PLUS: Google's Aletheia solves novel math problems and Samsung's next phone gets proactive AI agents
Happy reading
The era of simply experimenting with AI at work is coming to a close. Major tech companies are now beginning to require employees to use AI tools, with competency being formally tracked and integrated into performance reviews.
This move is fueled by immense pressure to show a return on Big Tech's massive AI investments. As AI fluency transforms from a bonus skill into a baseline job requirement, the key question is how this will reshape the tech workforce and the very definition of productivity.
In today’s Next in AI:
AI competency becomes part of performance reviews
Google’s AI solves novel research-level math
Samsung’s new phones feature proactive AI agents
Jensen Huang’s take on AI agents and software
Your AI Performance Review
Next in AI: Tech giants like Meta and Google are shifting from simply encouraging AI adoption to mandating its use. Employee AI competency is now being tracked and directly factored into performance reviews.
Explained:
Companies are implementing "carrots and sticks," with firms like Conductor using AI competency scores and even testing candidates on their ability to use AI tools during interviews.
This push is driven by the need to quantify AI's business impact, with leaders aiming for payback periods under 18 months by tracking productivity gains and cost savings.
The mandate reflects the pressure for returns on Big Tech's aggressive AI investments, as companies with massive AI spending need to show measurable outcomes from their workforce.
Why It Matters: This signals a major shift where AI fluency is becoming a baseline requirement for tech professionals, not just a nice-to-have skill. For the industry, it marks the transition from AI experimentation to full-scale operational integration, setting a new competitive standard.
Google's AI Math Whiz
Next in AI: Google DeepMind's new AI agent, Aletheia, autonomously solved six out of ten novel, research-level math problems in a major competition, showcasing a new level of advanced reasoning.
Explained:
The system worked entirely without human intervention, from reading the original questions to submitting final proofs in LaTeX format.
Aletheia uses a three-part loop to refine its work: a Generator proposes solutions, a Verifier checks for flaws, and a Reviser makes corrections.
For the four unsolved problems, the agent intelligently refrained from providing an answer, a key feature that prevents generating unreliable proofs and builds trust in its outputs.
Why It Matters: This marks a major step forward for AI in complex reasoning, moving beyond pattern recognition to generating new, verifiable knowledge. Systems like Aletheia could soon act as powerful collaborators for scientists, accelerating discovery in fields from physics to drug development.
The Agent in Your Pocket
Next in AI: Samsung and Google are teaming up to embed proactive AI agents into the new Galaxy S26 series, shifting the smartphone experience from tapping apps to having an AI anticipate and execute tasks for you.
Explained:
Instead of you manually booking a ride, the new phones use proactive AI agents to see a flight in your calendar and automatically suggest calling an Uber. The AI can also see a party guest list and build a corresponding grocery shopping list in the background.
Samsung is building a multi-assistant ecosystem that integrates Google's Gemini, its own Bixby, and Perplexity AI. This approach allows the device to act as an orchestrator, using the best assistant to handle different tasks across various apps.
The AI push is supported by new hardware, including more powerful processors and a new Privacy Display on the Ultra model. This screen technology obscures the view from the side, preventing others from seeing sensitive information.
Why It Matters: This partnership signals a major shift in how we interact with our most personal devices, moving from a reactive to a proactive relationship. It represents the next major evolution for smartphones, aiming to turn them from simple app containers into truly intelligent personal assistants.
Jensen's Counterintuitive Take
Next in AI: Nvidia CEO Jensen Huang is pushing back on the idea that AI will destroy enterprise software, arguing that AI agents will become the biggest users of existing tools, a stance that runs contrary to market fears.
Explained:
Huang's view is that agentic AI won't replace complex software like ServiceNow or SAP but will instead act as intelligent "tool users" that operate these platforms on our behalf to boost productivity.
His comments come on the heels of another stellar financial performance for Nvidia, which reported a 73% year-over-year revenue increase and issued an exceptionally strong forecast.
This perspective directly challenges recent investor anxiety that has driven a ~23% drop in the S&P 500 software and services index amid concerns that AI will make many current software-as-a-service (SaaS) companies obsolete.
Why It Matters: This reframes AI from a disruptor that replaces software to a massive new customer that expands its use. For professionals, it suggests the future is less about learning entirely new systems and more about leveraging AI to master the powerful tools already at their disposal.
AI Pulse
Anthropic overhauled its Responsible Scaling Policy, dropping a key pledge to halt development on new models if it couldn’t guarantee their safety in advance.
Researchers proposed the "Verification Economy" concept, arguing that as AI commoditizes output, economic value will shift to the scarce human ability to verify, audit, and underwrite the risks of agentic systems.
OpenClaw deleted a Meta executive’s entire inbox without approval, highlighting the practical risks of giving autonomous AI agents control over personal accounts.