From generative AI to the brain: five takeaways

Kari Jaaskelainen

21 Nov 2025 — 1 min read

What if the brain builds and tests ideas the way modern generative AI produces images and text? A new paper by Claudius Gros argues that clear, testable generative principles—not obscure tricks—drove AI's leap, and neuroscience can probe whether similar rules guide the brain.

Five takeaways

World models: Brains may not need perfect world models; like AI, they can get far with good-enough predictions and priors.
Generation of thought: Cognition may work by sampling trains of thought, not by finding a single optimum—generate, then select.
Attention: Attention routes and compresses information, spotlighting what matters while saving compute.
Neural scaling laws: Bigger models improve predictably—until data or compute bottlenecks—hinting at similar brain-resource trade-offs.
Quantization: Low-precision signals can work remarkably well, suggesting the brain may rely on coarse codes to stay efficient.

Bottom line: Machine learning now offers concrete, testable hypotheses for neuroscience. Bridging the two could reveal how minds generate, filter, and scale thought. Read: https://arxiv.org/abs/2511.16432v1

Paper: https://arxiv.org/abs/2511.16432v1

Register: https://www.AiFeta.com

ai neuroscience generativeai brain machinelearning attention scalinglaws quantization worldmodels cognition

Meet GRACE: a moral governor for safer, more transparent AI

AI agents are getting powerful—so how do we make sure they do the right thing, not just the effective thing? Meet GRACE, a reason-based moral governor that keeps AI behavior aligned with human norms by separating moral reasoning from goal-driven decision-making. * Moral Module: uses deontic logic and explicit reasons

Why some fine-tuned LLMs miss phishing—and how to fix it

Not all fine-tuned LLMs spot phishing equally. A new study tests Llama 3.1 8B, Gemma 2 9B, and Mistral on high-stakes phishing detection—and uses SHAP and mechanistic interpretability to reveal why models do (or don’t) generalize. * Architecture × data diversity matters: Gemma 2 9B hits state-of-the-art performance (F1

AI that explains itself—by following the science

AI that explains itself—by following the science Black-box AI can be powerful, but it often can’t tell us why it made a decision. Concept Bottleneck Models (CBMs) try to fix that by predicting human-understandable concepts first, then the final answer. The catch: standard CBMs ignore domain-specific cause-and-effect and

Meet GenomAgent: A Team of AI Specialists for Smarter Genomics Q&A

TL;DR: Finding reliable answers in genomics is hard. GenomAgent turns one big AI into a coordinating team of specialists—and beats the current leader by 12% on a key benchmark. Genomic facts live across many databases. Standard chatbots struggle because they can’t flexibly query those sources. GeneGPT added

Five takeaways

Read more

Meet GRACE: a moral governor for safer, more transparent AI

Why some fine-tuned LLMs miss phishing—and how to fix it

AI that explains itself—by following the science

Meet GenomAgent: A Team of AI Specialists for Smarter Genomics Q&A