[{"@context":"https:\/\/schema.org\/","@type":"BlogPosting","@id":"https:\/\/blog.terabox.com\/insights\/shinka-evolve-ia-evolutiva-descubrimiento-cientifico#BlogPosting","mainEntityOfPage":"https:\/\/blog.terabox.com\/insights\/shinka-evolve-ia-evolutiva-descubrimiento-cientifico","headline":"Shinka Evolve: C\u00f3mo la IA evolutiva transforma la ciencia","name":"Shinka Evolve: C\u00f3mo la IA evolutiva transforma la ciencia","description":"\ud83d\udcfa V\u00eddeo de estudio recomendado hoy: https:\/\/www.youtube.com\/watch?v=EInEmGaMRLc La Evoluci\u00f3n de la Inteligencia: C\u00f3mo los LLMs est\u00e1n Re-inventando el M\u00e9todo Cient\u00edficoShinka Evolve y la Computaci\u00f3n EvolutivaEl “Problema del Problema” y la AperturaEl Cient\u00edfico de IA V2: Autonom\u00eda TotalConclusiones clavePreguntas y Respuestas... ","datePublished":"2026-06-23","dateModified":"2026-06-23","author":{"@type":"Person","@id":"https:\/\/blog.terabox.com\/author\/flextech-admin\/#Person","name":"flextech-admin","url":"https:\/\/blog.terabox.com\/author\/flextech-admin\/","image":{"@type":"ImageObject","@id":"https:\/\/secure.gravatar.com\/avatar\/ad516503a11cd5ca435acc9bb6523536?s=150&d=mm&r=gforcedefault=1","url":"https:\/\/secure.gravatar.com\/avatar\/ad516503a11cd5ca435acc9bb6523536?s=150&d=mm&r=gforcedefault=1","height":96,"width":96}},"publisher":{"@type":"Organization","name":"terabox","logo":{"@type":"ImageObject","@id":"http:\/\/blog.terabox.com\/wp-content\/uploads\/2021\/11\/logo\u4ea7\u54c1\u540d-\u7ad6\u7248.png","url":"http:\/\/blog.terabox.com\/wp-content\/uploads\/2021\/11\/logo\u4ea7\u54c1\u540d-\u7ad6\u7248.png","width":900,"height":900}},"image":{"@type":"ImageObject","@id":"https:\/\/img.youtube.com\/vi\/EInEmGaMRLc\/maxresdefault.jpg","url":"https:\/\/img.youtube.com\/vi\/EInEmGaMRLc\/maxresdefault.jpg","height":"","width":""},"url":"https:\/\/blog.terabox.com\/insights\/shinka-evolve-ia-evolutiva-descubrimiento-cientifico","video":{"@context":"http:\/\/schema.org\/","@type":"VideoObject","@id":"https:\/\/www.youtube.com\/watch?v=EInEmGaMRLc#VideoObject","contentUrl":"https:\/\/www.youtube.com\/watch?v=EInEmGaMRLc","name":"When AI Discovers the Next Transformer \u2014 Robert Lange","description":"Robert Lange, founding researcher at Sakana AI, joins Tim to discuss *Shinka Evolve* \u2014 a framework that combines LLMs with evolutionary algorithms to do open-ended program search. The core claim: systems like AlphaEvolve can optimize solutions to fixed problems, but real scientific progress requires co-evolving the problems themselves.\n\nGTC is coming, the premier AI conference, great opportunity to learn about AI. NVIDIA and partners will showcase breakthroughs in physical AI, AI factories, agentic AI, and inference, exploring the next wave of AI innovation for developers and researchers. Register for virtual GTC for free, using my link and win NVIDIA DGX Spark (https:\/\/nvda.ws\/4qQ0LMg)\n\nIn this episode:\n\n\u2022 Why AlphaEvolve gets stuck \u2014 it needs a human to hand it the right problem. Shinka tries to invent new problems automatically, drawing on ideas from POET, PowerPlay, and MAP-Elites quality-diversity search.\n\n\u2022 The *architecture* of Shinka: an archive of programs organized as islands, LLMs used as mutation operators, and a UCB bandit that adaptively selects between frontier models (GPT-5, Sonnet 4.5, Gemini) mid-run. The credit-assignment problem across models turns out to be genuinely hard.