Larry Dignan

OpenAI outlined its five principles to AGI and said "power in the future can either be held by a small handful of companies using and controlling superintelligence, or it can be held in a decentralized way by people."

The company said it favors democratization, empowerment, universal prosperity, resilience and adaptability. You can find the flowery prose, daily dose of irony and probably a chuckle or two here.

China has reportedly blocked Meta's acquisition of Manus, which is registered in Singapore, but operates in China. The South China Morning Post reported that China's National Development and Reform Commission asked the parties to cancel the deal. The paper noted "that Beijing feared the Manus case could set an uncomfortable precedent for other AI companies, leading more to move their operations abroad."

• Meta hopes to accelerate its AI agent plans with Manus acquisition
• Meta releases Muse Spark model, aims to get back into LLM race

DeepSeek released its latest flagship models--V4 Pro and V4 Flash. The open-source models are garnering attention because they are closing the gap on closed models.

Reuters reports that Thoma Bravo is nearing an agreement to hand over software company Medallia to its lenders. The move will wipe out $5.1 billion in equity for Thoma Bravo and its co-investors.

Thoma Bravo acquired Medallia for $6.4 billion in 2021. At the time, Thoma Bravo said: "Through this transaction, Medallia will become a private company with additional resources and greater flexibility to build on its innovation leadership and expand its customer impact."

It's safe to say things have changed.

Microsoft said Copilot's agentic features in Word, Excel and PowerPoint are generally available and enabled by default for 365 Copilot and 365 Premium.

CrowdStrike launched Project QuiltWorks, a coalition aimed at continuously fixing vulnerabilities being discovered by frontier models. The effort includes Accenture, EY, IBM Cybersecurity Services, Kroll, OpenAI and CrowdStrike.

Elastic said Google Distributed Cloud will embed the company's agentic security operations platform with analytics, automation and XDR for air-gapped environments.

• Elastic Security bets on native Workflows, data-centric approach
• Elastic sees early traction with Agent Builder, context play, hybrid architecture

Here's Tesla's deck for the first quarter. Tesla's earnings were better than expected, but the enthusiasm on Wall Street revolved around the following excerpts.

• "We expect volume production of both Cybercab and the Tesla Semi this year."
• "Progress continued at the new Megafactory outside Houston, which will produce the Megapack 3 for Megablock. Start of production is on track for later this year."
• "Preparations for our first large-scale Optimus factory will begin shortly in Q2."

AI agents require forward deployed engineers, partners and integrators. Google Cloud is spending $750 million to develop its partner network for agentic AI as it announced a bevy of partnerships with integrators as well as enterprise software giants such as Salesforce and SAP.

The bottom line is Google Cloud is building out its ground game for AI agent deployment. Key announcements at Google Cloud Next include:

• Google Cloud will spend $750 million to accelerate customer deployments, AI prototyping and using Google FDEs.
• Deloitte launched a Google Cloud Agentic Transformation practice and Accenture expanded its Google Cloud efforts on agentic AI. BCG and Google Cloud also expanded a partnership as did McKinsey.
• Salesforce and Google Cloud said their agents will be able to collaborate with context and end-to-end workflows across their platforms.
• SAP said Google Cloud outlined a multi-agent integration pact with Gemini Enterprise and SAP’s Joule agents. Customers can deploy Joule Agents in SAP CX Solutions to build, launch and optimize marketing campaigns. Gemini Enterprise acts as a central hub for agents to take action across SAP and Google Cloud platforms.

OpenAI launched workspace agents in ChatGPT in what the company called "an evolution of GPTs." The idea is that ChatGPT and Codex become more tightly integrated and Codex will create agents for teams. Workspace agents will be available in research preview in ChatGPT Business, Enterprise, Edu and Teachers plans.

Cursor and SpaceX will partner in a deal that will give the AI company xAI's Colossus infrastructure to scale up its models. SpaceX will pay Cursor $10 billion for the joint models or pay $60 billion for the Cursor. Think of it as a circular deal meets try-before-you-buy acquisition.

As you’d imagine there’s a lot of commentary (some of it comical) on this arrangement, but the core themes are:

• Why did SpaceX buy xAI if it needs Cursor? We know the answer to that one—Elon. Another note: Why did SpaceX pay $250 billion for xAI? Same answer.
• SpaceX needs to monetize xAI’s Colossus infrastructure somehow.
• Cursor needs its own models to survive and they need to be cutting edge. xAI by the way needs models too.
• And xAI hired two Cursor co-founders so why not get the rest?

In any case, Cursor is doing good work and hopefully that’ll keep coming. We can have Claude and Codex running away with all the code.

• Cursor 3 to manage multiple AI coding agents
• Cursor launches Composer 2
• Infosys, Cursor form strategic pact

Google now offers two research agents including Deep Research, which replaces its preview release, and Deep Research Max. Both are available via Gemini API and paid tiers. Both models run on Gemini 3.1 Pro.

Deep Research is designed for speed and efficiency while Deep Research Max is designed to be complete.