On AI, Impatience, and Our Short Time Horizons

The pace of AI development continues to march forwards but at the same time the noise surrounding it is becoming increasingly tiring. I spend a fair amount of time consuming financial news and, particularly during earnings season, the same familiar pattern repeats itself.

Each quarter the spotlight falls on capex forecasts from hyperscalers, forward guidance from chip and memory manufacturers, and of course the latest earnings from Nvidia. The figures change, the context shifts, but the reactions remain strikingly similar.

If capex projections soften, the narrative shifts to is the bubble about to burst? If spending remains elevated or increases, the framing flips to billions are being poured in but where is the payoff?  

The immediate response in markets and media commentary has become increasingly binary. Excessive excitement or excessive disillusionment. Aggressive buying or aggressive selling. After a few days the dust settles, and then, on cue, the cycle begins again a few months later.

AI is undeniably ushering in a new era. The pace of technical progress over the past few years has been extraordinary. And yet the prevailing tendency is still to judge its success, failure, or promise on quarterly cycles, as though anything of real consequence could reasonably be proven or disproven on a three month rolling basis.

“But it’s been three years since ChatGPT was released.”

Anyone who has operated inside large organisations or lived through genuine technology transformations knows how little that timeframe really represents. Three years is barely enough to introduce complex systems, let alone embed them into core workflows, change behaviours, and extract sustained value.

It feels like we’re treating time and patience as liabilities rather than prerequisites for the long term benefits and promise that AI can deliver. In a world that moves at ever faster speeds, perhaps that shouldn’t be surprising.

Ultimately our impatience says less about the advance of the technology and more about us.

Why We Struggle With Waiting

We live in a world that increasingly rewards immediacy. Fast feedback loops, instant metrics, and near real time validation.

From social media likes and engagement to financial markets and trading platforms, we’ve conditioned ourselves to expect immediate results, whether they’re positive or negative. Our expectations on timelines for transformation have seemingly fallen in line accordingly.

It seems high time preference has become the primary, dominant lens through which progress is now to be judged.

AI and many of its applications deliver rapid feedback and impressive short term results, which only reinforces the expectation that the overall transformation it’s bringing should move just as fast. Yet the most far reaching benefits require carefully considered upfront decisions, sustained investment, and patience before value becomes broadly visible.

We want answers NOW! Payback on investment NOW! Productivity boosts NOW!

Yet this also quickly runs into the wall of reality, colliding with the physical and organisational constraints that no amount of optimism can wish away.

AI requires infrastructure before utility. Infrastructure requires capital, energy, materials, people and time. What we want immediately is constrained by what can realistically be built, deployed and absorbed.

Our unquenchable need for neat explanations to everything further amplifies the challenge of realistic expectations and patience. Markets demand reasons for price movements. Commentators need narratives to justify conviction. Media needs headlines capture the public’s attention whether through intrigue or fear, optimism or pessimism.

Instead of seeking to get to the end state now a better question to ask is if the promise of AI was delivered immediately tomorrow would we be ready for it?

Adoption Isn’t a Technical Problem, It’s a Human One

Let’s play a hypothetical game and set infrastructure, energy and scaling constraints aside. We have unlimited compute, a revolutionary new low cost, continual energy source and a perfect supply chain for everything that powers and supports AI infrastructure. We’d still be left with a stubborn bottleneck that doesn’t come off an assembly line or require physical labour and investment to build.

People.

It seems that we’re both massively overestimating our readiness for change and underestimating the scale of change full scale AI adoption will introduce. There’s a reason full industries exist to support organisations in implementing change from organisational structural change through to adoption of new tools and frameworks.

We’re resistant. We’re habitual. We have early adopters and laggards. We build identities around how work gets done.

New ways of working don’t propagate simply because they’re available. They require sustained effort, trust, incentives, accountability, and a willingness to unlearn behaviours that may have served people well for years or decades. Even within a single organisation, that process can take years.

Now scale that challenge across entire industries and society.

We’re not just integrating a new system or rolling out another platform. If AI delivers on its promise, we’re reshaping how work is performed, value is created, and how organisations operate at a fundamental level. These are paradigm shifts, not system upgrades.

Seen through that lens, the pace of adoption becomes less a frustration and more of a necessity. Even with rapid technical progress, human systems move more slowly, and in many cases, they’re still trying to understand what this change actually means.

Taking A Longer View

We often want transformative outcomes without transitional phases. Structural change without disruption. Certainty without waiting. When those expectations aren’t met, we oscillate between hype and the burst of a bubble.

Do I expect the noise to subside? No. Markets will continue to overreact in both directions. Commentators will keep reaching for definitive answers. Narratives that sell will continue to crowd out those that ask for patience.

But for those willing to zoom out, to resist the urge to judge everything on a quarterly cadence, there’s a calmer and more realistic way to view what’s happening. AI isn’t dying, nor is it instantly paying for itself. It’s embedding itself gradually into how we work and how organisations operate.

In an age that seems increasingly allergic to waiting, perhaps the most grounded response is to accept that some changes are too large to be rushed, too complex to be summarised, and too important to be reduced to quarterly forecasts and profits.