gshc2020.com

The organizations getting this right are not asking how few people they need. They are asking what their people need to become.

I have a specific vantage point on this. At APPLY, we aim to have every new client engagement start with a minimum of 50% agents working alongside human practitioners from day one. Not as an experiment, but as the delivery model. That means we have had to answer the talent question in practice, not in theory. And what we have learned is that the shift AI creates is not subtraction. It is transformation.

The trajectory of that transformation follows a pattern we have mapped explicitly. Today, most organizations are human-led with AI handling a thin layer of routine execution. In the near term, that flips into an assisted model: AI takes on more of the execution, and humans focus on judgment and direction.

Further out, agentic delivery becomes the norm, with AI doing the heavy lifting and humans guiding outcomes rather than producing them. The North Star is autonomous delivery, not a replacement story, but a leverage story. Total capacity expands at every stage. As AI absorbs more at the bottom, humans move up the value stack. The implication for talent isn’t smaller teams. It’s teams with a higher floor for what good looks like.

What Actually Changes

When agents handle execution, content generation, data processing, workflow execution and routine decision-making, the human role does not disappear. It elevates. We worked with a retail client whose merchandising team was spending the majority of their time producing weekly performance reports. Once agents handled the assembly, those same people were making ranging decisions that had previously been delegated to gut feel. The output was better. So was the job. The work that is left when AI handles execution is not less. It is harder, and it matters more.

Most people assume AI simplifies work. In our experience, the baseline expectation for what a human contribution looks like goes up when the mechanical parts are handled. That is uncomfortable at first. It is also where the real growth is.

The skill that separates people in this environment is not speed. It is judgment: knowing when the agent is wrong, being able to diagnose why and having enough genuine domain expertise to course-correct. The gap is not access to technology. It is the quality of the question being asked and the credibility to evaluate the answer.

The Ingenuity Premium

There is a pattern I have noticed across our engagements. As AI takes on more at the bottom of an organization, such as the repeatable, the predictable and the high-volume, human ingenuity expands at the top. The ceiling on what a person can contribute gets higher when the floor is handled.

What ingenuity actually looks like in an agentic environment is the ability to define a problem precisely enough that a system can act on it, to recognize when the system’s output is technically correct but strategically wrong and to hold accountability for results that were largely produced by something you directed rather than something you built yourself. That combination of domain mastery, systems fluency and outcome ownership is genuinely rare. The organizations building it deliberately are pulling away from those waiting for it to develop on its own.

This shows up in how we think about hiring and onboarding at APPLY. We are not looking for people who are fast at the tasks agents can handle. We are looking for people who can work at the level the agents make possible, who can direct, evaluate and own the outcome of work that is partly or mostly generated by systems they are responsible for. That is a different skill set and it takes deliberate effort to build.

The organizations that understand this are investing in human capability alongside their AI investment. They are training their people not just to use the tools but to work at the level the tools make possible. And they are not waiting for the workforce to adapt on its own; they are designing for it.

The Talent Needed In Five Years

When I think about the talent we will need five years from now, most of those roles do not exist today in their current form. We will need practitioners who are masters of their craft and fluent in agentic systems simultaneously, people who can set the direction a dozen agents execute, who can tell when the agent is wrong and why and who can design experience ecosystems rather than individual touchpoints.

The leadership question is harder still. Executives leading AI-native organizations will need a genuine tolerance for systems they do not fully control, a bias toward outcome measurement over process management and the conviction to hold their organizations accountable to a standard the rest of the industry is still debating.

That profile is rare today. Building it deliberately is one of the most important investments any organization can make right now, not because AI is taking jobs, but because AI is raising the bar on what the best jobs actually require.

The future of work is not fewer people doing the same things more efficiently. It is different people doing different things at a level that was not previously possible. The organizations that understand this are building a genuine competitive advantage. The ones treating it as a cost story will find themselves short not just of the right technology, but of the people capable of using it well.​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?

It’s tempting, in an industry built on differentiation, to greet a launch like this with caveats. I want to do the opposite. Coinbase just did something this industry has needed for years: It made tokenization mainstream.

This is the market education nobody else could buy.

Tokenization has spent most of its life as an insider’s topic, debated at conferences and written about in trade publications that never reach the people who decide where capital goes. A household brand, trading under a ticker millions of retail investors already recognize, just told the market that owning a share of Nvidia onchain is real and usable today.

That reaches audiences no RWA-focused startup could reach alone. Every credible entrant adds legitimacy to the category: Regulators pay closer attention, institutional allocators ask their own teams why they haven’t looked at this yet and retail investors who’d never search “real world asset tokenization” now ask their exchange app what “tokenized stock” means.

Wrappers are the on-ramp, not the destination.

