Shows
Gli articoli di Krisis.infoLettera di Natale ai lettoridi Elisabetta BurbaCare lettrici, cari lettori,alla fine del primo anno di vita di Krisis, sento il bisogno di rivolgermi a voi non solo per gli auguri, ma per condividere qualche pensiero sul percorso fatto insieme.In 12 mesi, abbiamo trovato una forma più definita: abbiamo chiarito il nostro perimetro d’azione e allargato la rete dei collaboratori, intrecciando relazioni a livello globale. Non abbiamo costruito solo un eccellente rapporto personale con Emmanuel Todd e Jeffrey Sachs. Il nostro network si estende ai quattro angoli del pianeta: dalla Croazia del vescovo serbo-ortodosso Jovan Ćulibrk, all’Australia della giornalista Caitlin Johnstone; dalla Svezia...
2025-12-2406 min
Daily Paper CastScaling Instruction-Based Video Editing with a High-Quality Synthetic Dataset
🤗 Upvotes: 37 | cs.CV
Authors:
Qingyan Bai, Qiuyu Wang, Hao Ouyang, Yue Yu, Hanlin Wang, Wen Wang, Ka Leong Cheng, Shuailei Ma, Yanhong Zeng, Zichen Liu, Yinghao Xu, Yujun Shen, Qifeng Chen
Title:
Scaling Instruction-Based Video Editing with a High-Quality Synthetic Dataset
Arxiv:
http://arxiv.org/abs/2510.15742v1
Abstract:
Instruction-based video editing promises to democratize content creation, yet its progress is severely hampered by the scarcity of large-scale, high-quality training data. We introduce Ditto, a holistic framework designed to tackle this fundamental challenge. At its heart, Ditto features a n...
2025-10-2120 min
Gli articoli di Krisis.infoAudio - «Smettetela di piangere sulla “trappola cinese”»di Zichen WangLa guerra commerciale fra Washington e Pechino ha intensificato il dibattito sul ruolo di Pechino sulla scena economica mondiale. Per capire il punto di vista cinese, Krisis propone un articolo in cui il fondatore della newsletter «Pekingnology» ribalta la narrazione occidentale che dipinge Pechino come artefice di una «trappola» per le aziende occidentali. Una lettura controcorrente che interpreta l’ascesa tecnologica cinese come risultato della competizione globale e non come risultato di una manovra ingannevole.Zichen Wang Esperto di comunicazione internazionale e politiche pubbliche, Zichen Wang è Research Fellow e Direttore per la Comunicazione Internazionale presso il Center for China a...
2025-10-1812 minDaily Paper CastAI for Service: Proactive Assistance with AI Glasses
🤗 Upvotes: 60 | cs.AI, cs.CL, cs.CV
Authors:
Zichen Wen, Yiyu Wang, Chenfei Liao, Boxue Yang, Junxian Li, Weifeng Liu, Haocong He, Bolong Feng, Xuyang Liu, Yuanhuiyi Lyu, Xu Zheng, Xuming Hu, Linfeng Zhang
Title:
AI for Service: Proactive Assistance with AI Glasses
Arxiv:
http://arxiv.org/abs/2510.14359v1
Abstract:
In an era where AI is evolving from a passive tool into an active and adaptive companion, we introduce AI for Service (AI4Service), a new paradigm that enables proactive and real-time assistance in daily life. Existing AI...
2025-10-1824 minDaily Paper CastEfficient Multi-modal Large Language Models via Progressive Consistency Distillation
🤗 Upvotes: 30 | cs.CV
Authors:
Zichen Wen, Shaobo Wang, Yufa Zhou, Junyuan Zhang, Qintong Zhang, Yifeng Gao, Zhaorun Chen, Bin Wang, Weijia Li, Conghui He, Linfeng Zhang
Title:
Efficient Multi-modal Large Language Models via Progressive Consistency Distillation
Arxiv:
http://arxiv.org/abs/2510.00515v1
Abstract:
Visual tokens consume substantial computational resources in multi-modal large models (MLLMs), significantly compromising their efficiency. Recent works have attempted to improve efficiency by compressing visual tokens during training, either through modifications to model components or by introducing additional parameters. However, they often overlook the inc...
