Home / Blog / Nvidia Drops a Bomb, Claude Gets Eyes, and Musk's AI Company Is Falling Apart - Your Friday AI Briefing

AI NEWS

Nvidia Drops a Bomb, Claude Gets Eyes, and Musk's AI Company Is Falling Apart - Your Friday AI Briefing

Nvidia's NemoClaw agent platform, Claude's inline visuals, Meta delays Avocado, xAI's cofounder exodus, and Perplexity's desktop agents - your complete Friday AI briefing.

By Robert Yeager, Founder and Full-Stack Developer · · 6 min read · 1,282 words

Nvidia Drops a Bomb, Claude Gets Eyes, and Musk's AI Company Is Falling Apart - Your Friday AI Briefing

The AI landscape just shifted again overnight - and if you blinked, you missed at least three major moves. From Nvidia muscling into the AI agent space to Anthropic giving Claude actual visual superpowers, and Meta quietly admitting their next model isn't ready... Friday the 13th is living up to its name for some companies and delivering gifts to others.

Let's break down what happened and what it means for anyone actually building with these tools.

Nvidia Is Coming for the AI Agent Market with NemoClaw

This is the big one. Ahead of Jensen Huang's GTC 2026 keynote on Monday, reports from Wired confirm that Nvidia is preparing to launch NemoClaw - an open-source AI agent platform designed to compete directly with tools like OpenClaw.

Here's what we know: Nvidia has been privately pitching NemoClaw to corporate partners including Salesforce, Cisco, Google, Adobe, and CrowdStrike. The platform will reportedly run on machines without Nvidia GPUs - a strategic move that prioritizes adoption over hardware lock-in.

Why it matters for business: The AI agent space is about to get a serious heavyweight competitor. OpenClaw proved the concept - always-on AI agents running locally on your machine, handling tasks autonomously. Now Nvidia wants to bring that to the enterprise with the kind of corporate partnerships and support structure that smaller open-source projects can't match. If you're a business exploring AI automation, the next 6 months are going to give you more options than ever.

Jensen Huang takes the stage at the SAP Center in San Jose on Monday, March 16th at 11 AM PT. Expect NemoClaw to be a centerpiece alongside rumored new inference chips. We'll be watching.

Anthropic's Claude Can Now Show You What It's Thinking - Literally

Anthropic rolled out a feature yesterday that sounds simple but changes how you interact with Claude: inline visual content. Claude can now generate charts, diagrams, timelines, and interactive visualizations directly inside your conversation.

Previously, if you wanted Claude to create something visual, it would generate an "artifact" in a side panel. Now, if Claude determines a visual would help explain its response, it creates one right in the flow of conversation. Think: asking about the periodic table and getting an interactive, clickable version embedded in the response. Asking about structural engineering and seeing a force diagram appear mid-explanation.

You can also explicitly ask Claude to generate charts or diagrams. The feature is rolling out to all users and is turned on by default.

Why it matters for business: This is Claude becoming a better communication tool, not just a text generator. When you're analyzing data, explaining concepts to clients, or building presentations, having an AI that can show you things instead of just telling you changes the workflow. We use Claude daily at Fusion Data Co for everything from data analysis to client reporting - inline visuals make those outputs immediately more useful.

It's worth noting that OpenAI and Google Gemini shipped similar features this week too. The race to make AI chatbots more visual is officially on, and users are the ones winning.

Meta Delays "Avocado" - Their Next AI Model Isn't Ready

Here's a story that tells you everything about the current AI arms race: Meta has pushed back the release of its next-generation AI model, codenamed "Avocado," from March to at least May.

The New York Times reports that performance falls short of competitors - particularly Google. This is significant because Meta has spent billions on AI infrastructure and talent, including hiring Scale AI's Alexandr Wang to lead their efforts. Despite all that investment, Avocado apparently can't keep pace with what Google's Gemini team is shipping.

Why it matters for business: If you've been building on Meta's Llama models (which are excellent for open-source deployments), this delay doesn't affect your current stack. But it does signal that the gap between the top-tier models (Claude, Gemini, GPT) and the open-source alternatives may be widening rather than closing. For businesses choosing AI partners, the proprietary models continue to pull ahead in raw capability. Choose your foundation wisely.

