--- schema: 1 type: video title: >- Self Driving Products: Product Signals to Pull Requests — Joshua Snyder, PostHog description: >- A rage click, a 2am error spike, a customer Slack message — today each sits until a developer notices, triages, tickets, and writes a fix. PostHog is building a pipeline that collapses that chain: signal arrives, a background agent groups it with related errors and session replays, researches the codebase, and opens a PR. You wake up to green PRs instead of dashboards. Three lessons from building it: off the shelf embedding models cluster signals by structural similarity rather than meaning, so errors land next to errors and Slack messages land next to Slack messages — the fix is to embed LLM generated queries rather than the signals themselves. Specificity determines whether the agent produces a useful PR or just fixes something at random; error tracking is immediately actionable, Slack and session replay usually are not. And start with agents even when it looks expensive — run the same problem through an agent 100 times, find the patterns, then collapse the expensive step into a one shot call. Speaker info: - https://x.com/joshsny -https://www.linkedin.com/in/joshsny/ --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/zMiSRliEzv4-self-driving-products-product-signals-to-pull-requests 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/self-driving-products-product-signals-to-pull-reque-b849fe61ab09 markdown: >- https://liquidgold.cc/channel/aidotengineer/video/self-driving-products-product-signals-to-pull-reque-b849fe61ab09.md llms: null parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-24T09:33:32.670Z' language: null children: [] videoId: zMiSRliEzv4 durationSeconds: 939 publishedAt: '2026-06-10T13:00:17.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 25 endSeconds: 129 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=25' heldReason: rights - ordinal: 1 text: null startSeconds: 25 endSeconds: 129 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=25' heldReason: rights - ordinal: 2 text: null startSeconds: 129 endSeconds: 140 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=129' heldReason: rights - ordinal: 3 text: null startSeconds: 140 endSeconds: 269 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=140' heldReason: rights - ordinal: 4 text: null startSeconds: 140 endSeconds: 269 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=140' heldReason: rights - ordinal: 5 text: null startSeconds: 280 endSeconds: 368 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=280' heldReason: rights - ordinal: 6 text: null startSeconds: 280 endSeconds: 368 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=280' heldReason: rights - ordinal: 7 text: null startSeconds: 368 endSeconds: 378 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=368' heldReason: rights - ordinal: 8 text: null startSeconds: 378 endSeconds: 513 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=378' heldReason: rights - ordinal: 9 text: null startSeconds: 378 endSeconds: 513 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=378' heldReason: rights - ordinal: 10 text: null startSeconds: 378 endSeconds: 513 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=378' heldReason: rights - ordinal: 11 text: null startSeconds: 513 endSeconds: 525 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=513' heldReason: rights - ordinal: 12 text: null startSeconds: 525 endSeconds: 620 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=525' heldReason: rights - ordinal: 13 text: null startSeconds: 525 endSeconds: 620 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=525' heldReason: rights - ordinal: 14 text: null startSeconds: 620 endSeconds: 631 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=620' heldReason: rights - ordinal: 15 text: null startSeconds: 631 endSeconds: 719 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=631' heldReason: rights - ordinal: 16 text: null startSeconds: 631 endSeconds: 719 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=631' heldReason: rights - ordinal: 17 text: null startSeconds: 729 endSeconds: 852 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=729' heldReason: rights - ordinal: 18 text: null startSeconds: 729 endSeconds: 852 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=729' heldReason: rights - ordinal: 19 text: null startSeconds: 864 endSeconds: 939 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=864' heldReason: rights - ordinal: 20 text: null startSeconds: 864 endSeconds: 939 watch: 'https://liquidgold.cc/watch/zMiSRliEzv4?t=864' heldReason: rights --- # Self Driving Products: Product Signals to Pull Requests — Joshua Snyder, PostHog A rage click, a 2am error spike, a customer Slack message — today each sits until a developer notices, triages, tickets, and writes a fix. PostHog is building a pipeline that collapses that chain: signal arrives, a background agent groups it with related errors and session replays, researches the codebase, and opens a PR. You wake up to green PRs instead of dashboards. Three lessons from building it: off the shelf embedding models cluster signals by structural similarity rather than meaning, so errors land next to errors and Slack messages land next to Slack messages — the fix is to embed LLM generated queries rather than the signals themselves. Specificity determines whether the agent produces a useful PR or just fixes something at random; error tracking is immediately actionable, Slack and session replay usually are not. And start with agents even when it looks expensive — run the same problem through an agent 100 times, find the patterns, then collapse the expensive step into a one shot call. Speaker info: - https://x.com/joshsny -https://www.linkedin.com/in/joshsny/ --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/zMiSRliEzv4-self-driving-products-product-signals-to-pull-requests 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/self-driving-products-product-signals-to-pull-reque-b849fe61ab09 · 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.