AI News Today (2026-06-25): The Hardware Shift, the Ethics Wars, and Why Philosophers Are Getting Hired

Quick Summary: Custom silicon, edge AI consolidation, and agentic computer use are redrawing the technical career map, while high-stakes disputes over model theft and government access create urgent demand for AI policy and security talent. Meanwhile, a quiet hiring trend reveals that the hottest new AI team member might have a philosophy degree.

OpenAI and Broadcom drop “Jalapeño,” the first custom inference chip

OpenAI took the wraps off its first in-house silicon, built with Broadcom, designed specifically to accelerate LLM inference. The move signals a strategic shift away from total reliance on Nvidia and toward vertically integrated AI infrastructure.

What it means for you: Hardware engineering, chip design, and AI infrastructure roles are about to get even hotter. If you work in DevOps or cloud architecture, expect inference cost curves to bend—and start building fluency in hardware-aware model optimization to stay ahead.

Qualcomm buys Modular, placing a massive bet on edge AI and the Mojo language

Qualcomm announced it will acquire Modular, the startup behind the Mojo programming language and a next-generation AI engine. The deal points to a future where AI runs pervasively on phones, wearables, and IoT devices, not just in the cloud.

What it means for you: Edge AI skills are moving from niche to necessary. Learning Mojo or optimizing models for mobile/embedded targets could differentiate you in the next wave of hiring. For mobile developers, the line between app dev and AI dev just vanished.

Google’s Gemini 3.5 Flash can now use computers—ushering in the agentic automation era

Gemini 3.5 Flash gained a “computer use” capability, allowing it to interact with UIs, click buttons, and navigate software like a human. This turns the model into a general-purpose automation agent that can operate legacy tools without APIs.

What it means for you: Roles in QA, RPA, operations, and even customer support will be reshaped. The skill to cultivate is not just prompting, but designing, supervising, and auditing agentic workflows. Start experimenting with computer-use APIs today to understand what’s about to be automated under your nose.

Big AI labs are on a philosopher hiring spree

Multiple frontier labs are quietly adding philosophers to their teams, seeking expertise in ethics, value alignment, and conceptual rigor that pure engineering teams lack. The trend reflects a maturing industry grappling with real-world impact.

What it means for you: Non-technical AI roles are proliferating. If you have a background in philosophy, ethics, policy, or the humanities, this is your moment to translate those skills into high-impact, well-compensated AI safety and alignment positions. For engineers, pairing technical chops with structured ethical reasoning can fast-track you to leadership.

Anthropic’s double dispute: NSA access cut and Alibaba model extraction claims

In a dramatic week, the NSA reportedly lost access to Anthropic’s Mythos intelligence tool amid an unresolved ethical dispute, while Anthropic separately accused Alibaba of illicitly extracting Claude’s capabilities. These twin events expose the volatile intersection of national security, IP protection, and AI ethics.

What it means for you: Demand is spiking for professionals who sit at the crossroads of AI, security, and policy. Roles in AI governance, model security, IP law, and government contracting will grow. If you can speak both technical and regulatory languages, you’re suddenly invaluable.

Reid Hoffman warns: xAI is a “complete train wreck,” SpaceX isn’t an AI company

Investor and LinkedIn co-founder Reid Hoffman publicly criticized xAI’s execution and dismissed SpaceX’s AI ambitions, reminding the market that not every company waving the AI flag is a safe career bet.

What it means for you: When evaluating job offers, look past the hype. Real AI companies are defined by research output, infrastructure ownership, and product traction—not just a famous founder. Vet your next employer ruthlessly, and prioritize roles where AI is the core business, not a slide deck buzzword.

RubyLLM gives Ruby devs a unified AI framework—and a career boost

A new open-source gem, RubyLLM, lets Ruby developers tap into all major AI providers through a single, elegant interface. It removes the friction that previously pushed AI-curious Rubyists toward Python.

What it means for you: If you’re a Ruby developer, this is your low-friction on-ramp to AI integration. Building AI features into Rails apps just became trivial, and that niche expertise can command a premium. For polyglot developers, it’s a reminder: AI isn’t just a Python game anymore.

The one thing to act on today

Spend 30 minutes with Gemini 3.5 Flash’s computer-use API (or any agentic framework) and automate one repetitive task you currently do manually. The goal isn’t the automation itself—it’s to build the muscle of thinking in terms of AI agents, because that mindset will separate the replacable from the indispensable in the next two years.

Join the SuperCareer AI career newsletter for your personalized roadmap.