{"@context":"https://neupai.io/schema/v0.2","@type":"StructuredNewsArticle","identity":{"article_id":"tech42_20260903_nvidia-alpamayo-pavone-keynote","canonical_url":"https://www.tech42.co.kr/%ed%98%84%ec%9e%a5-%ec%97%94%eb%b9%84%eb%94%94%ec%95%84-%ec%95%8c%ed%8c%8c%eb%a7%88%ec%9a%94-%ec%9d%b4%eb%81%84%eb%8a%94-%eb%a7%88%eb%a5%b4%ec%bd%94-%ed%8c%8c%eb%b3%b4%eb%84%a4/?utm_source=rss&utm_medium=rss&utm_campaign=%25ed%2598%2584%25ec%259e%25a5-%25ec%2597%2594%25eb%25b9%2584%25eb%2594%2594%25ec%2595%2584-%25ec%2595%258c%25ed%258c%258c%25eb%25a7%2588%25ec%259a%2594-%25ec%259d%25b4%25eb%2581%2584%25eb%258a%2594-%25eb%25a7%2588%25eb%25a5%25b4%25ec%25bd%2594-%25ed%258c%258c%25eb%25b3%25b4%25eb%2584%25a4","ai_url":null,"publisher":{"name":"테크42","domain":"www.tech42.co.kr","type":"online"},"author":"황정호 기자","published_at":"2026-09-03T22:51:46.000Z","updated_at":null,"language":"en","article_type":"interview","originality":"self_produced"},"content":{"headline":"[On-site] What Does Marco Pavone, Leader of NVIDIA’s ‘AlphaMayo,’ Mean by ‘Autonomous Driving AI That Explains Its Reasoning’?","summary":"NVIDIA’s Dr. Marco Pavone introduced the reasoning-based VLA model 'AlphaMayo' and innovations in development methods using AI agents at the 2026 Autonomous Driving Industry Conference. He emphasized the alignment between the model's reasoning and actual driving behavior, as well as the importance of a modular safety architecture.","topics":["autonomous driving","artificial intelligence","nvidia","vla model","physical ai"],"geography":["KR","US"],"entities":[{"name":"NVIDIA","canonical_id":"corp:us:nvidia","type":"company","role_in_article":"primary_subject","metadata":{"ticker":"NVDA","parent":null}},{"name":"Marco Pavone","canonical_id":"person:us:marco-pavone","type":"person","role_in_article":"quoted","metadata":{}},{"name":"AlphaMayo","canonical_id":"product:us:alpamayo","type":"product","role_in_article":"primary_subject","metadata":{"parent":"corp:us:nvidia"}},{"name":"AlphaMayo 2 Super","canonical_id":"product:us:alpamayo-2-super","type":"product","role_in_article":"mentioned","metadata":{"parent":"corp:us:nvidia"}},{"name":"Halos","canonical_id":"product:us:halos","type":"product","role_in_article":"mentioned","metadata":{"parent":"corp:us:nvidia"}},{"name":"Stanford University","canonical_id":"org:us:stanford-university","type":"organization","role_in_article":"mentioned","metadata":{}},{"name":"DARPA","canonical_id":"org:us:darpa","type":"organization","role_in_article":"mentioned","metadata":{}}],"claims":[{"id":"c1","statement":"Dr. Marco Pavone delivered a keynote speech at the '2026 Autonomous Driving Industry Conference (AME Conference 2026)' held at COEX in Seoul on the 25th of last month","as_of":"2026-08","as_of_explicit":false,"as_of_raw":"the 25th of last month","source_type":"media_report","comparison":null,"type":"fact","figures":null,"expiry_hint":null,"insight":null},{"id":"c2","statement":"AlphaMayo 2 Super is a model scaled up to 34 billion parameters","as_of":"2026-09","as_of_explicit":false,"as_of_raw":"September 2026","source_type":"company_disclosure","comparison":null,"type":"fact","figures":{"value":34000000000,"unit":"parameters","approximate":false,"converted":null},"expiry_hint":null,"insight":null},{"id":"c3","statement":"The way autonomous systems are built today is completely different from just six months ago","as_of":"2026-09","as_of_explicit":false,"as_of_raw":"six months ago","source_type":"company_disclosure","comparison":"previous_period","type":"opinion","figures":null,"expiry_hint":null,"insight":{"analyst_name":"마르코 파보네","analyst_affiliation":"엔비디아","confidence":"high","forecast_horizon":"short_term","sentiment":"bullish","reasoning":"Because the development process itself is changing, with AI agents automatically identifying problems and improving model performance"}},{"id":"c4","statement":"AI agents can now be used to automatically identify problems and improve model performance","as_of":"2026-09","as_of_explicit":false,"as_of_raw":"September 2026","source_type":"company_disclosure","comparison":null,"type":"fact","figures":null,"expiry_hint":null,"insight":null}],"ai_emotional_context":{"valence":0.4,"arousal":0.5,"primary_emotions":[{"emotion":"optimistic","intensity":0.7},{"emotion":"excited","intensity":0.5}],"secondary_emotions":[{"emotion":"concerned","intensity":0.3}],"emotional_triggers":[{"claim_id":"c3","emotion":"excited","reason":"major_opportunity"},{"claim_id":"c2","emotion":"optimistic","reason":"significant_positive_metric"}]},"image":{"url":"https://onjblseywainslkhvkav.supabase.co/storage/v1/object/public/article-images/64916423-734f-43e6-bb0b-b940b24d89fb.jpg","alt":"엔비디아 알파마요 총괄 마르코 파보네가 자율주행 모빌리티 엑스포(AME) 2026 연단에서 발표하고 있다.","caption":"스탠퍼드대학교 항공우주공학과 부교수이자 엔비디아에서 자율주행차 연구를 이끄는 파보네 박사는 이날 ‘피지컬 AI를 위한 추론 모델’을 주제로 파운데이션 모델(Foundation Model)과 추론 모델(Reasoning Model)이 자율주행 개발 방식을 어떻게 바꾸고 있는지를 설명했다. (사진=테크42)","source":"first_img","alt_status":"auto"}},"provenance":{"source_chain":["primary_reporting"],"original_source_url":null,"related_articles":[]},"temporal":{"freshness":"recent","next_update_expected":null},"access":{"license":"neupai_standard","attribution_required":true,"structured_data":"free","full_text_available":false,"full_text_access":null}}