The xAI Exodus: 9 of 11 Cofounders Have Now Left Musk's AI Company

This one's hard to spin positively for Elon Musk. Business Insider reports that two more xAI cofounders - Zihang Dai and Guodong Zhang - have departed, leaving only 2 of the original 11 engineers who started the company with Musk in 2023.

Zhang was a particularly significant loss - he reported directly to Musk and led two of xAI's key projects: Grok Code (their coding agent) and Grok Imagine (image and video generation). His departure comes just one day after Musk publicly admitted at the Abundance Conference that "Grok is currently behind in coding."

Adding fuel to the fire: xAI has reportedly hired two employees from Cursor, the AI-powered code editor, suggesting Musk is trying to rebuild his coding capabilities from scratch.

Why it matters for business: If you've been considering Grok for business applications, this instability is a red flag. Leadership continuity matters in AI development - models need consistent architectural vision to improve. Nine of eleven cofounders leaving in roughly a year suggests fundamental disagreements about direction. For now, Claude, Gemini, and GPT remain the safer bets for production business use.

Perplexity Launches "Personal Computer" - Desktop AI Agents for Everyone

Perplexity is taking aim at the personal AI agent market with "Personal Computer" - a desktop application that gives AI agents direct access to your local files and apps. Available in early access by invite, it runs on Mac and lets users describe high-level objectives while AI agents figure out the specifics.

Sound familiar? It should. OpenClaw pioneered this exact concept - letting AI agents operate autonomously on your personal machine. Perplexity's version appears to be a more polished, user-friendly take on the same idea, with a clean dockable interface, multi-task tracking, and remote access from any device.

Why it matters for business: The personal AI agent category is exploding. OpenClaw proved the demand, Nvidia's NemoClaw is targeting enterprise, and now Perplexity is going after the consumer/prosumer market. Within a year, having an AI agent running on your machine will be as normal as having a web browser. Businesses that start building workflows around these tools now will have a massive head start.

Bonus: Gemini Task Automation Goes Live on Samsung and Pixel

Google's Gemini AI assistant now has real task automation capabilities on the Samsung Galaxy S26 and Google Pixel phones. The Verge reports it's arriving in beta - and it's described as "wild." We're talking about the AI assistant actually doing things on your phone, not just answering questions about them.

This is the kind of AI-meets-mobile integration that's been promised for years. It appears Google is finally delivering.

The Big Picture: AI Agents Are the Story of 2026

If there's one thread connecting all of today's news, it's this: the AI industry has moved past chatbots and into agents. Nvidia is building agent infrastructure. Perplexity is putting agents on your desktop. Google is putting agents on your phone. Claude is becoming more visually intelligent. And the companies that can't keep up (Meta, xAI) are showing the cracks.

For businesses, the message is clear: AI is no longer about asking questions and getting answers. It's about deploying autonomous systems that handle work. The companies adapting to this shift right now are the ones that will dominate their industries.


Fusion Data Co builds with these tools every day - Claude, Gemini, OpenClaw, and more. We don't just write about AI; we deploy it for real businesses with real results. If you want AI working FOR your business, not just reading about it, visit fusiondataco.com or call (916) 534-0915.

About the author

Robert Yeager is the Founder and Full-Stack Developer of Fusion Data Co. He builds the whole stack himself: database, backend, front end, voice agents and the automation between them. Reach him at rob@fusiondataco.com or book a 30 minute call.

Related articles

AI NEWS

GPT-5.4 Just Dropped, Cursor's Agents Run Themselves Now, and the Pentagon Is at War With Claude

GPT-5.4 launches with 1M token context, Cursor's Automations run coding agents on autopilot, and the Pentagon just declared war on Claude.

AI NEWS

GPT-5.4 Drops, Anthropic's Code Review Changes Everything, and Meta Just Bought an AI Social Network

GPT-5.4 drops with native computer use, Anthropic launches multi-agent Code Review, Meta buys Moltbook, and LeCun raises $1B for world models.

AI NEWS

xAI Is Imploding, Meta Can't Ship, and NVIDIA's About to Drop a Bomb - Your Saturday AI Briefing

xAI loses cofounders and poaches from Cursor, Meta delays its Avocado model, NVIDIA GTC 2026 kicks off Monday, and Morgan Stanley says the AI breakthrough is...