Recurrence and Attention for Long-Context Transformers with Jacob Buckman - #750

2025-10-07 · 57 min · episode 750 · 13 entities

episode page audio

Asserted relationships

→ appeared on The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence podcast

0.68

evidence rules-v4
Recurrence and Attention for Long-Context Transformers with Jacob Buckman - #750
→ hosted by Sam Charrington person

0.55

evidence rules-v4
Feed author/publisher: Sam Charrington
→ discusses Science concept

0.40

evidence rules-v4
Feed category: Science
→ discusses Technology company

0.40

evidence rules-v4
Feed category: Technology
→ discusses News concept

0.40

evidence rules-v4
Feed category: News
→ discusses Tech News concept

0.40

evidence rules-v4
Feed category: Tech News
→ hosted by TWIML company

0.40

evidence rules-v4
Feed author/publisher: TWIML

Entities found in this episode

concepts 5

mentioned Science concept

0.50

evidence rules-v4
Feed category: Science
mentioned Tech News concept

0.50

evidence rules-v4
Feed category: Tech News
discusses Science concept

0.40

evidence rules-v4
Feed category: Science
discusses News concept

0.40

evidence rules-v4
Feed category: News
discusses Tech News concept

0.40

evidence rules-v4
Feed category: Tech News

companys 4

mentioned TWIML company

0.70

evidence rules-v4
Feed author/publisher: TWIML
mentioned Technology company

0.50

evidence rules-v4
Feed category: Technology
discusses Technology company

0.40

evidence rules-v4
Feed category: Technology
hosted by TWIML company

0.40

evidence rules-v4
Feed author/publisher: TWIML

persons 3

mentioned Jacob Buckman person

0.72

evidence rules-v4
Recurrence and Attention for Long-Context Transformers with Jacob Buckman - #750
mentioned Sam Charrington person

0.70

evidence rules-v4
Feed author/publisher: Sam Charrington
hosted by Sam Charrington person

0.55

evidence rules-v4
Feed author/publisher: Sam Charrington

podcasts 1

appeared on The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence podcast

0.68

evidence rules-v4
Recurrence and Attention for Long-Context Transformers with Jacob Buckman - #750

Episode description as stored

Today, we're joined by Jacob Buckman, co-founder and CEO of Manifest AI to discuss achieving long context in transformers. We discuss the bottlenecks of scaling context length and recent techniques to overcome them, including windowed attention, grouped query attention, and latent space attention. We explore the idea of weight-state balance and the weight-state FLOP ratio as a way of reasoning about the optimality of compute architectures, and we dig into the Power Retention architecture, which blends the parallelization of attention with the linear scaling of recurrence and promises speedups of >10x during training and >100x during inference. We review Manifest AI’s recent open source projects as well: Vidrial—a custom CUDA framework for building highly optimized GPU kernels in Python, and PowerCoder—a 3B-parameter coding model fine-tuned from StarCoder to use power retention. Our chat also covers the use of metrics like in-context learning curves and negative log likelihood to measure context utility, the implications of scaling laws, and the future of long context lengths in AI applications. The complete show notes for this episode can be found at https://twimlai.com/go/750.