Gemma
includes Gemma 4, Gemma 2, Gemma 2B, Gemma 27B, Gemma 3, Gemma 3n, Gemma 9B, Gemma 1, Gemma 2 2B, Gemma Scope, Gemma 3.1, Gemma 7B and 3 more
8 statements across 5 episodes · 4 bullish · 1 bearish · 6 people on the record · first statement Mar 23, 2025 by Rishabh Agarwal · said 109 times in 20 episodes since 2024 · across every show →
Mentions by year, the whole family
brought up most by Omar Sanseviero (32), Alessio Fanelli (17), Shawn Wang (12), Rishabh Agarwal (11), Vibhu (Viboo) (8), Emmanuel Ameisen (4), Peter Robicheaux (2), Jack Morris (2)
tap a year for its mentions
2026 58 mentions in 8 episodes 7 per episode
- ⚡️ Google's Open AI Strategy — Omar Sanseviero, Google DeepMind
- Next 100x in AI: Inference, Networking, & Self-Optimizing Models — Philip Kiely & Ali Taha, Baseten
- The $15B Physical AI Company: Simulation, Autonomy OS, Neural Sim, & 1K Engineers—Applied Intuition
- The Inference Frontier: from 100 to 10,000 tokens per second — Sean Lie, Cerebras CTO
- The AI Frontier: from Gemini 3 Deep Think distilling to Flash — Jeff Dean
- Goodfire AI’s Bet: Interpretability as the Next Frontier of Model Design — Myra Deng & Mark Bissell
- The AI Frontier: from open weights to open research — Eiso Kant, Poolside AI
- Inside xAI: Building Grok Imagine in 3 Months, Videogen vs World Models, and Video Agents— Ethan He
- every mention in 2026, scene by scene →
2025 37 mentions in 9 episodes 4 per episode
- The Utility of Interpretability — Emmanuel Amiesen
- The Magic of LLM Distillation — Rishabh Agarwal, Google DeepMind
- Information Theory for Language Models: Jack Morris
- Better Data is All You Need — Ari Morcos, Datology
- ⚡ Open Model Pretraining Masterclass — Elie Bakouch, HuggingFace SmolLM 3, FineWeb, FinePDF
- A Technical History of Generative Media
- Gemini 2.0 Flash and Flash Thinking: the new SOTA models for the agentic era
- Why is everyone cloning Deep Research?
- 1 more episode that year, every mention in 2025 →
2024 14 mentions in 3 episodes 5 per episode
Everything said about Gemma, oldest first
Mar 23, 2025
Agarwal: Filtered 9B Synthetic Data Outperforms 27B Self-Generated Data
“One thing we found consistently, so here what we had two models, nine Gemma, nine B and Gemma, 27 B, and we found consistently that actually generating data from nine B in a compute match setting is always better, even better for distilling or actually improvi…”
Jun 6, 2025 positive
Ameisen: Multi-Hop Reasoning Circuits Are Extremely Similar Across Small and Large Models
“The way the circuit looks in Gemma, like a really small model is extremely similar to the way that it looks like a huge model, which that in itself is, I think like a pretty novel discovery. It's like, oh, you have these models that are like super different. Y…”
Jun 6, 2025 positive
Vibhu: Gemma Activates Abstract Behavioral Traits Over Simple Token Completion
“It also shows internally that there's more than just token completion of, you know, this plus this equals this. No, it has some under understanding of characteristics, right? Like this is a pretty stubborn dog. It has a stubborn feature. Pretty high up that ac…”
Jul 2, 2025 negative
Morris: Open model creators do not use differential privacy or anonymization
“I would be extremely surprised if they do any type of like private training. Like there are these mechanisms for doing like differentially private language model training, or even just anonymization in the pre-training pipeline. I bet they don't do any of that…”
Aug 29, 2025 positive
May 24, 2026 positive
May 24, 2026 neutral