Most tokenized equity products today, including Coinbase’s, work the same way: A regulated custodian holds the underlying shares, a special purpose entity issues a token representing a claim on them and the token trades onchain while legal title sits upstream in a trust structure. It’s a wrapper, a well-built one, but a wrapper nonetheless, representing ownership rather than serving as the instrument of it.

That’s not a criticism. It’s a description of where this industry sits in its own development, and it’s exactly the right place to start. Stablecoins followed the same arc, as wrapped claims on bank deposits, before the market matured enough to demand more sophisticated structures underneath.

The more people who interact with a wrapped tokenized stock, the more who will eventually ask why ownership still routes through a trust deed and a custodian’s balance sheet instead of existing natively onchain from the moment it’s authorized. That question only gets asked at scale once people have experienced the wrapper first. Coinbase isn’t competing with that future. It’s building the audience for it.

Bank charters are another door into the same building.

Coinbase isn’t the only major player redrawing this map. On August 14, 2026, the Office of the Comptroller of the Currency (OCC) granted conditional approval for World Liberty Trust Company, National Association, the banking arm of Trump family-backed World Liberty Financial, to operate as a national trust bank that would issue its USD1 stablecoin, manage the reserves backing it and provide digital-asset custody services under federal supervision.​

That structure matters more than the headline suggests. Money transmission is regulated state by state, and most crypto platforms spend years assembling money transmitter licenses one jurisdiction at a time. A national bank charter changes that math. Federally chartered banks can benefit from National Bank Act preemption, which can displace state licensing requirements when those requirements prevent or significantly interfere with the exercise of federally authorized banking powers.

In June 2026, OCC Interpretive Letter 1192 specifically confirmed that national trust banks are not required to obtain state money-transmitter licenses for federally authorized activities, including certain payments-related activities. The interpretation was discussed by law firms including Gibson Dunn and Norton Rose Fulbright.​

World Liberty is joining, not originating, this pattern. In December 2025, the OCC conditionally approved national trust bank applications from Ripple and approved applications from BitGo, Fidelity Digital Assets and Paxos to convert their existing state trust companies into national trust banks.

​One federal charter beats 50 state licensing regimes. That’s five major entrants choosing banking infrastructure as their on-ramp into regulated onchain finance, alongside Coinbase choosing exchange infrastructure as its own.

Different door, same building: Exchanges prove the demand side, while bank charters prove the rails side, that money movement can run through one federally supervised structure instead of a 50-state patchwork. Neither gets the industry to native issuance alone, but both are necessary groundwork for it.

Every champion builds the runway for what comes next.

I coined the term RWA+ to describe the next phase of this industry: infrastructure where the token is the security from the moment it’s created, not a receipt layered on top of one, moving through rails federally supervised by design rather than assembled state by state after the fact.

That vision only becomes reality if enough capital, attention and regulatory comfort flow into tokenization first, on both the demand side and the rails side. Coinbase, World Liberty and the other major players moving equities, stablecoins and custody onchain in 2026 are all doing that work simultaneously, from different starting points.

The tokenization industry doesn’t need fewer champions right now. It needs more of them, moving more asset classes onchain, building more of the underlying rails, faster than anyone expected a couple of years ago. Every wrapped stock that trades successfully today, and every charter proving regulated money movement can run through a single federal structure, makes the case for native issuance easier to make tomorrow.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?

Then ask which business outcomes that spend produced, which workloads generated the growth or whether the same results could have been achieved for materially less. Most executives don’t have the answer to those questions. In fact, according to a PwC survey from January 2026, only 12% of CEOs say that AI has delivered both cost and revenue benefits for their organizations.

While many companies diagnose this as a reporting failure and send the problem to finance, I believe the more likely culprit is deploying a production dependency on an architecture that was never designed to be operated in this way.

Cloud was never really about remote servers.

Cloud was the default answer to nearly every enterprise technology question for over a decade. While the initial benefit was moving servers elsewhere, the real strategic value came from creating a programmable operating model: on-demand infrastructure, standardized interfaces, automation and, critically, a common control plane to see and govern everything underneath.​

AI is reaching a similar inflection point. The early phase of enterprise AI adoption was model-centric: compare benchmarks, choose a provider, build a proof of concept and measure response quality. ​

Now, a single production workflow may call several models, retrieve governed data, invoke tools, maintain memory, apply policy, verify outputs and escalate exceptions to a person. An agent may invoke another agent. One user request can generate dozens of inference calls across multiple systems. The model is just one component in a long chain, and most enterprises still lack control over the rest.​​​

The tooling you own cannot see the full picture.

Many organizations still view AI as a portfolio of models, co-pilots and pilots. That worked when AI produced drafts and suggestions, but it is too narrow now that AI powers customer service, software delivery, risk, operations and decisions.

