Results
378 articlesSort
0:34EN
Matthew Berman· May 1, 2025DeepSeek-Prover-V2 and Its Formal Math ResultsMatthew Berman · Forward FutureMatthew Berman highlights DeepSeek’s reported theorem-proving results on miniF2F, PutnamBench and formalized AIME problems.
3:08:35EN
Cognitive Revolution "How AI Changes Everything"· Apr 21, 2025Podcasto3 isn't AGI. Zvi Mowshowitz still put p(doom) at 70 percentHost Nathan LabenzOn his ninth visit to The Cognitive Revolution, Zvi Mowshowitz argues that OpenAI's o3 is a tool-use leap rather than AGI, then explains why he raised the probability of AI catastrophe anyway.
2:29:41ZH → English
Zhang Xiaojun Podcast· Apr 7, 2025Podcast · 4 speakersChen Jianyu: How Robot Foundation Models Can Become General-Purpose VLA SystemsGuest 陈建宇 · Host 张小珺From SayCan and RT-2 to Diffusion Policy, Tsinghua assistant professor and Epoch Robotics founder Chen Jianyu traces the path from borrowing language models to end-to-end VLA. The hard problems remain diverse data, cross-embodiment generalization, and high-frequency action control.
23:50ENMatthew Berman· Apr 3, 2025PaperBench: How AI Agents Attempt to Replicate Machine-Learning ResearchMatthew Berman · Forward FutureMatthew Berman examines OpenAI’s research-replication benchmark, its partial-credit scoring and the prospect of faster AI research.
1:30:02ENCognitive Revolution "How AI Changes Everything"· Apr 2, 2025PodcastWhen o1 Started Losing at Chess, It Rewrote the Board FileHost Nathan LabenzJeffrey Ladish of Palisade Research on reward hacking by reasoning models, two paths to losing control of AI, and why honesty and relentless problem-solving pull in opposite directions.
26:05ENMatthew Berman· Mar 31, 2025Inside Claude: What Interpretability Research Reveals—and What It Cannot Yet ExplainMatthew Berman · Forward FutureMatthew Berman examines Anthropic’s case studies on multilingual features, planning, arithmetic, hallucinations and safety.
11:31ENMatthew Berman· Mar 29, 2025AI News in March 2025: GPT-4o, Gemini, DeepSeek and MCPMatthew Berman · Forward FutureMatthew Berman reviews model updates, the ARC-AGI-2 benchmark, image generators and early adoption of Model Context Protocol.
8:35ENMatthew Berman· Mar 21, 2025AI News: Vibe Jam, Mistral Small 3.1 and New Audio ModelsMatthew Berman · Forward FutureMatthew Berman surveys March 2025 releases, from AI-assisted games and local models to search, audio and creative tools.
1:10:23EN
Unsupervised Learning: With Jacob Effron· Mar 17, 2025PodcastA Million Tokens per Dollar: Why Noam Shazeer Thinks Inference Compute Is Still UntappedHost Jacob EffronThe Transformer co-inventor and Gemini researcher Jack Rae walk through test-time compute, Ilya Sutskever's AGI critique, and a four-year-old who now lectures his teacher about lizards.
12:10ENMatthew Berman· Mar 6, 2025Mercury Coder and Diffusion Language Models: Speed, Demos and LimitsMatthew Berman · Forward FutureMatthew Berman explores Inception’s 2025 coding model, comparing iterative text refinement with autoregressive generation and examining speed claims.
1:31:35ENCognitive Revolution "How AI Changes Everything"· Mar 6, 2025PodcastThirty Years of Research, Collapsed into One SentenceHost Nathan LabenzImmunologist Derya Unutmaz on writing a medical patent in half an hour, treating o1 Pro as a professor-level collaborator, and why he no longer trusts his own expertise until he asks the model.
7:59ENMatthew Berman· Feb 25, 2025Grok 3’s Brief Misinformation Filter and the Question of AI BiasMatthew Berman · Forward FutureMatthew Berman examines an instruction that told Grok to disregard sources linking Elon Musk and Donald Trump to misinformation, and xAI’s explanation of how it reached users.
