Grabal

Weekly Talk #32 : Agentic World Models: Cross-Entropy Is Not Intelligence

Saturday, January 24, 2026 | 12–1 PM GMT

Most of what we call “AI progress” today is optimization under a risk-neutral illusion.

Cross-entropy. Log-likelihood. Expected loss. Expected reward.
All of them compute averages before asking whether the outcome is survivable, stable, or causally valid.

This is not a technical footnote.
It is a structural failure.

At AI Mali, we will make a blunt claim:
"Any agent trained only to minimize expected loss is unfit to act in the real world."

Why?

Data is not evidence. Acquisition across environments (biochemical reactions, physical systems, economic markets, digital ecosystems) is selective, incomplete, and often contaminated by incentives, omissions, or adversarial interference. Passive raw measurements alone cannot guarantee situational awareness.

Assimilation is critical. Observations must be integrated, reconciled, and contextualized before perception. Failing to assimilate the blockchained data pipeline (measurement → observation → perception → decision) renders downstream intelligence brittle.

Knowledge is not correlation. Representations that cannot answer counterfactuals or simulate interventions are epistemically hollow.

Training objectives hide tail risk. Averages wash away precisely the rare but high-impact events that determine real-world success or failure.

Deployment breaks IID assumptions. The moment an agent acts, the world reacts. Historical data no longer reflects current reality.

Usage creates feedback loops. Agents reshape their environment and invalidates yesterday's “alignment,” particularly in economic, digital, or ecological systems.
An agent that does not model causality across acquisition, assimilation, perception, and action is not intelligent.
It is a high-dimensional curve fitter with authority.
Cross-entropy (being an expectation under an implicit risk-neutral measure) is incoherent once agents influence their own data-generating process.
The uncomfortable question we will ask:

"What world does your model believe it inhabits, and what happens when that belief fails under intervention?"

Agentic World Models must internalize, inside the learning objective:
Causal structure across environment and event types (biochemical, physical, economic, digital):
Counterfactual reasoning,
Coherent risk measures,
Data acquisition and assimilation pipelines from measurement → observation → perception → decision.
Not after training.
Not in post-hoc safety layers.
Inside the model.

This talk transcends slogans: it confronts the structural limits of current AI.
It is about whether our current paradigm can survive contact with audio reality.

Agentic World Models at AI Mali:
Abandon averages. Embrace causality. Enable Risk-awareness. Survival demands it.



Register here : https://forms.gle/RebGz5TCMjUZsFya9

Talk #31 IA Audio & Inclusion Digitale

Building Meaningful AI-enabled Robotics for Africa: From Coopetition to Community Impact

AI-enabled Robotics in Africa requires a fundamentally different approach.

At AI Mali, our priority is to develop AI and blockchained technologies that deliver tangible value to local populations, not to pursue superficial demonstrations or promotional showcases involving virtual reality.

We emphasize practical, impactful AI-enabled robotics applications that address real needs in Malian cities and communities.

Examples include:

🚀 AI-assisted robotic systems for motorcycle engine cleaning, diagnostics, and assembly

👀 Autonomous robots for removing dust and waste from urban areas

➡️ Robotic systems for clearing water-drainage channels and roadside pipelines

❌ Disinfection robots designed for hospitals and public facilities

Our vision rejects the idea of unstructured competitions that often replicate foreign models without local relevance.

Instead, we advocate co-opetition, a blend of cooperation and competition that fosters mutual learning and collective problem-solving among participating schools, institutions, and innovators.

We want AI and robotics that matter for our population, not copy-paste exercises that fail to reflect our context.

For instance, thousands of motorcycles require maintenance in our cities every month; this is a practical domain where AI-enabled robotics-based co-opetition could directly support economic needs and workforce development.

Education and training programs must be aligned with the realities of our job markets.

Otherwise, we risk producing high-school graduates who are unable to secure employment in Mali and who are unprepared for advanced robotics programs abroad.

Since 2010, as a member of the AI & Robotics Center, we have not seen a single student trained locally who successfully passed the entrance evaluations for graduate programs in robotics in the United States or China.

This reality underscores the urgent need for a new, context-driven strategy.

Some of our papers in AI & Robotics:

➡️

https://dl.acm.org/doi/10.1145/3200947.3201015

🚀 https://ieeexplore.ieee.org/document/8394924

Questions? Reach out : https://www.aimali.ai/