What businesses are really building is a platform coordinating models, agents, data, accelerators, memory, policy and human accountability in real time. Most enterprises are assembling it by accident, but AI will only become a viable enterprise platform when intelligence can be operated intentionally.​​

​There are a few factors preventing executives from answering the question necessary to oversee AI: “How do we operate intelligence across the business at acceptable cost, latency, reliability and accountability—and prove it?”​​

Resource control planes manage resources, but AI has to manage outcomes.

Existing control planes are built around instances, containers, storage and networks. AI introduces a different unit of production: useful work.

The same task can run on different models, inference engines and hardware, each with different quality, latency, energy and cost profiles. Provisioning capacity is not the same as deciding how intelligence should be produced.

Application monitoring stops at the boundary that matters.

Traditional observability shows whether a service is up, how much compute it uses and where latency occurs. Operating AI requires deeper context: which model produced the answer, which prompts and retrieval steps shaped it, where time was lost, whether a cheaper route could meet the same service level and whether policy was enforced.

Without that lineage, enterprises can measure AI activity but cannot explain or improve it.

Static architecture cannot track moving economics.

Model prices fall, new accelerators arrive, serving engines improve and workloads shift. A configuration that works today may be wasteful months later.

Since any fixed design will become outdated, AI infrastructure must benchmark, route and adapt continuously—becoming self-optimizing rather than permanently hand-tuned.

Four Disciplines Of Operating Intelligence

While these challenges cannot be solved overnight, there are a few steps that executives can take now to get visibility into the impact AI is having on their organization:​​

1. Adopt a unit of production that means something to the business.

Tokens are a technical measure, not a business outcome. A cheaper model that creates rework is more expensive, and a faster model that breaches policy is not productive.

Leaders need to adopt metrics such as cost per resolved case, accepted code change or completed task. Optimize the full cost of achieving an outcome, not the cost of generating text.​

2. Compile policy into execution.

Governance cannot live only in review boards and documents. It must become runtime behavior: which data an agent can retrieve, which tools it can call, which actions require approval, which models are permitted for regulated data and what evidence must be retained.

The safe path should be the default, enforced by the platform rather than individual teams.

3. Instrument the whole path, not the layers.

Optimization requires visibility from the request through the model and down to utilization, memory pressure, network delay and storage behavior.

When each layer is measured separately, teams optimize locally and move the bottleneck rather than remove it. Shared lineage is also what converts an AI incident from a mystery into an investigation.

4. Make every workload teach the next one.

Each request generates signals about latency, quality, cache behavior, contention, tool success and business value. These should feed a loop that improves model selection, workload placement, memory allocation, demand prediction and cost optimization.

This is where durable advantage accumulates. Rather than focusing only on the model, build the operating knowledge that determines how every model, agent and accelerator should be used.​

What Comes Next

The enterprise AI platform will not replace everything with a single product. Instead, it will emerge as an operating layer connecting what already exists: a system of record for AI workloads, a policy engine, full-stack telemetry and an optimization loop matching business intent to the right models and infrastructure.​

As models keep improving and become interchangeable for many tasks, capacity will grow. Organizations can differentiate themselves by developing the ability to turn business intent into reliable AI execution—safely, repeatedly and economically.​​

Cloud made infrastructure programmable. The next transition makes intelligence operable. Companies that recognize it early will build an operating advantage from every AI workload while others are still comparing benchmarks.​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?

Players certainly noticed this and made their feelings clear, even before Marathon launched. At the time, Bungie responded to concerns about its steep, unwelcoming learning curve to say that while the game was a “dangerous climb,” its difficulty could be overcome. The Sony-owned studio tried to make the point that players should play the long game, slowly improving their base stats and gaining access to better items by completing contracts and progressing faction upgrades. But unfortunately for Bungie, many players were immediately put off by the constant failure they experienced simply trying to get to grips with the game. The release of brutally difficult and complex endgame raid, Cryo Archive, in April only doubled down on the sense that Marathon was built for only the most hardcore of shooter fans.

Even former professional Counter-Strike player, Shroud, expressed concern about Marathon’s difficulty, saying that while Cryo Archive offered an incredible experience, it was too hard for casual players. "Cryo Archive is insane,” he said in April. “It's the most elaborate extraction shooter map I've ever seen in a game ever. The loop that they made is truly something special. The problem is, is it too elaborate? Is it too complex? Is it too much of a grind? Is your 9-5 grandma and grandpa going to be able to do it? I don't know."

Now, in detailing radical changes coming to Marathon as part of Bungie’s bid to save the game, developers have talked about what went wrong.

“Launch didn’t go the way we wanted it to,” Kevin Yanes, senior design lead, said in a new video. “We all built a game that we believed would resonate. But I think when you’re so close to something it’s hard to see the bigger picture objectively. And so we made a game that was overly difficult I think for a lot of folks. And that constant loop of failure I don’t think spoke to as many people as we hoped it would.

