AI Picks Up the Pace: From Strategy Shifts to Full Automation

AI is shifting from novelty to infrastructure, powering global systems, reshaping strategy, and igniting bold, sometimes unsettling visions. Today’s cues explore where that pace is accelerating, where it’s faltering, and what leaders are doing next.

Here’s what’s happening in AI today:

• AI is speeding up the global economy, but many companies aren’t built to keep up

• China showcases humanoid robots in a half-marathon, spotlighting physical AI ambitions

• Johnson & Johnson shifts its AI strategy, focusing on smaller, practical tools over big bets

• A new AGI startup wants to replace all human workers, sparking backlash and existential questions

• OpenAI’s newest models improve at reasoning, but hallucinate more, raising trust concerns

AI is powering an ‘always-on’ economy

The Cue:
AI is speeding up everything, markets, services, decision-making, and businesses that can’t match that pace risk falling behind

The Details:
The WSJ explores how generative AI is ushering in an “always-on” economy, where companies operate around the clock, responding to customers, markets, and data in near real time. From banks that use AI to monitor transactions 24/7, to manufacturers that dynamically adjust supply chains overnight, AI is collapsing traditional business cycles.

Leaders from McKinsey and Salesforce note that AI is shifting the pace of business from “timely” to continuous. But many companies aren’t ready: outdated systems, risk aversion, and cultural inertia are slowing down adoption - especially in highly regulated sectors like banking and healthcare.

Why it matters?

• AI is no longer just about productivity, it’s about temporal advantage

• The companies that win won’t just be the smartest, they’ll be the fastest

• Business leaders must rethink ops, staffing, and decision rights in a world where AI is always on

• Laggards won’t just be slower, they’ll be invisible

China puts humanoid robots to the test in a half marathon

The Cue:
In a public tech flex, China debuted humanoid robots in a half-marathon, showing off physical coordination, battery life, and national ambitions.

The Details:
During the recent WRC Marathon Challenge in Beijing, more than 10 Chinese-built humanoid robots walked a 13-mile course in what the event billed as the first endurance competition of its kind. The robots moved slowly and awkwardly, one finished in 5.5 hours, but it wasn’t about speed. It was a stage to show how far China’s domestic robotics sector has come.

The event was organized by China's Ministry of Industry and Technology and featured robots from top national players including Unitree and Fourier Intelligence. Engineers monitored battery endurance, gait, balance, and structural performance. The challenge highlighted a coordinated push to become a global leader in intelligent robotics, across both hardware and control systems.

Why it matters?

• hina isn’t just chasing language models, it’s heavily investing in physical AI

• Humanoids are being shaped as future co-workers, logistics bots, and public safety assets

• For businesses, this signals where industrial and service automation may scale next

• This is also soft power, China is publicly branding its AI leadership beyond the datacenter

Johnson & Johnson rethinks its AI game plan

The Cue:
J&J is shifting from big AI projects to more focused, department-level tools — after early bets didn’t deliver as expected.

The Details:
Johnson & Johnson’s Chief Data Science Officer, Najat Khan, told WSJ that the company is pivoting away from broad, foundational AI initiatives. Instead, it’s now targeting more tightly scoped applications, like document summarisation, clinical trial data parsing, and patient-facing insights. The shift comes after a period of internal experimentation with larger models that proved difficult to scale or integrate effectively.

J&J isn’t stepping back from AI, it’s just moving from ambition to execution. The company is building out a catalog of "approved" AI tools and taking a modular approach: embedding smaller AI capabilities into specific team workflows, rather than chasing transformative, all-in-one platforms.

Why it matters?

• This is a real-world case of AI strategy maturing from experimentation to utility

• Big players are discovering that practical wins often beat grand visions

• For leaders: success in AI may come from alignment, not ambition — start narrow, go deep

• Modular, decentralized AI deployment could become the dominant enterprise model

AI pioneer’s startup aims to replace ‘All Human Workers’

The Cue:
Ben Goertzel, a prominent AI researcher, has launched a startup with a bold and controversial goal: full human job automation.

The Details:
Goertzel, known for his work in artificial general intelligence (AGI) and past projects like SingularityNET, unveiled a new venture, Rejuve AI Systems, which seeks to develop an AGI-powered labor force that could eliminate the need for human workers. In his own words, the aim is to “replace all human workers, everywhere, as soon as possible.”

The startup is focusing on humanoid robotics, decentralized AGI training, and open-source infrastructure. While Goertzel frames the goal as liberation from labor, critics see echoes of techno-utopianism that downplay economic, ethical, and societal disruption. The announcement sparked debate across social media, academic circles, and industry leadership forums.

Why it matters?

• This is AGI ambition pushed to its sharpest edge, total labor automation

• It forces tough questions: Who controls these systems? Who benefits? Who gets displaced?

• For orgs experimenting with AI: there’s a fine line between efficiency and obsolescence

• The narrative matters, even bold visions can spook regulators, workers, and investors if framed poorly

OpenAI’s latest models are smarter, but also less trustworthy

The Cue:
OpenAI’s new models are better at reasoning, but they also hallucinate more, raising fresh concerns over reliability.

The Details:
According to TechCrunch, OpenAI’s latest generation of models (part of the GPT-4.1 family) show improvements in advanced reasoning tasks, including math, programming, and structured logic. However, internal tests revealed a concerning trend: these models also hallucinate (i.e., generate false but plausible information) more often than their predecessors.

The issue seems to be linked to model complexity, as reasoning improves, so does the tendency to “overcommit” to incorrect facts. This creates a challenge for users relying on AI for high-stakes or fact-sensitive outputs. OpenAI says it is actively tuning system prompts and fine-tuning methods to reduce hallucinations without sacrificing capability.

Why it matters?

• Better logic doesn’t mean better truth, AI accuracy is not a given

• For teams using LLMs in regulated industries or customer-facing workflows, trust boundaries matter

• Highlights the growing need for AI verification layers, audits, and fact-checking pipelines

• “Smarter” AI may also require “smarter” governance, and more human oversight, not less

What else is happening in AI today?

• Meta will start training its AI models on public content from EU users: despite pushback, the company claims it complies with GDPR, though critics say consent remains unclear.

• Huawei announces an AI hardware breakthrough: its new Ascend chip and compute cluster aim to rival Nvidia, signalling China’s push for AI self-sufficiency amid ongoing sanctions.

• Meta’s FAIR lab unveils five new AI projects: including models that mimic human learning and memory, reinforcing Meta’s long-term AGI ambitions through open research initiatives.

As the gap widens between AI ambition and readiness, the signal is clear: speed, scale, and trust are the new battlegrounds. Stay sharp, this week will move fast.

Like The Cue?
Share it with someone who wants fast, smart AI updates — without the fluff.