--- schema: 1 type: video title: 'The AI Bill Is Too Damn High: Tokenomics 101 — Eugene Cheah, Featherless AI' description: "If your token spend keeps going up and productivity isn't keeping pace, start here. Eugene Cheah lays out four steps for cutting AI inference costs at scale: measure, right-size, buy capacity well and check your processes. Along the way he covers splitting API keys per team, agentic vs. process workloads, open models that match frontier models on specific tasks, classifiers, fine-tuning, the speed-vs-cost tradeoff and \"slop grenades.\" Recorded live at AI Engineer NYC 2026. \U0001F39F️ Join us at the next one: AIE CODE Summit → https://aie.codes Speaker Eugene Cheah — X: https://x.com/picocreator Featherless AI — https://featherless.ai Timestamps 0:00 Intro: Eugene Cheah, Featherless AI 1:48 What \"scale\" looks like: 2T tokens/day 2:13 Disclaimers: who this is for 3:13 The 4 moves 3:23 Move 1: Measure (stop sharing one API key) 5:12 Agentic vs. process workloads 9:27 Move 2: Right-size your models 12:42 Classifiers: a new class of cheap models 14:37 Plan with the smart model, build with the cheap one 15:47 Nobody switches models to save money 19:27 Give teams model labels, not model names 25:27 Bootstrapping your own fine-tuned model 27:42 Shadow traffic to keep improving 28:57 Move 3: Buying GPU capacity 30:07 2x the speed = 3x the cost 32:22 Bursty traffic & dedicated capacity 34:32 Move 4: Check your processes 37:07 Tokens are not work done 42:07 Don't throw slop grenades 44:27 Q&A: Fine-tuning & data sets 48:02 Q&A: Golden traces 50:04 Q&A: Escaping harness lock-in 52:37 Q&A: Caching & auto-routing" canonical: >- https://liquidgold.cc/channel/aidotengineer/explore/video/the-ai-bill-is-too-damn-high-tokenomics-101-eugene-363f8998522c markdown: >- https://liquidgold.cc/channel/aidotengineer/video/the-ai-bill-is-too-damn-high-tokenomics-101-eugene-363f8998522c.md llms: null parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-23T22:23:37.313Z' language: null children: [] videoId: Ti-BmGU1_mQ durationSeconds: 3325 publishedAt: '2026-10-17T09:00:27.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 12 endSeconds: 23 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=12' heldReason: rights - ordinal: 1 text: null startSeconds: 23 endSeconds: 112 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=23' heldReason: rights - ordinal: 2 text: null startSeconds: 112 endSeconds: 124 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=112' heldReason: rights - ordinal: 3 text: null startSeconds: 124 endSeconds: 136 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=124' heldReason: rights - ordinal: 4 text: null startSeconds: 136 endSeconds: 146 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=136' heldReason: rights - ordinal: 5 text: null startSeconds: 146 endSeconds: 200 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=146' heldReason: rights - ordinal: 6 text: null startSeconds: 200 endSeconds: 211 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=200' heldReason: rights - ordinal: 7 text: null startSeconds: 211 endSeconds: 222 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=211' heldReason: rights - ordinal: 8 text: null startSeconds: 222 endSeconds: 313 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=222' heldReason: rights - ordinal: 9 text: null startSeconds: 313 endSeconds: 324 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=313' heldReason: rights - ordinal: 10 text: null startSeconds: 324 endSeconds: 569 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=324' heldReason: rights - ordinal: 11 text: null startSeconds: 569 endSeconds: 581 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=569' heldReason: rights - ordinal: 12 text: null startSeconds: 581 endSeconds: 769 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=581' heldReason: rights - ordinal: 13 text: null startSeconds: 769 endSeconds: 781 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=769' heldReason: rights - ordinal: 14 text: null startSeconds: 781 endSeconds: 880 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=781' heldReason: rights - ordinal: 15 text: null startSeconds: 880 endSeconds: 891 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=880' heldReason: rights - ordinal: 16 text: null startSeconds: 891 endSeconds: 957 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=891' heldReason: rights - ordinal: 17 text: null startSeconds: 957 endSeconds: 968 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=957' heldReason: rights - ordinal: 18 text: null startSeconds: 968 endSeconds: 1167 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=968' heldReason: rights - ordinal: 19 text: null startSeconds: 1167 endSeconds: 1178 