Data Science and Governance

“Musk Will Get Richer, People Will Get Unemployed”: Hinton on AI

Massimo2025-12-01033 views

The recent interview with Geoffrey Hinton — often called the “godfather of AI” — delivers a stark warning: the runaway acceleration of AI could enrich a small elite while displacing massive numbers of workers. Hinton argues that humanity is constructing something far more dangerous than a mere tool: a super-intelligence that, in a decade or so, may leave us obsolete. His metaphor of an alien invasion arriving in ten years is chilling — unless we fundamentally rethink how we coexist with artificial minds. Enterprises investing trillions today are betting on replacing human labour, not enhancing it. Hinton notes that while some firms (such as Anthropic and Google DeepMind) claim to take safety seriously, many others prioritise dominance and profit. In a world where machines outsmart their creators, notions like “hiring more” or “creating new jobs” become hollow. Unless social structures and governments intervene, the result may be vast unemployment, intensified inequality, and a future where only a few — the “Musk-class” of investors — prosper.

Google Nested Learning – AI memorizes like our brain

Massimo2025-11-25069 views

Google Research’s Nested Learning paradigm reframes the age-old dichotomy of architecture vs optimiser into a unified, hierarchical system of nested learning loops. By deploying multiple modules updating at varied frequencies, the continuum memory system enables long-context retention and mitigates catastrophic forgetting. Their HOPE architecture exemplifies this, outperforming standard models in continual-learning tasks. For AI agents, this suggests a transition from static tools to evolving systems. The real frontier isn’t larger models — it’s learning better models.

SpikingBrain: a revolutionary brain-inspired Chatgpt made in China

Massimo2025-09-180357 views

The Chinese SpikingBrain is a new family of brain-inspired large language models that reimagines how AI can process information more efficiently. SpikingBrain models adopt a biological principle: neurons remain idle until an event triggers them to fire. This event-driven design reduces unnecessary computation, cuts energy use, and enables faster responses. SpikingBrain achieves over 100× speedup in “time to first token” for sequences up to 4 million tokens. Energy consumption drops by 97% compared to traditional LLMs.

AI-Native Developers: The New Divide in Software Engineering

Massimo2025-09-050106 views

AI-native development is creating a paradox: soaring demand for AI engineers, and unemployment for traditional CS graduates. Businesses want developers skilled in prompting, RAG, evals, and agentic workflows—yet most universities still teach 2022-style coding. The best engineers today pair computer science fundamentals with cutting-edge AI fluency. Like the shift from punchcards to terminals, AI-native coding is becoming the new baseline. Those who adapt will thrive. Those who don’t risk obsolescence.

Markov Chains, MDPs, and Memory-Augmented MDPs: The Mathematical Core of Agentic AI

Massimo2025-09-010147 views

Markov Chains, Markov Decision Processes (MDP), and Memory-augmented MDPs (M-MDP) form the mathematical backbone of decision-making under uncertainty. While Markov Chains capture stochastic dynamics, MDPs extend them with actions and rewards. Yet, real-world tasks demand memory—this is where M-MDPs shine. By embedding structured memory into the agent’s state, M-MDPs enable agentic AI systems to reason, plan, and adapt across long horizons. This blog post explores the mathematics, technicalities, and the disruptive role of M-MDPs in modern AI architectures.

Agentic SEO – When AI Shops for You: How Autonomous Agents Are Rewiring E-Commerce

Massimo2025-09-01094 views

AI agents are overtaking search: shopping visits driven by generative AI surged 4,700%, while retailers like Walmart deploy “super agents” that guide purchasing end-to-end. But agents bring risks—less visible brands, opaque decisions, and emerging trust deficits. To thrive, businesses must reorganise for agent interaction: reengineer SEO through semantic structures, track agent-led conversions, and build accountability into the agent flow. In short, we’re moving into a world where your brand needs to speak agent, not just user.

Why 90% of Generative AI Projects Fail — and How to Avoid Becoming a Statistic

Massimo2025-08-230203 views

MIT’s 2025 report finds 95% of enterprise GenAI pilots fail, blocked by a “learning gap.” Tools that don’t adapt, remember, or integrate into workflows stall, while adaptive, embedded systems cross the GenAI Divide. The winners are startups, not big Companies then, focusing on narrow but high-value use cases, embedding in workflows, and scaling through learning. Again, generic SaaS tools and in-house builds fail. Leaders must focus on strategic partnerships with startups, adaptive systems, back-office ROI, and agentic readiness to ensure AI delivers measurable impact—not hype.

