{"generated_at": "2026-07-22T01:07:09.347803Z", "slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "name": "Latent Space: The AI Engineer Podcast", "episode_count": 18, "avg_signal": 81.1, "median_signal": 82.8, "top_signal": 88.2, "latest_episode_at": "2026-07-16T13:30:00Z", "earliest_episode_at": "2026-04-15T00:31:14Z", "category_mode": "ai", "cover_image_url": "https://substackcdn.com/feed/podcast/1084089/post/207109360/a94f3718358ca28b17f17b7b9cb63144.jpg", "rank_score": 224.134, "episodes": [{"episode_id": "ep_latent_space_the_ai_engineer_podcast_592a8c489a40", "episode_title": "\ud83d\udd2c The Lab of the Future Should Feel Like a Data Center \u2014 Andy Beam & Rafa G\u00f3mez-Bombarelli, Lila Sciences", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-07-16T13:30:00Z", "overall_score": 81.0, "score_breakdown": {"clarity": 59.3, "originality": 55.5, "hype_penalty": 6.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/207109360/a94f3718358ca28b17f17b7b9cb63144.jpg", "source_link": "https://www.latent.space/p/the-lab-of-the-future-should-feel", "audio_url": "https://api.substack.com/feed/podcast/207109360/bfeb234f52e25f505c732000262964af.mp3", "listen_url": "https://www.latent.space/p/the-lab-of-the-future-should-feel", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, know, think, including: We are all in on the bitter lesson and scale.", "summary": "Not just tech bio, what do you do in terms of science?. We think that methods that scale and that are general beat those that are not."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_e137351cd3f7", "episode_title": "Why AI Infrastructure must evolve for Agent Experience \u2014 Akshat Bubna, Modal CTO", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-07-08T22:55:07Z", "overall_score": 83.9, "score_breakdown": {"clarity": 50.0, "originality": 49.7, "hype_penalty": 0.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/205716015/02a47ebb7738eeed6d5cc48ec3f820b1.jpg", "source_link": "https://www.latent.space/p/modal2026", "audio_url": "https://api.substack.com/feed/podcast/205716015/3060ccd83d88111aef06dc1ba0e2b411.mp3", "listen_url": "https://www.latent.space/p/modal2026", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, yeah, it's, including: Your party yesterday was amazing.", "summary": "We're here with Akshat of Modo, CTO of Modo, together with Vibu."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_22854cf2d004", "episode_title": "Why the Frontier Ecosystem must be Open \u2014 Matei Zaharia and Reynold Xin, Databricks", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-06-24T18:53:16Z", "overall_score": 83.3, "score_breakdown": {"clarity": 55.9, "originality": 55.5, "hype_penalty": 3.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/203293676/75f08b4693dd342e455aded8c15273aa.jpg", "source_link": "https://www.latent.space/p/databricks", "audio_url": "https://api.substack.com/feed/podcast/203293676/dd012c693ce8a0df476abdb30707f110.mp3", "listen_url": "https://www.latent.space/p/databricks", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, yeah, think, including: Thanks for having us.", "summary": "Mattia and Reynolds from Databricks, welcome to Lien Space. You have your Databricks Data AI Summit going on."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_f260a1dbec68", "episode_title": "Red-Teaming after Mythos \u2014 Zico Kolter & Matt Fredrikson, Gray Swan", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-06-22T21:06:55Z", "overall_score": 84.5, "score_breakdown": {"clarity": 49.2, "originality": 53.5, "hype_penalty": 0.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/202758604/87156a501f9d66123a10009102e359a9.jpg", "source_link": "https://www.latent.space/p/gray-swan", "audio_url": "https://api.substack.com/feed/podcast/202758604/450849f7afd2f8a05057e4744d0eba68.mp3", "listen_url": "https://www.latent.space/p/gray-swan", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, right, think, including: Thanks for having us.", "summary": "Okay, we're here in the studio with Grace Wan, Matt and Zico."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_f2713f36f435", "episode_title": "The Professor of Outputmaxxing \u2014 Anjney Midha, AMP", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-06-18T17:30:00Z", "overall_score": 83.1, "score_breakdown": {"clarity": 62.5, "originality": 47.8, "hype_penalty": 3.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/202359797/ea2fc7069d01c9b88f01facdf07d26a8.jpg", "source_link": "https://www.latent.space/p/anj", "audio_url": "https://api.substack.com/feed/podcast/202359797/7d6863592b786561d5ce8ed820585ddb.mp3", "listen_url": "https://www.latent.space/p/anj", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, know, it's, including: Thanks for having me.", "summary": "We're in Periodic Labs with Anjimidha, CEO, founder of Amp. So there's two types of utilization usually, right, that you're measuring in these clusters."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_6a21d7fca0c2", "episode_title": "\ud83d\udd2c The Self-Driving Lab \u2014 Joseph Krause, Radical AI", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-06-17T17:58:06Z", "overall_score": 86.1, "score_breakdown": {"clarity": 66.2, "originality": 59.2, "hype_penalty": 3.