Mo Gawdat says artificial intelligence is already rewriting the rules, most people have not noticed

Mo Gawdat, an Egyptian software engineer and former chief business officer at Google X, delivered a blunt message in a presentation over the last 24 hours: artificial intelligence (AI) is not a future disruption. It is a present one. He argues that it is already reshaping geopolitics, economics and the day-to-day mechanics of modern life, while much of the public still treats it as a novelty.

Gawdat’s core claim is simple. AI has been building for years, largely out of view, and it is now reaching a point where it can improve itself. If that feedback loop holds, it changes the pace of everything. It also changes who keeps up.

The long build, then the sudden lurch

Gawdat frames AI as a project that has been in motion since the early 2000s. In his telling, the current moment looks sudden only because most people are seeing it properly for the first time.

He points to recent breakthroughs as evidence of how far the field has moved. He cites work he describes as contributing to a cure for cancer, a deeper understanding of the human genome, and even a redefinition of mathematics. Whether you agree with the framing or not, his intent is clear. He wants his audience to stop thinking of AI as a chatbot layer on top of the internet.

He is arguing that AI is becoming a general tool for discovery. If that is right, the impacts will not stay confined to the technology sector.

When the machine improves the machine

The most striking section of the talk is his focus on AI improving its own performance. He describes a recent episode where an AI system identified weaknesses in a key piece of traditional mathematics used in computing, matrix multiplication.

Matrix multiplication is not a niche concept in this context. It underpins a large share of modern computing, including the training and running of AI models. Gawdat says the system flagged issues in an approach that has been relied upon for 56 years, then produced new mathematics that improved Google’s AI performance by 20% to 23%.

That type of claim is designed to land a specific point. If AI can meaningfully optimise the fundamentals beneath AI, progress can compound. It also becomes harder for non-specialists to track what is changing and why.

The practical result is not just faster models. It is faster cycles of capability.

Cities are the next interface

From there, Gawdat moves outward. He argues that AI will not remain trapped in screens. He expects it to reshape cities first, because cities are systems full of repeatable patterns.

Urban mobility is his lead example. If AI becomes reliable at perceiving and predicting real-world movement, it can coordinate transport, logistics and services. That vision tends to come with a familiar set of applications: self-driving vehicles, more automated public infrastructure, and software that orchestrates physical assets in real time.

This is the point where AI stops being a workplace tool and becomes a civic one. It becomes the control layer for the built environment. That has benefits, but it also raises immediate questions about governance, safety and accountability.

Gawdat does not treat those questions as distant. He treats them as deployment issues.

Commerce will be rebuilt around prediction

He also argues that commerce is set for a reset. In his framing, AI changes how markets work because it changes how prediction works.

AI can price risk differently. It can personalise offers at scale. It can reduce friction in transactions and service. Over time, it can compress the cost of decision-making across supply chains, marketing and customer support.

This is not just automation as cost cutting. It is automation as competitive advantage. It will favour organisations that can integrate AI into their workflows quickly, and it will punish those that cannot.

That dynamic rarely stays confined to one industry. It tends to cascade.

The robot arrives quietly

Gawdat’s most concrete consumer-facing forecast is the arrival of capable home robots. He suggests that robots will become both cheaper and more useful, moving from novelty devices to practical household systems.

He notes that some models are already available at around $9,000. His larger point is not the price tag today. It is the direction of travel. Once the software becomes reliable and the hardware becomes mass-produced, the economics change.

A robot that can handle chores, simple care tasks, or home management is not just a gadget. It changes time budgets, labour markets and expectations of what a home can do.

It also forces a cultural shift. People will need to decide what they are comfortable delegating to machines, inside private spaces.

The gap between reality and perception

Running through the presentation is a consistent theme: the public’s mental model of AI is lagging the real state of development.

Gawdat’s warning is not only about capability. It is about readiness. If AI is already shaping geopolitics and economics, then misunderstanding it becomes a risk in its own right.

In practical terms, that can look like policy that moves too slowly, businesses that bet against change, and individuals who treat AI as optional knowledge. If AI becomes embedded in transport, commerce and domestic life, it will not stay optional for long.

What to take seriously, even if you doubt the details

You do not need to accept every claim in the talk to take the structure of the argument seriously.

First, AI is already a broad disruptor, not a single product category. Second, the integration story matters as much as the model story. Third, if AI can accelerate its own improvement, timelines shrink.

The key question is not whether AI will touch cities, commerce and homes. The question is how quickly it happens, who controls the systems, and what failures look like when they occur.

Gawdat’s message is that the world is on the cusp of this shift. The harder message is that the cusp may already be behind us.