Inside the AceReason-Nemotron LLM of NVIDIA

Massimo2025-05-270269 views

AceReason-Nemotron is a groundbreaking AI model developed by NVIDIA that redefines how we train large language models (LLMs) for math and coding tasks. Unlike traditional models trained through distillation, AceReason uses reinforcement learning (RL) guided by strict verification and binary rewards to push reasoning capabilities further—particularly for small and mid-sized models. Starting with math-focused RL and later fine-tuning on code, the model shows impressive cross-domain generalization: math-only training significantly boosts code performance before even seeing code-related tasks. The new strategies help AceReason-14B outperform strong baselines like DeepSeek-R1-Distill, OpenMath-14B, and OpenCodeReasoning-14B on benchmarks like AIME and LiveCodeBench. It even approaches the capabilities of frontier models like GPT-4 and Qwen-32B in specific reasoning domains. For AI researchers and recruiters, AceReason is a compelling case study in how reinforcement learning—when combined with rigorous training design—can unlock reasoning in smaller models that once seemed exclusive to ultra-large systems.

S1: The Open-Source AI Model Challenging Industry Giants

Massimo2025-02-210278 views

The landscape of AI language models has been dominated by proprietary systems requiring massive computational resources. However, a new contender, S1, is redefining what’s possible with efficient training techniques and open-source transparency. Developed by researchers from Stanford University, the University of Washington, and the Allen Institute for AI, S1 showcases a novel approach to improving reasoning capabilities without exponential increases in computational cost. It seems the next breakthrough will come to the optimization of the reasoning methodologies. I envision two different engineering paths we should follow to better inferencing LLM models: prompt engineering reasoning engineering (I wrote a post about this). Technical Overview S1 employs a test-time scaling approach, allowing the model to enhance its reasoning capabilities dynamically during inference rather…

The Rise of Reasoning Engineering: optimizing reasoning beyond prompting

Massimo2025-02-150263 views

Reasoning Engineering is the next frontier in AI, optimizing how AI agents collaborate to enhance structured reasoning rather than relying solely on prompt engineering. This approach designs reasoning models, where multiple agents interact to refine inference depth, self-awareness, and response modulation.

For instance, to simulate shyness, an AI system combines emotional perception, self-consciousness modeling, uncertainty processing, and inhibition mechanisms. A RoBERTa model detects emotional triggers, a Bayesian agent estimates social scrutiny, and a GPT-4-based processor introduces hesitation. Finally, a Transformer inhibition model restricts emotional output, ensuring reserved, self-conscious responses, replicating human-like shyness in AI-driven interactions.

AI and the Death of Critical Thinking: A Looming Crisis

Massimo2025-01-090252 views

How Our Reliance on Artificial Intelligence Risks Eroding Human Reasoning and Shaping a Passive Future Artificial intelligence (AI) is heralded as a transformative force, reshaping industries and augmenting human capabilities. Yet, emerging research warns of a darker undercurrent: the erosion of critical thinking. A study by the Swiss Business School reveals a troubling pattern—frequent AI tool users, particularly younger individuals, exhibit markedly lower critical thinking scores. The problem is cognitive offloading, where people rely on AI for mental tasks, reducing their own thinking efforts. This dependency is most pronounced among younger users, who, raised in an AI-saturated environment, often rely on algorithms to answer questions, make decisions, and even form opinions. The trade-off is stark: efficiency and convenience at the expense of…

A New Frontier in AI: Introspection and the Changing Dynamics of Learning

Massimo2024-11-210290 views

Extract knowledge from LLMs for training. Introspection might change the dynamics of learning The landscape of training language models (LLMs) is on the brink of a dramatic transformation. Insights into how LLMs can introspect—access and utilise their own internal knowledge—promise to reshape the costs and strategies of AI development. The implications are profound: the cost of training could collapse in the coming months, accelerating innovation and democratising access to cutting-edge AI technologies. A Past Vision Revisited: Rethinking How LLMs Learn Years ago, I delved into the challenge of optimizing how LLMs acquire and refine knowledge. The central question was whether we could fundamentally alter the training phase itself, bypassing traditional methods that rely on ever-larger datasets and increasingly computationally expensive…