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/202058620/7148465d0072305423313a9e4cb28933.jpg", "source_link": "https://www.latent.space/p/radical-ai", "audio_url": "https://api.substack.com/feed/podcast/202058620/cbf3e49d692276339a44a698b75ed914.mp3", "listen_url": "https://www.latent.space/p/radical-ai", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, that's, think, including: If you look at bio or maybe small molecules as a more broad category, you look at selfies and smile strings, right?.", "summary": "If you look at bio or maybe small molecules as a more broad category, you look at selfies and smile strings, right?. And that's because you know the elements and then you know the bonds."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_c3d33be3dd2e", "episode_title": "Reality: The Final Eval \u2014 Lukas Petersson and Axel Backlund of Andon Labs", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-06-04T20:39:18Z", "overall_score": 76.5, "score_breakdown": {"clarity": 48.0, "originality": 59.5, "hype_penalty": 9.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/200614482/aa6ef7356b22d140befd10a0e44bb4a5.jpg", "source_link": "https://www.latent.space/p/andon", "audio_url": "https://api.substack.com/feed/podcast/200614482/b8155f27572ea1cda50b33bd50e67f32.mp3", "listen_url": "https://www.latent.space/p/andon", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, yeah, think, including: Thank you for having us.", "summary": "Thank you for having us. Yeah, I'm Lucas and I'm Axel."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_2af898201fc2", "episode_title": "\ud83d\udd2cScaling Past Informal AI - Carina Hong, Axiom Math", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-06-03T19:27:49Z", "overall_score": 81.3, "score_breakdown": {"clarity": 56.9, "originality": 59.4, "hype_penalty": 6.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/199994886/c288c681026bb569cba00020c5003b25.jpg", "source_link": "https://www.latent.space/p/axiom", "audio_url": "https://api.substack.com/feed/podcast/199994886/ba6647aab39199a4ad19cb037e6a8011.mp3", "listen_url": "https://www.latent.space/p/axiom", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, think, know, including: Either it's human-AI collaboration.", "summary": "But it's for the first time now, I think, verified AI is to open up collaboration. Either it's human-AI collaboration."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_cd1d64e9b24d", "episode_title": "\u26a1\ufe0fSatya Nadella: No Priors x Latent Space Crossover Special at Microsoft Build", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-06-03T17:13:57Z", "overall_score": 83.0, "score_breakdown": {"clarity": 66.4, "originality": 43.7, "hype_penalty": 3.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/200432443/1cf5a66a6ee1300ce0cc0121e7c664d8.jpg", "source_link": "https://www.latent.space/p/satya-2026", "audio_url": "https://api.substack.com/feed/podcast/200432443/dff0614ab10e21be66d39c4847157a18.mp3", "listen_url": "https://www.latent.space/p/satya-2026", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of right, think, it's, including: I'm so excited to be here.", "summary": "Music Music Music Music Music Music Music Music Music Music Music Please welcome, SWIX, Sarah Goa, Alad Gill, and Chairman and Chief Executive Officer of Microsoft, Satya Nadella."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_94f5fa20401d", "episode_title": "GitHub's plan for Agents \u2014 Kyle Daigle, GitHub", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-06-02T16:48:21Z", "overall_score": 78.3, "score_breakdown": {"clarity": 50.4, "originality": 50.9, "hype_penalty": 6.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/200249307/e4ca4e6d5f3b40aa381e5f5f554e1a04.jpg", "source_link": "https://www.latent.space/p/github", "audio_url": "https://api.substack.com/feed/podcast/200249307/b006a03717ddcc40eb59c02b49063879.mp3", "listen_url": "https://www.latent.space/p/github", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, know, think, including: Thanks for having me.", "summary": "Okay, we're here with Kyle Bagel, CEO of GitHub. People know you as that."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_22dc831445f9", "episode_title": "The Age of Async Agents \u2014 Cognition's Walden Yan & OpenInspect's Cole Murray", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-05-28T18:41:24Z", "overall_score": 78.8, "score_breakdown": {"clarity": 75.0, "originality": 92.0, "hype_penalty": 3.0, "actionability": 65.0, "technical_depth": 82.0, "information_density": 75.