--- schema: 1 type: video title: >- The Unreasonable Effectiveness of Separating the Task from the Model — Maxime Rivest & Isaac Miller description: >- By declaring a task's inputs and outputs without initially considering model capability, you create the space needed to determine execution later. DSPy's promise is that AI engineering should happen above a particular prompt template or provider API shape: the Signature. That remains useful in a world of tools, RLMs, and Skills. Define a task strictly through its inputs and outputs, and the underlying implementation becomes flexible: experiment with models, settings, weights, templates, and output formats without touching the workflow. The talk covers DSPy 3.5 and previews DSPy 4.0, where models can write code beneath a signature and programs can learn directly from interactions with users while still respecting the signature's inputs and outputs. Speakers: Maxime Rivest — Core Contributor, DSPy Maxime builds tools and content that make LLMs more accessible and powerful. He is a DSPy core contributor and an open-source Python library author. X/Twitter: https://x.com/MaximeRivest LinkedIn: https://linkedin.com/in/maximerivest Website: https://maximerivest.com Isaac Miller — Lead Maintainer of DSPy; Co-Founder, cmpnd Isaac leads DSPy and co-founded cmpnd, building an open-source framework for self-improving, modular AI systems. X/Twitter: https://x.com/isaacbmiller1 LinkedIn: https://www.linkedin.com/in/miller-isaac/ --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/GgLQ02aO-hs-the-unreasonable-effectiveness-of-separating-the-task-from-the-model Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- canonical: >- https://liquidgold.cc/channel/aidotengineer/explore/video/the-unreasonable-effectiveness-of-separating-the-ta-aa9bb07ae32e markdown: >- https://liquidgold.cc/channel/aidotengineer/video/the-unreasonable-effectiveness-of-separating-the-ta-aa9bb07ae32e.md llms: null parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-24T06:36:16.251Z' language: null children: [] videoId: GgLQ02aO-hs durationSeconds: 1031 publishedAt: '2026-07-23T17:45:06.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 22 endSeconds: 157 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=22' heldReason: rights - ordinal: 1 text: null startSeconds: 22 endSeconds: 157 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=22' heldReason: rights - ordinal: 2 text: null startSeconds: 22 endSeconds: 157 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=22' heldReason: rights - ordinal: 3 text: null startSeconds: 167 endSeconds: 257 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=167' heldReason: rights - ordinal: 4 text: null startSeconds: 167 endSeconds: 257 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=167' heldReason: rights - ordinal: 5 text: null startSeconds: 167 endSeconds: 257 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=167' heldReason: rights - ordinal: 6 text: null startSeconds: 167 endSeconds: 257 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=167' heldReason: rights - ordinal: 7 text: null startSeconds: 268 endSeconds: 354 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=268' heldReason: rights - ordinal: 8 text: null startSeconds: 365 endSeconds: 452 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=365' heldReason: rights - ordinal: 9 text: null startSeconds: 452 endSeconds: 462 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=452' heldReason: rights - ordinal: 10 text: null startSeconds: 462 endSeconds: 569 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=462' heldReason: rights - ordinal: 11 text: null startSeconds: 580 endSeconds: 714 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=580' heldReason: rights - ordinal: 12 text: null startSeconds: 580 endSeconds: 714 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=580' heldReason: rights - ordinal: 13 text: null startSeconds: 725 endSeconds: 813 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=725' heldReason: rights - ordinal: 14 text: null startSeconds: 824 endSeconds: 924 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=824' heldReason: rights - ordinal: 15 text: null startSeconds: 935 endSeconds: 1031 watch: 'https://liquidgold.cc/watch/GgLQ02aO-hs?t=935' heldReason: rights --- # The Unreasonable Effectiveness of Separating the Task from the Model — Maxime Rivest & Isaac Miller By declaring a task's inputs and outputs without initially considering model capability, you create the space needed to determine execution later. DSPy's promise is that AI engineering should happen above a particular prompt template or provider API shape: the Signature. That remains useful in a world of tools, RLMs, and Skills. Define a task strictly through its inputs and outputs, and the underlying implementation becomes flexible: experiment with models, settings, weights, templates, and output formats without touching the workflow. The talk covers DSPy 3.5 and previews DSPy 4.0, where models can write code beneath a signature and programs can learn directly from interactions with users while still respecting the signature's inputs and outputs. Speakers: Maxime Rivest — Core Contributor, DSPy Maxime builds tools and content that make LLMs more accessible and powerful. He is a DSPy core contributor and an open-source Python library author. X/Twitter: https://x.com/MaximeRivest LinkedIn: https://linkedin.com/in/maximerivest Website: https://maximerivest.com Isaac Miller — Lead Maintainer of DSPy; Co-Founder, cmpnd Isaac leads DSPy and co-founded cmpnd, building an open-source framework for self-improving, modular AI systems. X/Twitter: https://x.com/isaacbmiller1 LinkedIn: https://www.linkedin.com/in/miller-isaac/ --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/GgLQ02aO-hs-the-unreasonable-effectiveness-of-separating-the-task-from-the-model Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- Canonical: https://liquidgold.cc/channel/aidotengineer/explore/video/the-unreasonable-effectiveness-of-separating-the-ta-aa9bb07ae32e · Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks > Missing: no stored document chunks are available.