Landmark research across intelligence science

Vaswani et al.

Introduced the Transformer, replacing recurrence with self-attention and reshaping modern AI.

He, Zhang, Ren, Sun

Residual connections enabled training of very deep networks, winning ImageNet 2015.

Silver et al.

AlphaGo combined deep networks and Monte Carlo tree search to defeat world-class Go players.

Kaplan et al.

Showed that LLM performance follows smooth, predictable power-law relationships with compute, data, and parameters.

Hoffmann et al.

Demonstrated that for a fixed compute budget, model size and training tokens should scale roughly equally.

Wei et al.

Argued that certain capabilities appear abruptly above a scale threshold rather than improving smoothly.

Amodei, Olah, Steinhardt, Christiano, Schulman, Mané

Foundational taxonomy of practical safety problems in modern ML systems.

Ouyang et al.

Introduced InstructGPT and the now-standard RLHF pipeline for aligning LLMs with human intent.

Bai et al.

Trained a helpful, harmless assistant using AI-generated critiques guided by a written constitution.

Karl Friston

Proposed that the brain minimizes a single quantity - variational free energy - across perception, action, and learning.

John O'Keefe, Lynn Nadel

Established the hippocampus as a spatial-mapping system, foundational to memory research.

Raichle et al.

Identified the default mode network - brain regions most active during rest and internal thought.

Willett et al.

An intracortical BCI decoded imagined handwriting into text at 90 characters per minute.

Anumanchipalli, Chartier, Chang

Synthesized intelligible speech directly from cortical activity in the vocal motor area.

Boyden, Zhang, Bamberg, Nagel, Deisseroth

Founding paper of optogenetics - using light to control genetically targeted neurons with millisecond precision.

Uta Frith, Francesca Happé

Proposed that autistic cognition favors local detail processing over global integration.

Sonuga-Barke, Castellanos

Proposed that ADHD attention lapses arise from failure to suppress default-mode network activity.

Snowling, Stanovich

Established phonological processing impairments as the core cognitive feature of developmental dyslexia.

Charles Spearman

Proposed the existence of a general intelligence factor (g) underlying performance across cognitive tasks.

Daniel Kahneman

Synthesized decades of research into a dual-process model: fast intuitive System 1 and slow deliberative System 2.

Raymond Cattell

Distinguished fluid intelligence (novel reasoning) from crystallized intelligence (accumulated knowledge).

Carhart-Harris et al.

Head-to-head Phase 2 trial comparing psilocybin-assisted therapy with a standard SSRI in moderate-to-severe depression.

Goodwin et al.

Multi-site dose-finding Phase 2b trial of a single psilocybin dose in treatment-resistant depression.

Mitchell et al.

First Phase 3 RCT of MDMA-assisted therapy for post-traumatic stress disorder.

Carhart-Harris, Friston

Theoretical synthesis proposing that psychedelics work by relaxing high-level priors in a hierarchical predictive brain.

Moda-Sava et al.

Two-photon imaging in mice showing rapid dendritic spine formation after ketamine and its causal role in sustained behavioral rescue.

Kaelen et al.

Demonstrated that the strength of patients' emotional response to music during psilocybin sessions predicts clinical outcome at follow-up.