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/199607874/d8e216e78ecdbcd9c655825359d96895.jpg", "source_link": "https://www.latent.space/p/cognition", "audio_url": "https://api.substack.com/feed/podcast/199607874/583bc5ffc268ed91c9847c65cd631492.mp3", "listen_url": "https://www.latent.space/p/cognition", "verdict": "must_listen", "why_listen": "Understand the real architectural trade-offs in building production-grade agent systems from teams shipping them today.", "summary": "Async agents are shifting from experimental tools to core infrastructure, with real-world deployments showing 7x growth in PRs. Cloud-based agents face architectural trade-offs around state, security, and orchestration, while local agents offer tighter control but require more work. The future lies in thoughtful agent design that balances autonomy, context, and integration without overfitting to brittle multi-agent systems."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_ecccf4e281ac", "episode_title": "\ud83d\udd2cESM: The Bitter Lesson is Coming for Proteins - Alex Rives, BioHub", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-05-27T17:46:16Z", "overall_score": 88.2, "score_breakdown": {"clarity": 85.0, "originality": 94.0, "hype_penalty": 2.0, "actionability": 75.0, "technical_depth": 92.0, "information_density": 85.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/199487836/75b5314446676e82174a649769e80f68.jpg", "source_link": "https://www.latent.space/p/esmfold2", "audio_url": "https://api.substack.com/feed/podcast/199487836/d44a44519bdf8c14dc458f47d0972ffc.mp3", "listen_url": "https://www.latent.space/p/esmfold2", "verdict": "must_listen", "why_listen": "Understand how scaling laws and data diversity are enabling a new paradigm in protein modeling that could unlock programmable biology without traditional structural priors.", "summary": "ESM (Evolutionary Scale Modeling) from BioHub demonstrates that scaling protein language models on massive, diverse metagenomic datasets enables high-fidelity representation of protein biology without relying on traditional priors like multiple sequence alignments. The model reveals intrinsic structural and functional patterns through self-supervised learning, achieving representational fidelity that extrapolates via clean scaling laws. Mechanistic interpretability techniques like sparse coding uncover biologically meaningful features, such as catalytic motifs, directly from the model's learned representations."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_55d5f0621636", "episode_title": "Giving Agents Computers \u2014 Ivan Burazin, Daytona", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-05-21T20:37:40Z", "overall_score": 67.0, "score_breakdown": {"clarity": 75.0, "originality": 85.0, "hype_penalty": 5.0, "actionability": 60.0, "technical_depth": 65.0, "information_density": 55.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/198688585/37f23851ec0ce5356367972e01540a0a.jpg", "source_link": "https://www.latent.space/p/daytona", "audio_url": "https://api.substack.com/feed/podcast/198688585/e48f923a6fb4c7f822f1f5013a97c41d.mp3", "listen_url": "https://www.latent.space/p/daytona", "verdict": "worth_your_time", "why_listen": "Understand how AI agents are reshaping infrastructure needs and why ultra-fast, secure sandboxing is becoming a critical layer in the AI stack.", "summary": "The episode explores Daytona's evolution from a developer tool into an AI-native platform that provides instant, high-performance sandboxes for AI agents. Key technical decisions include building on Firecracker VMs, optimizing for 60ms cold starts, and enabling GPU-efficient sandboxing at scale. The discussion highlights how AI agents are driving new infrastructure demands around speed, isolation, and cost."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_5f8eed013b54", "episode_title": "Railway: The Agent-Native Cloud \u2014 Jake Cooper", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-05-20T22:42:06Z", "overall_score": 76.8, "score_breakdown": {"clarity": 75.0, "originality": 94.0, "hype_penalty": 5.0, "actionability": 60.0, "technical_depth": 85.0, "information_density": 75.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/198575235/f919f0d9b5303f9324a789862a0c2623.jpg", "source_link": "https://www.latent.space/p/railway", "audio_url": "https://api.substack.com/feed/podcast/198575235/fbd2be59f0c8aad59a022f3709b9b930.mp3", "listen_url": "https://www.latent.space/p/railway", "verdict": "must_listen", "why_listen": "Understand how agent-native infrastructure will reshape cloud platforms, and why owning the full stack from metal to CLI is critical for the next generation of AI systems.", "summary": "Railway is redefining cloud infrastructure for AI agents by building agent-native infrastructure that eliminates traditional compute constraints, moving from public cloud to bare metal data centers for tighter control and performance. The platform enables frictionless, versioned deployments with safe iteration and blast radius controls, treating long-running agent workflows as first-class citizens. This shift challenges legacy paradigms like static VM allocation and Git-based versioning, anticipating a future where infrastructure is dynamically optimized for autonomous systems."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_959716592984", "episode_title": "The Autonomous Drone Tech Stack & Economics of Drones \u2014 Yaroslav Azhnyuk, The Fourth Law & Guest Host Noah Smith, Noahpinion", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-05-18T13:45:32Z", "overall_score": 83.3, "score_breakdown": {"clarity": 54.2, "originality": 57.5, "hype_penalty": 3.