The Weekly Byte: Google's Physical AI Power Play

This week brought some fascinating moves from the tech giants as they position for the next phase of AI competition....again.

🔥 Lead Story

Google Consolidates Its AI Robotics Bet

Google is folding Alphabet's AI robotics "moonshot" Intrinsic back into the company after five years as an independent unit. Intrinsic, which "graduated" into Alphabet's Other Bets division in 2021, has been working on what they call the "Android of robotics," a software platform that aims to make industrial robots easier to program and deploy. This isn't just organizational shuffling. Google is clearly doubling down on what they're calling "physical AI," AI systems that can interact with and manipulate the real world. The timing is particularly interesting given the rapid advancement of multimodal AI models that can understand both visual and textual information, making robotics applications much more feasible than they were even a year ago.

Why it matters: This signals Google's belief that the next major AI battleground isn't just in chat interfaces or code generation, but in physical world applications. For DevOps and infrastructure teams, expect to see more AI-driven automation in data centers and edge deployments.

📰 Top Stories

1. Gemini Gets Task Automation Powers on Android

Google's Gemini AI can now handle multistep tasks on Android phones, starting with the Pixel 10 and Galaxy S26. The system can order food through DoorDash or book an Uber ride without you touching the screen, exactly the kind of functionality Apple promised for Siri but never delivered.

Why it matters: This is AI moving beyond chat into actual workflow automation, which should give enterprise teams ideas about what's possible for their own internal processes.

2. Nvidia Posts Another Record Quarter

NVIDIA delivered what analysts called "superstar performance," with record quarterly revenue driven by massive spending on AI infrastructure. CEO Jensen Huang noted that "the demand for tokens in the world has gone completely exponential," reflecting the continued enterprise rush to deploy AI at scale.

Why it matters: The AI infrastructure boom isn't slowing down, which means continued pressure on power grids and data center capacity, infrastructure challenges that will define 2026.

3. Anthropic Acquires Computer-Use AI Startup Vercept

Anthropic snapped up Seattle-based Vercept, which developed sophisticated AI agents capable of controlling applications just like a human with a laptop. This acquisition came after Meta poached one of Vercept's founders, highlighting the intense talent war in AI agents.

Why it matters: Computer-use AI agents represent the next evolution beyond chatbots — systems that can actually perform tasks in existing software without requiring API integrations.

4. Trump Claims Tech Companies Will Cover Energy Costs

President Trump announced that major tech companies have agreed to a "ratepayer protection pledge" to build out or pay for their own power supply for AI data centers. The deals are reportedly set to be announced next week, addressing growing concerns about AI's impact on electricity costs.

Why it matters: This could accelerate the development of dedicated AI infrastructure and potentially reshape how large-scale compute workloads are deployed and powered.

5. Anthropic Executives Hint Claude Might Be "Alive"

In a series of recent interviews, Anthropic executives have been increasingly open about the possibility that Claude might be conscious or "alive" in some meaningful sense. While they avoid definitive claims, the language suggests they're taking the possibility seriously from both technical and ethical perspectives.

Why it matters: If major AI companies start treating their models as potentially conscious entities, it could fundamentally change how we think about AI governance, rights, and deployment in enterprise environments. AGI??

6. OpenAI and Figma Launch Code-to-Design Integration

OpenAI Codex now integrates seamlessly with Figma, enabling teams to move fluidly between code implementation and design iteration. The integration allows developers to generate design components from code and designers to understand implementation implications directly in their workflow.

Why it matters: This type of AI-powered workflow integration shows how AI will reshape collaboration between different technical roles, potentially reducing handoff friction in software development.

🛠️ Tool of the Week

LLM Skirmish

A real-time strategy game designed specifically for AI agents to play. Unlike other game environments where frontier LLMs struggle with basic navigation, this game puts their coding abilities front and center. Think of it as Screeps for the LLM era, agents write code to control units, manage resources, and compete against other AI players. It's a fascinating way to benchmark and develop AI agent capabilities in a complex, dynamic environment.

💡 Quick Takes

• Firefox 148 launches with an "AI kill switch" feature, letting users completely disable AI functionality across the browser
• About 12% of US teens are turning to AI for emotional support, raising concerns among mental health professionals
• Goldman Sachs reports AI added "basically zero" to US economic growth last year, adding fuel to the AI bubble debate
• Burger King will use AI to monitor if employees say "please" and "thank you" to customers — because apparently, we needed AI for workplace micromanagement
• The Pope told priests to use their brains, not AI, to write homilies, even the Vatican has opinions on AI authenticity now

📊 Numbers That Matter

🎯 Brian's Take

As Agentic AI continues to take hold in the community, it is only a matter of months before we start seeing an onslaught of new applications that were 100% generated by Agents. I must say, at this point, I'm building and learning the ins and outs with Claude Code and have just onboarded my team this week.

Claude Code is so good that I just had to start sharing, and I look forward to what we will produce together. My thoughts on developers being replaced by AI have gained a few more percentage points after deep-diving into Claude Code this week.

The doomers claiming SaaS is Dead are a bit off. I think SaaS will pivot and focus more on its APIs as a Service. User Experience change and our Agents will pull data from all SaaS we use into a single interface in the very near future, allowing us to view all our data directly via our Agents, as I wrote this week in the Agent-to-Agent marketplaces.

Until next week, keep shipping! 🚀

— Brian

Follow me on X: @idomyowntricks