Inference, not prediction — Prof. Michael I. Jordan on what modern AI is still missing
TL;DR
Professor Michael I. Jordan critiques the hype around AGI and prediction-based LLMs, arguing that modern AI lacks economic and social thinking; he advocates for 'inference' systems grounded in game theory and market dynamics that respect human agency and create collective value.
🎭 The AGI Hype Problem 3 insights
AGI is a distortionary PR term
Jordan dismisses Artificial General Intelligence as marketing language that confuses young researchers and diverts attention from meaningful engineering challenges.
Alarmist rhetoric demoralizes young engineers
Industry leaders' false dichotomy of imminent superintelligence or existential doom discourages 20-year-olds from building helpful technology by suggesting all important problems are either solved or too dangerous.
Anthropomorphizing intelligence is harmful fiction
Describing systems as 'understanding' or mimicking human cognition constitutes unnecessary science fiction that distracts from practical mathematical and economic frameworks.
💰 Intelligence as Economic and Social 3 insights
Intelligence requires game theory and markets
True intelligence is contextual and social, emerging from aggregation, culture, and economic interaction rather than isolated prediction algorithms.
Systems must respect human producers and consumers
AI rests on data from billions of humans and should serve billions, requiring formal mathematical models of value exchange that return economic benefits to data creators.
Technology should create jobs, not just automate
The goal must be building economic ecosystems that generate opportunities for human talent, collaboration, and creativity rather than merely providing 'secretary on your shoulder' automation.
⚙️ Engineering Discipline and Accountability 3 insights
AI lacks foundational engineering principles
Unlike chemical or electrical engineering built on Maxwell's or Newton's equations, modern AI relies on ad-hoc gradient descent without theoretical guardrails for safety or scalability.
Current systems are prediction, not inference
Large language models remain statistical prediction engines rather than reasoning systems capable of economic inference needed for healthcare, finance, and transportation infrastructure.
Accountability requires outcome-based explanation
For decisions like loan denials, systems should explain outcomes by referencing similar cases and embeddings rather than uninterpretable internal neural circuits.
Bottom Line
Replace the current hype-driven prediction paradigm with economically-grounded inference systems that mathematically model human agency, market dynamics, and collective value creation at global scale.
More from Machine Learning Street Talk
View all
The Benchmark With No Instructions — Tufa Labs (ARC-AGI-3)
Researchers from Tufa Labs dissect the ARC-AGI-3 benchmark, which evaluates AI agents on learning abstract goals from raw pixel inputs without instructions. They reveal that while LLM-guided coding agents currently outperform pure reinforcement learning approaches, success on the benchmark may exploit latent training priors rather than represent genuine abstraction acquisition, raising fundamental questions about the relationship between language, exploration, and general intelligence.
The Thermodynamic AI Chip · Thomas Ahle
Thomas Ahle explores thermodynamic computing chips that harness physical noise for probabilistic ML, while detailing how AI agents can democratize prohibitively expensive chip design tools—though this raises urgent questions about verification and 'understanding debt' when AI generates complex systems humans no longer fully comprehend.
He won a Nobel here for AlphaFold. Then he left. - John Jumper
Nobel laureate John Jumper explains how AlphaFold solved the 50-year protein structure prediction problem by collapsing years of experimental work into minutes, while emphasizing its narrow scope as a starting point for biological research rather than a universal model of life.
The Ex-Congressman Who Says AI Isn't Unstoppable — Brad Carson
Former Congressman and Pentagon official Brad Carson argues that AI development is not inevitable and can be controlled through strategic regulation, particularly by treating AI as products subject to liability laws rather than granting them human rights, while leveraging chip controls and mandatory testing to shape the technology's future.