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=1167' heldReason: rights - ordinal: 20 text: null startSeconds: 1178 endSeconds: 1281 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=1178' heldReason: rights - ordinal: 21 text: null startSeconds: 1281 endSeconds: 1291 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=1281' heldReason: rights - ordinal: 22 text: null startSeconds: 1291 endSeconds: 1401 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=1291' heldReason: rights - ordinal: 23 text: null startSeconds: 1401 endSeconds: 1411 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=1401' heldReason: rights - ordinal: 24 text: null startSeconds: 1411 endSeconds: 1528 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=1411' heldReason: rights - ordinal: 25 text: null startSeconds: 1528 endSeconds: 1540 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=1528' heldReason: rights - ordinal: 26 text: null startSeconds: 1540 endSeconds: 1663 watch: 'https://liquidgold.cc/watch/Ti-BmGU1_mQ?t=1540' heldReason: rights --- # The AI Bill Is Too Damn High: Tokenomics 101 — Eugene Cheah, Featherless AI If your token spend keeps going up and productivity isn't keeping pace, start here. Eugene Cheah lays out four steps for cutting AI inference costs at scale: measure, right-size, buy capacity well and check your processes. Along the way he covers splitting API keys per team, agentic vs. process workloads, open models that match frontier models on specific tasks, classifiers, fine-tuning, the speed-vs-cost tradeoff and "slop grenades." Recorded live at AI Engineer NYC 2026. 🎟️ Join us at the next one: AIE CODE Summit → https://aie.codes Speaker Eugene Cheah — X: https://x.com/picocreator Featherless AI — https://featherless.ai Timestamps 0:00 Intro: Eugene Cheah, Featherless AI 1:48 What "scale" looks like: 2T tokens/day 2:13 Disclaimers: who this is for 3:13 The 4 moves 3:23 Move 1: Measure (stop sharing one API key) 5:12 Agentic vs. process workloads 9:27 Move 2: Right-size your models 12:42 Classifiers: a new class of cheap models 14:37 Plan with the smart model, build with the cheap one 15:47 Nobody switches models to save money 19:27 Give teams model labels, not model names 25:27 Bootstrapping your own fine-tuned model 27:42 Shadow traffic to keep improving 28:57 Move 3: Buying GPU capacity 30:07 2x the speed = 3x the cost 32:22 Bursty traffic & dedicated capacity 34:32 Move 4: Check your processes 37:07 Tokens are not work done 42:07 Don't throw slop grenades 44:27 Q\&A: Fine-tuning & data sets 48:02 Q\&A: Golden traces 50:04 Q\&A: Escaping harness lock-in 52:37 Q\&A: Caching & auto-routing Canonical: https://liquidgold.cc/channel/aidotengineer/explore/video/the-ai-bill-is-too-damn-high-tokenomics-101-eugene-363f8998522c · Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks --- #### Chunk 0: Document excerpt (0:12 to 2:04) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 1: Document excerpt (2:04 to 3:31) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 2: Document excerpt (3:31 to 5:13) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 3: Document excerpt (5:13 to 9:29) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 4: Document excerpt (9:29 to 12:49) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 5: Document excerpt (12:49 to 14:40) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 6: Document excerpt (14:40 to 19:27) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 7: Document excerpt (19:27 to 21:21) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 8: Document excerpt (21:21 to 23:31) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 9: Document excerpt (23:31 to 25:40) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 10: Document excerpt (25:40 to 27:54) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 11: Document excerpt (27:54 to 32:24) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 12: Document excerpt (32:24 to 34:36) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 13: Document excerpt (34:36 to 37:10) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 14: Document excerpt (37:10 to 38:58) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 15: Document excerpt (38:58 to 40:58) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 16: Document excerpt (40:58 to 42:16) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 17: Document excerpt (42:16 to 44:30) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 18: Document excerpt (44:30 to 48:10) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 19: Document excerpt (48:10 to 50:08) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 20: Document excerpt (50:08 to 52:48) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 21: Document excerpt (52:48 to 55:25) > Held: transcript rights do not permit this chunk's text.