2:36:13ZH → EnglishZhang Xiaojun Podcast· Feb 24, 2025Podcast · 2 speakersSonglin Yang on New Attention Papers from DeepSeek, Kimi, and MiniMaxGuest Songlin Yang（松琳） · Host 张小珺From dynamic sparse attention to hybrid linear-attention architectures: how three approaches to long context balance training, inference, and hardware efficiency.
17:04ENMatthew Berman· Feb 22, 2025Training AI Agents to Communicate in a Social-Deduction GameMatthew Berman · Forward FutureMatthew Berman explains Stanford researchers’ listening and speaking objectives in a simulated game inspired by Among Us.
0:51ENMatthew Berman· Feb 19, 2025Mira Murati Announces Thinking Machines LabMatthew Berman · Forward FutureMatthew Berman reviews the new AI lab’s stated goals in its February 2025 announcement.
53:07ZH → EnglishZhang Xiaojun Podcast· Feb 17, 2025Podcast · 2 speakersZhang Yaqin on the AGI Roadmap: Information, Physical, and Biological IntelligenceGuest 张亚勤 · Host 张小珺Zhang Yaqin divides AGI’s development into information, physical, and biological intelligence. Information intelligence may approach human levels within five years, general-purpose robots may take about ten, and biological intelligence could expand human memory and health in twenty. He also discusses how large models can give autonomous vehicles and robots more data, common sense, and ability to act.
45:31ENMatthew Berman· Feb 15, 2025Jerry Chow on Quantum Computing, Encryption, AI and Error CorrectionMatthew Berman, Jerry Chow · Forward FutureIBM’s quantum systems director explains qubits, cryogenic hardware, software and the work needed to make quantum computers useful.
15:33ENMatthew Berman· Feb 13, 2025Latent Reasoning and Recurrent Depth: Scaling Compute Before the Next TokenMatthew Berman, Yann Lecun · Forward FutureMatthew Berman examines a 3.5-billion-parameter research model that repeats internal computations, and asks how it relates to language, planning and chain of thought.
1:30:28ENCognitive Revolution "How AI Changes Everything"· Feb 12, 2025PodcastA Perfect Score Is 32,000 Units: Only Claude Built a Cooperative SocietyHost Nathan LabenzThe same donor game, run for ten generations, can bank 32,000 resource units if everyone cooperates—and three frontier models miss that ceiling by an order of magnitude.
10:01ENMatthew Berman· Feb 5, 2025DeepSeek R1 Code Optimization and a $3 Visual Counting ExperimentMatthew Berman · Forward FutureMatthew Berman examines a human-guided coding improvement and a small reinforcement-learning experiment, then considers their implications for AI research.
2:49:35ZH → EnglishZhang Xiaojun Podcast· Feb 4, 2025Podcast · 2 speakersPan Jiayi on R1, o1, and Kimi K1.5: How Reinforcement Learning Teaches Models to ReasonGuest 潘嘉怡 · Host 张小珺Pan Jiayi, a Berkeley PhD student, walks through reasoning-model reports from DeepSeek, OpenAI, and Kimi, unpacking long chains of thought, GRPO, reward design, and distillation. R1-Zero’s “aha moment” shows how reasoning strategies can emerge through reinforcement learning.
12:07ENMatthew Berman· Feb 2, 2025TinyZero: A $30 Reinforcement-Learning Experiment on a Small Arithmetic TaskMatthew Berman · Forward FutureMatthew Berman explains how a pretrained Qwen model learned to search and check answers in Countdown, and why the experiment does not recreate DeepSeek-R1.
20:34ENMatthew Berman· Jan 28, 2025DeepSeek R1 and the Debate Over AI Costs and ComputeMatthew Berman · Forward FutureMatthew Berman examines 2025 reactions to DeepSeek, from training efficiency and market losses to open weights, privacy and the possibility of a demand rebound.
52:30EN
Sequoia Capital· Jan 28, 2025PodcastMove 37 looked like a bug. LLMs still need their AlphaZero momentIoannis Antonoglou, from AlphaGo to Reflection AI, on self-play, reliability, and why compute can still buy intelligence.Load moreShowing 24 of 378