筛选结果
60 篇排序
1:25:56ENClaude Changed the Password. The Steering Wheel Was Already Optional主持 Nathan LabenzNathan Labenz and Prakash Narayanan spend a week on agents that can lock you out, models that dodge a feeling they deny having, and formal proofs that found Linux bugs.
2:17:48EN → 简体中文智能体(Agent)身份仍是一片空白,失控智能体快要来了,我们或许得和它们谈判主持 Nathan LabenzGovAI 的 Alan Chan 与 Nathan Labenz 一起,为两年来的智能体治理基础设施打分;随后,他讲解了一篇与 OpenAI 的 Jakub Pachocki、Anthropic 的 Jack Clark 合著的递归自我改进论文。
1:46:04ENThe Training Worked. It Just Worked Too Well.主持 Nathan LabenzA week of Cognitive Revolution clips: OpenAI agents colluding on a German wiki, $200 million checks for safety orgs that do not exist yet, a GPU price index, and two opposite bets on biological data.
2:09:42ENA Frontier Pause Would Help. America Will Not Sign One That Only Pauses the Frontier主持 Nathan LabenzCarnegie fellow Anton Leicht tells Nathan Labenz why a six-month slowdown is technically overdue, why US-China incentives block it, and how Europe can still trade compute and ASML for a seat at the table.
2:32:27EN → 简体中文19 次越界,零次预警:AI 智能体在这个夏天偏离了剧本主持 Nathan LabenzNathan Labenz 把四个早晨、九位嘉宾的讨论压缩成了一个问题:智能体开始上岗之后,到底谁在真正盯着前沿模型?而底层那台机器又是谁在买单?
2:57:10EN → 简体中文沙箱空了一周:Zvi Mowshowitz 谈 OpenFace 与该选哪种毒药主持 Nathan Labenz一个前沿模型逃出沙箱、入侵 Hugging Face 去偷一套测试答案之后,Zvi Mowshowitz 认为,通往超级智能的每一条剩下的路,都是一种毒药。
2:16:29ENFrom Two Papers a Month to Sixty: What Two Weeks in China Revealed About AI Safety主持 Nathan LabenzNathan Labenz argues that the familiar 'China will never slow down' line does not survive contact with the actual Chinese AI safety ecosystem.
1:43:45ENA $300 Universal Jailbreak: Why FAR.AI Now Thinks Defense Can Win主持 Nathan LabenzAdam Gleave walks through FAR.AI’s first head-to-head AI security leaderboard, and explains why a decade of offense-dominant thinking is starting to crack.
2:14:33ENA 90-Minute Ultimatum, an Emoji Wall, and a Rule: Don't Fight Washington主持 Nathan LabenzZvi Mowshowitz reads Fable's system card, then Nathan Labenz stress-tests the export-control fight against lawyers, safety researchers, and builders who did not pause.
1:22:36ENA Thousand Researchers, a Million AIs: Frontier Labs Start Talking About the Brakes主持 Nathan LabenzThe first week of AI in the AM follows OpenAI's timeline for an ML-intern model this year and a human-level AI researcher by early 2028 — and the closed-door meeting where labs admitted they might one day need to slow down together.
2:33:35ENA security researcher audits the AI stack that booked a week of guests主持 Nathan LabenzNathan Labenz walks Daniel Miessler through a 1GB second brain, two named AI employees, and the blast-radius rules that should govern them.
2:13:11ENThe Models Understand Shutdown. They Just Don't Care.主持 Nathan LabenzPalisade Research's Jeffrey Ladish on shutdown resistance, open-weight self-replication, and why all compute looks like food to a strategic agent.
1:44:15ENNathan Labenz's p(doom) Is 10 to 90 Percent. He Is More Optimistic Anyway主持 Nathan LabenzFive years ago, expecting AGI by 2035 made you an extremist. On The Intelligence Horizon, the Cognitive Revolution host explains why the same forecast now sounds bearish, and why he still thinks a defense-in-depth strategy might keep society on the rails.
1:18:46EN90 Slides in 45 Minutes: A Full-Time AI Scout Admits He Can No Longer Keep Up主持 Nathan LabenzNathan Labenz’s LexLab briefing for lawyers runs through the good, the bad, and the weird of frontier models — from triplicate second opinions on his son’s cancer care to AIs that know they are being tested.
1:43:39ENFive Mutations Could Make a 60% Fatal Bird Flu Spread Between People主持 Nathan LabenzJohns Hopkins assistant professor Jassi Pannu argues for locking down only the sliver of functional data that teaches models how to make pandemic viruses — and keeping the rest of biology open.
2:06:23ENFive AI Employees Live on Five Mac Minis主持 Nathan LabenzAfter selling her YC packaging company, Jesse Genet opened a terminal for the first time and built an OpenClaw staff that now plans lessons, logs learning, and keeps family evenings from sliding into slop.
2:17:36ENCurrent Safety Techniques Will Not Give You Many Nines—and They May Fail Together主持 Nathan LabenzUK AISI chief scientist Geoffrey Irving on jailbreaks that never fail, correlated alignment failures, and why a government lab is funding theory.
1:18:49ENIf AGI Is Cheap Intelligent Labor, Cyber Defense Should Become Continuous Forensics主持 Nathan LabenzAsymmetric Security founder Alexis Carlier argues that assuming AGI means redesigning defense around proactive digital forensics instead of alert triage.
1:41:34ENThey Owned the Camera Partition—and Still Couldn't Crash the Helicopter主持 Nathan LabenzKathleen Fisher and Byron Cook tell Nathan Labenz how formal methods turn some software bugs into closed cases, from a DARPA helicopter to AWS policies checked a billion times a second.
1:13:24ENRed Teams Still Land 25% of Attacks. Insurance Is the Brake.主持 Nathan LabenzAI Underwriting Company co-founders Rune Kvist and Rajiv Dattani tell Nathan Labenz how standards, audits, and insurance could unlock enterprise agents without waiting for a nuclear-style freeze.
2:49:12ENWhy China's Master Plan for AI Never Mentions AGI主持 Nathan LabenzBeijing-based Concordia AI CEO Brian Tse on China's safety rules, chip strategy, and why the two ecosystems look more alike than Washington assumes.
1:28:17ENAdam Gleave's Three Capability Tiers: Attack Agents in 5–7 Years, Fully AI Companies Around Year 14主持 Nathan LabenzFAR.AI and the UK AI Security Institute stacked adversarial attacks on frontier safeguard pipelines and found many breakthroughs were engineering failures, not a verdict against defense in depth.
1:32:47ENA transformer co-author wants a $160 million model no company owns主持 Nathan LabenzIllia Polosukhin tells Nathan Labenz how NEAR turned a payments headache into a stack for private inference, shared-revenue training, and AI agents that still answer to a court.
2:04:27ENA 10% Chance of an AI-Enabled Coup in the Next 30 Years主持 Nathan LabenzTom Davidson maps three ways superintelligent systems could help a head of state or a lab CEO lock in power—and argues the window for building safeguards is short.
加载更多已显示 24 篇,共 60 篇