“While death is the first step for many players, death was the only step.”

Game director Del Chafe III called out Marathon’s PvP, which he called “sweaty.”

“One of the things we learned at launch is that our game is pretty niche,” he said. “It’s for sweaty people who like PvP. And, to be fair, a lot of us on the team like those experiences.”

“We all thought we were gonna bring extractions to the masses, and we didn’t, right?” Yanes added. “We built very much an extraction shooter as it was, as you’ve ever seen one.”

The question now for Bungie and for Sony is, how can the developers turn Marathon around? Certainly part of the plan is to make it more accessible, but adding dedicated PvE modes is the key. Bungie will also add a team deathmatch mode, and expand Perimeter, essentially its tutorial level, to be closer in scope to other full zones.

Bungie's goal seems to be limiting the friction that new players feel and letting them filter themselves into more familiar gameplay types. That should, in theory, let people ease themselves into the traditional extraction mode that's proven so divisive.

The changes to Marathon were revealed in the same breath as Bungie announced a surprise return to the world of Destiny 2. Having initially ditched the game, Bungie will now unvault content, although it is not returning to full live service support. Bungie has said it’s got more Destiny in its future, though. What this turns out to be remains to be seen.

Last week, Bungie denied a rumor that it planned to merge the Destiny and Marathon IPs amid continued questions about the future of both franchises. Based on this week’s news, the plan is to soldier on with Marathon, at least into 2027. The spectre in all this is of course Sony’s ongoing commitment to Marathon. It has backed the game publicly, but some fans are worried about its future. Steam concurrents do not paint the full picture of a multiplatform game, but it’s worth noting that Marathon had a 24-hour peak of 3,285 on Valve’s platform, down from an all-time peak of 88,337 set when the game launched in March.

Wesley Yin-Poole

Director, News

Wesley is Director, News at IGN. Find him on Twitter at @wyp100. You can reach Wesley at [email protected] or confidentially at [email protected]. %!s()

Due in 2027, Day Drinker is Depp's second upcoming project to be backed by a major studio, following hot on the heels of Ebeneezer, a starry retelling of Charles Dickens classic A Christmas Carol that arrives in theaters this November.

While Depp has been busy with a few smaller projects, the Pirates of the Caribbean star has not appeared in a mainstream Hollywood movie for years, following a series of high-profile legal battles with his ex-wife Amber Heard.

Now, however, Depp appears to be mounting a comeback, via movies which feel well chosen for his current status. In Ebeneezer, Depp will play a hated figure who learns the error of his ways and ultimately ends up beloved by all. And in this first look at Day Drinker, Depp plays a mysterious alcoholic who seems to become entangled with the law while helping expose a criminal operation.

"I've made a living out of misbehaviour," Depp's character says in the trailer as he's cuffed and forced to the ground, before he affords the camera a little smile. "I see no tangible reason to stop now. Don't you sometimes want to misbehave?

"My story? Perhaps it's a tale of revenge," Depp's character continues, apparently in an Irish accent. "Or perhaps something worse. What if all this is a ghost story? Or maybe I'm just another day drinker."

"You thought I was f***ing done?" he concludes at the trailer's end. "I'm done when I say I'm done."

Day Drinker stars Depp as the eponymous alcoholic, alongside Madelyn Cline and Penelope Cruz. Noteworthy talent behind the camera includes The Amazing Spider-Man director Marc Webb (looking for his own comeback after helming Disney's widely-panned live-action Snow White), and Fast X co-writer Zach Dean. The movie is currently set to launch on March 26, 2027.

One of the big questions surrounding Depp is whether he'd return for a potential new Pirates of the Caribbean movie. Last month, Pirates of the Caribbean producer Jerry Bruckheimer reignited hope of Depp reprising his iconic role as Captain Jack Sparrow one more time while issuing a fresh update on Disney's long-in-development plans for Pirates of the Caribbean 6. Bruckheimer confirmed that conversations were underway with Depp for the actor to potentially return as the infamous pirate.

Bruckheimer has long expressed interest in Depp donning his Captain Jack Sparrow costume once again, even when it seemed unlikely, as Disney mulled other, alternative routes forward for the franchise. Depp's past legal issues with Heard saw Disney largely distance itself from the star, leading to some choice words from Depp in response. Meanwhile, long-gestating plans have placed Barbie star Margot Robbie at the forefront of Pirates 6 if and when it does ever materialize.

Image credit: Lionsgate/YouTube.

Tom Phillips

News Editor

Tom Phillips is IGN's News Editor. You can reach Tom at [email protected] or find him on Bluesky @tomphillipseg.bsky.social

Get IGN's Biggest Stories FIRST. Set Us as Your Source.Add Source %!s()