Better Data is All You Need — Ari Morcos, Datology

1530 snips

Aug 29, 2025

Ari Morcos, CEO of Datology and former research scientist at Google and Meta, discusses the vital but overlooked area of data curation in AI. He emphasizes that high-quality data is essential for developing efficient models, arguing that the focus should shift from scaling models to improving data quality. Morcos recounts his journey from neuroscience to understanding that as data scales, the importance of model architecture diminishes. He advocates for automating data curation to enable organizations to harness the full potential of AI effectively.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

00:00 / 00:00

Models Are What They Eat

Ari Morcos: "Models are what they eat" captures the core insight that training data defines model quality.
Automating data curation at scale is essential to train faster, better, and smaller models.

00:00 / 00:00

From Neuroscience To Deep Learning Science

Ari Morcos described his path from neuroscience to ML and wanting a science of deep learning.
He ran experiments to understand representations but found it hard to turn insights into consistently causal model improvements.

00:00 / 00:00

Data Overtakes Inductive Bias At Scale

Several papers convinced Ari that at scale inductive biases matter less and data dominates.
As datasets grow, learned posteriors from data outweigh architecture priors.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

Our chat with Ari shows that data curation is the most impactful and underinvested area in AI. He argues that the prevailing focus on model architecture and compute scaling overlooks the "bitter lesson" that "models are what they eat." Effective data curation—a sophisticated process involving filtering, rebalancing, sequencing (curriculum), and synthetic data generation—allows for training models that are simultaneously faster, better, and smaller. Morcos recounts his personal journey from focusing on model-centric inductive biases to realizing that data quality is the primary lever for breaking the diminishing returns of naive scaling laws. Datology's mission is to automate this complex curation process, making state-of-the-art data accessible to any organization and enabling a new paradigm of AI development where data efficiency, not just raw scale, drives progress.

Timestamps

00:00 Introduction

00:46 What is Datology? The mission to train models faster, better, and smaller through data curation.

01:59 Ari's background: From neuroscience to realizing the "Bitter Lesson" of AI.

05:30 Key Insight: Inductive biases from architecture become less important and even harmful as data scale increases.

08:08 Thesis: Data is the most underinvested area of AI research relative to its impact.

10:15 Why data work is culturally undervalued in research and industry.

12:19 How self-supervised learning changed everything, moving from a data-scarce to a data-abundant regime.

17:05 Why automated curation is superior to human-in-the-loop, citing the DCLM study.

19:22 The "Elephants vs. Dogs" analogy for managing data redundancy and complexity.

22:46 A brief history and commentary on key datasets (Common Crawl, GitHub, Books3).

26:24 Breaking naive scaling laws by improving data quality to maintain high marginal information gain.

29:07 Datology's demonstrated impact: Achieving baseline performance 12x faster.

34:19 The business of data: Datology's moat and its relationship with open-source datasets.

39:12 Synthetic Data Explain

ed: The difference between risky "net-new" creation and powerful "rephrasing."

49:02 The Resurgence of Curriculum Learning: Why ordering data matters in the underfitting regime.

52:55 The Future of Training: Optimizing pre-training data to make post-training more effective.

54:49 Who is training their own models and why (Sovereign AI, large enterprises).

57:24 "Train Smaller": Why inference cost makes smaller, specialized models the ultimate goal for enterprises.

01:00:19 The problem with model pruning and why data-side solutions are complementary.

01:03:03 On finding the smallest possible model for a given capability.

01:06:49 Key learnings from the RC foundation model collaboration, proving that data curation "stacks."

01:09:46 Lightning Round: What data everyone wants & who should work at Datology.

01:14:24 Commentary on Meta's superintelligence efforts and Yann LeCun's role.