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/198200418/58c14553cecb4dff0fac2a4d478b7bdd.jpg", "source_link": "https://www.latent.space/p/the-fourth-law", "audio_url": "https://api.substack.com/feed/podcast/198200418/d911388480315e62363f89b6bb57a310.mp3", "listen_url": "https://www.latent.space/p/the-fourth-law", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, drones, drone, including: Ukraine is not the most industrious nation in the world.", "summary": "Last year, Ukraine produced 4 million FPV drones. China can produce 4 billion of these FPV drones."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_f729d27f067b", "episode_title": "AI-Native Healthcare: 100M Doctor Visits, 10\u201320 Hours Saved, Prior Auth in Minutes \u2014 Janie Lee & Chai Asawa, Abridge", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-05-14T22:05:31Z", "overall_score": 82.6, "score_breakdown": {"clarity": 50.4, "originality": 57.6, "hype_penalty": 3.0, "actionability": 100.0, "technical_depth": 100.0, "information_density": 100.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/197417280/be8a15c7305ab9f50ed3be9c8f08283c.jpg", "source_link": "https://www.latent.space/p/abridge", "audio_url": "https://api.substack.com/feed/podcast/197417280/2d79546fa3b8f67062f71fa1b134c52c.mp3", "listen_url": "https://www.latent.space/p/abridge", "verdict": "worth_your_time", "why_listen": "It goes beyond the title with direct discussion of like, think, actually, including: Very excited to do this.", "summary": "Okay, this is a special crossover in space on Supervised Learning Pod."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_16b3920a36ea", "episode_title": "\ud83d\udd2c Training Transformers to solve 95% failure rate of Cancer Trials \u2014 Ron Alfa & Daniel Bear, Noetik", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-04-20T16:17:17Z", "overall_score": 84.4, "score_breakdown": {"clarity": 85.0, "originality": 94.0, "hype_penalty": 2.0, "actionability": 75.0, "technical_depth": 82.0, "information_density": 76.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/194810752/3a87f1aa9c0d9002c08d675c5183aae1.jpg", "source_link": "https://www.latent.space/p/noetik", "audio_url": "https://api.substack.com/feed/podcast/194810752/1b92bce4d49858354007a47c48e4e6d4.mp3", "listen_url": "https://www.latent.space/p/noetik", "verdict": "must_listen", "why_listen": "You'll understand how foundation models applied to real patient biology\u2014not synthetic cell lines\u2014could fix the broken paradigm of cancer drug development.", "summary": "Noetik is building foundation models trained on multimodal patient data to solve the 95% failure rate of cancer trials by redefining patient selection, not drug design. The company argues that most oncology drugs fail not due to poor pharmacology but because they're tested on poorly characterized patient populations, often modeled on artificial cell lines that don't reflect human biology. Their approach uses transformer models trained on real tumor samples to identify functional disease subtypes and match drugs to patients based on underlying biology."}, {"episode_id": "ep_latent_space_the_ai_engineer_podcast_d7653a4b08db", "episode_title": "Notion\u2019s Token Town: 5 Rebuilds, 100+ Tools, MCP vs CLIs and the Software Factory Future \u2014 Simon Last & Sarah Sachs of Notion", "podcast_name": "Latent Space: The AI Engineer Podcast", "podcast_slug": "latent_space_the_ai_engineer_podcast", "source_id": "src_latent_space_the_ai_engineer_podcast", "category": "ai", "publish_date": "2026-04-15T00:31:14Z", "overall_score": 76.8, "score_breakdown": {"clarity": 75.0, "originality": 82.0, "hype_penalty": 3.0, "actionability": 65.0, "technical_depth": 82.0, "information_density": 75.0}, "podcast_cover_url": "https://substackcdn.com/feed/podcast/1084089/post/194195821/ee214f89894a45cabf7a528470753d02.jpg", "source_link": "https://www.latent.space/p/notion", "audio_url": "https://api.substack.com/feed/podcast/194195821/10f31b379fbf657fd80757e3b4244e4f.mp3", "listen_url": "https://www.latent.space/p/notion", "verdict": "must_listen", "why_listen": "It reveals how a leading product-led company navigates the gap between early AI promise and shippable reality, with hard-won lessons on when to persist, pivot, or wait for model improvements.", "summary": "Notion rebuilt its agent system five times between 2022 and 2024, iterating through non-existent tooling, poor model reasoning, and short contexts before reliable function calling and improved models made custom agents viable. The team emphasizes a portfolio approach to product development\u2014balancing maintenance, near-term wins, and speculative AGI-pilled projects like software factories where agents collaboratively build, debug, and maintain codebases. They frame Notion not as a wrapper but as a collaboration-layer platform akin to Datadog\u2019s relationship to AWS: leveraging foundational infrastructure while delivering differentiated UX expertise."}], "category_breakdown": [{"category": "ai", "count": 18}]}