\n\n\u2022 Concrete results \u2014 state-of-the-art circle packing with dramatically fewer evaluations, second place in an AtCoder competitive programming challenge, evolved load-balancing loss functions for mixture-of-experts models, and agent scaffolds for AIME math benchmarks.\n\n\u2022 Are these systems actually thinking outside the box, or are they parasitic on their starting conditions? When LLMs run autonomously, \"nothing interesting happens.\" Robert pushes back with the stepping-stone argument \u2014 evolution doesn't need to extrapolate, just recombine usefully.\n\n\u2022 The AI Scientist question: can automated research pipelines produce real science, or just workshop-level slop that passes surface-level review? Robert is honest that the current version is more co-pilot than autonomous researcher.\n\n\u2022 Where this lands in 5-20 years \u2014 Robert's prediction that scientific research will be fundamentally transformed, and Tim's thought experiment about alien mathematical artifacts that no human could have conceived.\n\nRobert Lange: https:\/\/roberttlange.com\/\n\n---\nTIMESTAMPS:\n00:00:00 Introduction: Robert Lange, Sakana AI and Shinka Evolve\n00:04:15 AlphaEvolve's Blind Spot: Co-Evolving Problems with Solutions\n00:09:05 Unknown Unknowns, POET, and Auto-Curricula for AI Science\n00:14:20 MAP-Elites and Quality-Diversity: Shinka's Evolutionary Architecture\n00:28:00 UCB Bandits, Mutations and the Vibe Research Vision\n00:40:00 Scaling Shinka: Meta-Evolution, Democratisation and the Three-Axis Model\n00:47:10 Applications, ARC-AGI and the Future of Work\n00:57:00 The AI Scientist and the Human Co-Pilot: Who Steers the Search?\n01:06:00 AI Scientist v2, Slop Critique and the Future of Scientific Publishing\n\n---\nREFERENCES:\npaper:\n[00:03:30] ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolution\nhttps:\/\/arxiv.org\/abs\/2509.19349\n[00:04:15] AlphaEvolve: A Coding Agent for Scientific and Algorithmic Discovery\nhttps:\/\/arxiv.org\/abs\/2506.13131\n[00:06:30] Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents\nhttps:\/\/arxiv.org\/abs\/2505.22954\n[00:09:05] Paired Open-Ended Trailblazer (POET)\nhttps:\/\/arxiv.org\/abs\/1901.01753\n[00:10:00] PowerPlay: Training an Increasingly General Problem Solver by Continually Searching for the Simplest Still Unsolvable Problem\nhttps:\/\/arxiv.org\/abs\/1112.5309\n[00:10:40] Automated Capability Discovery via Foundation Model Self-Exploration\nhttps:\/\/arxiv.org\/abs\/2502.07577\n[00:15:30] Illuminating Search Spaces by Mapping Elites (MAP-Elites)\nhttps:\/\/arxiv.org\/abs\/1504.04909\n[00:47:10] Automated Design of Agentic Systems (ADAS)\nhttps:\/\/arxiv.org\/abs\/2408.08435\n[00:49:50] Discovering Preference Optimization Algorithms with and for Large Language Models (DiscoPOP)\nhttps:\/\/arxiv.org\/abs\/2406.08414\n[00:57:00] The AI Scientist v2: Automating the Full Research Pipeline\nhttps:\/\/arxiv.org\/abs\/2504.08066\nbook:\n[00:06:48] Why Greatness Cannot Be Planned\nhttps:\/\/link.springer.com\/book\/10.1007\/978-3-319-15524-1\nbenchmark:\n[00:47:10] ALE-Bench: A Benchmark for Long-Horizon Objective-Driven Algorithm Engineering\nhttps:\/\/arxiv.org\/abs\/2506.09050\n[00:50:50] On the Measure of Intelligence (ARC-AGI)\nhttps:\/\/arxiv.org\/abs\/1911.01547\n\n---\nLINKS:\nDownload PDF transcript: https:\/\/app.rescript.info\/api\/sessions\/b8a9dcf60623657c\/pdf\/download\nFull Transcript: https:\/\/app.rescript.info\/public\/share\/SDOD_3oXOcli3zTqcAtR8eibT5U3gam84oo4KRtI-Vk","thumbnailUrl":["https:\/\/i.ytimg.com\/vi\/EInEmGaMRLc\/default.jpg","https:\/\/i.ytimg.com\/vi\/EInEmGaMRLc\/mqdefault.jpg","https:\/\/i.ytimg.com\/vi\/EInEmGaMRLc\/hqdefault.jpg","https:\/\/i.ytimg.com\/vi\/EInEmGaMRLc\/sddefault.jpg","https:\/\/i.ytimg.com\/vi\/EInEmGaMRLc\/maxresdefault.jpg"],"uploadDate":"2026-03-13T21:00:10+00:00","duration":"PT1H18M7S","embedUrl":"https:\/\/www.youtube.com\/embed\/EInEmGaMRLc","publisher":{"@type":"Organization","@id":"https:\/\/www.youtube.com\/channel\/UCMLtBahI5DMrt0NPvDSoIRQ#Organization","url":"https:\/\/www.youtube.com\/channel\/UCMLtBahI5DMrt0NPvDSoIRQ","name":"Machine Learning Street Talk","description":"MLST is the leading highly technical AI podcast. Subscribe now! Welcome! We bring you the latest in advanced AI research, from the best AI experts in the world. Our approach is unrivalled in terms of scope and rigour \u2013 we believe in diversity of ideas (which is to say, not just LLMs!) and we also cover other promising alternative paths to AGI, as well as CogSci, CompSci, Neuro, Mathematics, Philosophy of Mind and Language.  \n\nSupport us on Patreon for early access, exclusive content, private Discord, biweekly calls and much more! \nhttps:\/\/www.patreon.com\/mlst \n\nDonate here: https:\/\/www.paypal.com\/donate\/?hosted_button_id=K2TYRVPBGXVNA\n\nPlease email us to learn about sponsorship packages and deals. \ntim at mlst.ai (please put your budget in the subject line)\n\nPodcast booking agencies - *don't contact us* - we wouldn't even interview anyone who needed a booking agent.\nMedia\/influence agencies - *don't contact us* - we only work directly with brands\/sponsors.\n","logo":{"url":"https:\/\/yt3.ggpht.com\/15Akj76BG8IsM5ctgqVwKXArl6IfIVFAbuGa1kOomoioRgJgXHHaLmMAW7iHTMRUoEfyjTtq8lg=s800-c-k-c0x00ffffff-no-rj","width":800,"height":800,"@type":"ImageObject","@id":"https:\/\/www.youtube.com\/watch?v=EInEmGaMRLc#VideoObject_publisher_logo_ImageObject"}},"potentialAction":{"@type":"SeekToAction","@id":"https:\/\/www.youtube.com\/watch?v=EInEmGaMRLc#VideoObject_potentialAction","target":"https:\/\/www.youtube.com\/watch?v=EInEmGaMRLc&t={seek_to_second_number}","startOffset-input":"required name=seek_to_second_number"},"interactionStatistic":[[{"@type":"InteractionCounter","@id":"https:\/\/www.youtube.com\/watch?v=EInEmGaMRLc#VideoObject_interactionStatistic_WatchAction","interactionType":{"@type":"WatchAction"},"userInteractionCount":28308}],{"@type":"InteractionCounter","@id":"https:\/\/www.youtube.com\/watch?v=EInEmGaMRLc#VideoObject_interactionStatistic_LikeAction","interactionType":{"@type":"LikeAction"},"userInteractionCount":724}]},"about":["Insights","\u300eSpanish\u300f"],"wordCount":1573},{"@context":"https:\/\/schema.org\/","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Insights","item":"https:\/\/blog.terabox.com\/insights\/#breadcrumbitem"},{"@type":"ListItem","position":2,"name":"Shinka Evolve: C\u00f3mo la IA evolutiva transforma la ciencia","item":"https:\/\/blog.terabox.com\/insights\/shinka-evolve-ia-evolutiva-descubrimiento-cientifico#breadcrumbitem"}]}]