By Damian Mathews & The Last Mile Team
Standard Chartered CEO Bill Winters made a comment in Hong Kong on Tuesday that has been circling the internet ever since. The bank announced plans to cut roughly 8,000 corporate support roles by 2030, replacing them with AI and automation. Standard restructuring. Nothing new in the financial industry.
But then he described what was happening in a way that turned the announcement into a global story.
“It’s not cost cutting. It’s replacing in some cases lower-value human capital with the financial capital and the investment capital we’re putting in.”
Lower-value human capital.
The reaction was immediate. Former Singapore president Halimah Yacob called the language disturbing and demeaning. LinkedIn lit up. By Wednesday, Winters was sending a memo to staff trying to walk it back. “I know this may be unsettling when reduced to simple headlines,” he wrote. “Where roles do fall away, it reflects changes in the work, not the value of our people.”
Too late. The phrase was out.
Coinbase CEO Brian Armstrong said something quite different when announcing his own 14% workforce reduction a couple weeks back. He’s not ditching low-value human capital, he’s “rebuilding Coinbase as an intelligence, with humans around the edge aligning it.” Same restructuring logic. Different lens.
One framing positions humans as the irreplaceable layer that gives the AI purpose. The other positions them as a line item being optimized out.
The substance of both decisions is similar. The downstream effects will not be.
In CX at least, the transition seems to be happening either way.
The companies who lead it well will keep their best people, attract the next generation of talent, and build cultures where AI gets embraced rather than sabotaged. The companies who lead it poorly will hollow out, watch their best people leave first, and end up with the people most likely to resist the change. Underneath the language question is a more important one we need to be honest about.
Knowledge workers who refuse to learn AI are in trouble. That’s real. The capability curve is steep, the adoption curve is steeper, and “I don’t want to learn this” is not a survivable position for the next ten years of office work. We wrote about this a few months ago in Operate in the World That Exists. The trajectory has not gotten kinder since.
But here’s the part that gets lost in the doom narratives: Knowledge workers who DO learn AI are not just safe. They are becoming the most valuable workforce in history. A person who knows what good looks like, who understands the business deeply, who has judgment built from years of experience, and who can now direct AI to execute their thinking at scale is worth more than they have ever been. Not less. More.
There is no such thing as lower-value human capital. Humans are intrinsically valuable. And AI makes their humanity more valuable, not less. But only us humans are doing the right tasks. And what’s right for humans to spend their time on is changing rapidly.
There are humans with AI fluency and humans without it. The first group is becoming extraordinary. The second is at real risk. The job of every leader right now, especially in CX where this transition is happening on the visible front line, is to help your people get to the first group as quickly as possible.
That starts with how you talk about them.
Kerry wrote about how we approached this transition internally in A1B: Customer Zero to AI-First. The whole thing rests on a phrase we kept coming back to: AI isn’t coming for your job. Someone using AI is.
Our teams get to be those people. They surround themselves with AI tools they helped design, and they replace their old job with a better one. That’s the entire point. Not human OR AI. Humans with AI, beating humans without.
Bill Winters will likely recover from this week’s headlines. The bigger question is whether his staff will trust his next message. The CX leaders reading this should take a hard look at how their organizations are talking about this transition. The substance is one thing. The story your people hear is another. Both matter, and the second one is harder to fix once you’ve broken it.
How are you talking about your team’s future?
— Damian
Here’s what went down this week.
Bleeding Edge
Early signals you should keep on your radar.
Google unveiled Gemini 3.5 Flash, Omni, and Spark at I/O this week, alongside a cut to the Ultra subscription. Gemini 3.5 Flash now hits frontier-level intelligence at $1.50 in and $9 out per million tokens with a one-million-token context window. Sundar’s roughly $190 billion capex bet looks less reckless when per-token pricing keeps collapsing this fast.
Dell launched PowerRack at Dell Technologies World, its first fully integrated rack-scale platform for AI and HPC. The turnkey unit bundles compute, networking, storage, and cooling, and can be running live workloads inside roughly six and a half hours of delivery. For CIOs tired of integrator math, a single-SKU AI factory could make on-prem inferencing competitive with the hyperscalers again.
Leading Edge
Proven moves you can copy today.
KPMG and Anthropic signed a global alliance to embed Claude inside KPMG’s Digital Gateway platform. All 276,000 KPMG employees get access, with tax clients and private equity firms first in line for the agentic workflows. Agent build cycles that used to take weeks reportedly now run in minutes, and Big Four consulting may not look the same for long.
xAI launched Grok Build, a coding agent and CLI aimed at professional software engineering work. The release pairs with new Grok Web Connectors for SharePoint, Outlook, GitHub, Notion, and Google Workspace, plus bring-your-own MCP servers. At $300 per month for SuperGrok Heavy access, xAI is pricing this for serious engineering teams, not the casual user base.
Off the Ledge
Hype and headaches we’re steering clear of.
An Oakland jury dismissed Elon Musk’s suit against Sam Altman and OpenAI in under two hours. The advisory panel found his breach-of-charitable-trust claims fell outside a three-year statute of limitations, and Judge Gonzalez Rogers adopted the verdict immediately. Musk called it a “calendar technicality” and vowed to appeal, but he just lost his cleanest legal lever to slow OpenAI’s commercial pivot.
Meta began cutting 8,000 jobs this week, with Singapore staff hit first at 4 a.m. local time. Another 6,000 open requisitions are cancelled and roughly 7,000 employees are being reassigned into AI-focused pods under the company’s $115 to $135 billion capex plan. Zuckerberg has now eliminated 25,000 roles since 2022, and the message to anyone outside Superintelligence Labs is hard to misread.
