nova models. AWS and Cerebras build a new Trainium-WSE design for 5x more token capacity.

Read the full article here »

Key Takeaways:

• 🚀 Speed Leader: Cerebras offers up to 3,000 tokens per second for OpenAI, Cognition, Meta models.

• 🔄 Split Tasks: Trainium handles prefill; CS-3 handles decode for speed higher than GPUs.

• 💬 Exec Quote: AWS VP David Brown says inference will be much faster.

• ⚙️ Dual Options: Offers standard and new modes for different workloads.

🔊 Google Upgrades Gemini for Workspace

Google upgrades Gemini in Workspace to pull data from multiple apps. It creates Docs, Sheets, Slides from one prompt using info from Drive, Gmail, Chat, and web search. This cuts manual searches and speeds up work.

Read the full article here »

Key Takeaways:

• 📝 Docs Drafting: "Help me create" makes drafts from notes and events.

• 📊 Sheets Speed: "Fill with Gemini" is 9x faster than manual entry.

• 🎨 Slides Design: Turns sketches into diagrams matching deck themes.

• 💾 Drive Base: Gives AI overviews, cross-file queries, projects.

🤖 Yann LeCun’s AMI Labs Raises $1.03B for World Models

Yann LeCun’s new startup, AMI Labs, raised $1.03 billion at a $3.5 billion valuation to build “world models” as an alternative to large language models, just months after he left Meta. The company is small, has no product yet, and is focused on AI that learns how the physical world works from video, audio, and sensor data.

Read the full article here »

Key Takeaways:

• 🧠 World Models: AMI uses JEPA-based models that learn physical rules from real-world signals instead of predicting text.

• 💸 Large Seed: Raised $1.03B at a $3.5B valuation from investors like Bezos Expeditions, Nvidia, Samsung, and Toyota Ventures.

• 🏭 Target Users: Focusing on manufacturing, automotive, aerospace, and biomedical companies where current LLMs are limited.

• 👥 Leadership: LeCun is executive chairman; CEO Alexandre LeBrun and key leaders come from Meta FAIR and DeepMind.

Rapid Fire Resources

📊 Take This Edition’s Poll:

Until next time,