2025-10-0721 minDaily Paper CastGEM: A Gym for Agentic LLMs
🤗 Upvotes: 53 | cs.LG, cs.AI, cs.CL
Authors:
Zichen Liu, Anya Sims, Keyu Duan, Changyu Chen, Simon Yu, Xiangxin Zhou, Haotian Xu, Shaopan Xiong, Bo Liu, Chenmien Tan, Chuen Yang Beh, Weixun Wang, Hao Zhu, Weiyan Shi, Diyi Yang, Michael Shieh, Yee Whye Teh, Wee Sun Lee, Min Lin
Title:
GEM: A Gym for Agentic LLMs
Arxiv:
http://arxiv.org/abs/2510.01051v1
Abstract:
The training paradigm for large language models (LLMs) is moving from static datasets to experience-based learning, where agents acquire skills via interacting with complex env...
2025-10-0325 minDaily Paper CastWinning the Pruning Gamble: A Unified Approach to Joint Sample and Token Pruning for Efficient Supervised Fine-Tuning
🤗 Upvotes: 57 | cs.CL
Authors:
Shaobo Wang, Jiaming Wang, Jiajun Zhang, Cong Wang, Yue Min, Zichen Wen, Fei Huang, Huiqiang Jiang, Junyang Lin, Dayiheng Liu, Linfeng Zhang
Title:
Winning the Pruning Gamble: A Unified Approach to Joint Sample and Token Pruning for Efficient Supervised Fine-Tuning
Arxiv:
http://arxiv.org/abs/2509.23873v1
Abstract:
As supervised fine-tuning (SFT) evolves from a lightweight post-training step into a compute-intensive phase rivaling mid-training in scale, data efficiency has become critical for aligning large language models (LLMs) under tight budgets. Existing data pruning methods suf...
2025-10-0220 minDaily Paper CastVariational Reasoning for Language Models
🤗 Upvotes: 51 | cs.CL, cs.AI, cs.LG
Authors:
Xiangxin Zhou, Zichen Liu, Haonan Wang, Chao Du, Min Lin, Chongxuan Li, Liang Wang, Tianyu Pang
Title:
Variational Reasoning for Language Models
Arxiv:
http://arxiv.org/abs/2509.22637v1
Abstract:
We introduce a variational reasoning framework for language models that treats thinking traces as latent variables and optimizes them through variational inference. Starting from the evidence lower bound (ELBO), we extend it to a multi-trace objective for tighter bounds and propose a forward-KL formulation that stabilizes the training of the var...
2025-09-3022 minDaily Paper CastLanguage Models Can Learn from Verbal Feedback Without Scalar Rewards
🤗 Upvotes: 48 | cs.CL, cs.AI, cs.LG
Authors:
Renjie Luo, Zichen Liu, Xiangyan Liu, Chao Du, Min Lin, Wenhu Chen, Wei Lu, Tianyu Pang
Title:
Language Models Can Learn from Verbal Feedback Without Scalar Rewards
Arxiv:
http://arxiv.org/abs/2509.22638v1
Abstract:
LLMs are often trained with RL from human or AI feedback, yet such methods typically compress nuanced feedback into scalar rewards, discarding much of their richness and inducing scale imbalance. We propose treating verbal feedback as a conditioning signal. Inspired by language priors in text-to-image gen...
2025-09-3023 minDaily Paper CastScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data
🤗 Upvotes: 81 | cs.CV
Authors:
Zhaoyang Liu, JingJing Xie, Zichen Ding, Zehao Li, Bowen Yang, Zhenyu Wu, Xuehui Wang, Qiushi Sun, Shi Liu, Weiyun Wang, Shenglong Ye, Qingyun Li, Zeyue Tian, Gen Luo, Xiangyu Yue, Biqing Qi, Kai Chen, Bowen Zhou, Yu Qiao, Qifeng Chen, Wenhai Wang
Title:
ScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data
Arxiv:
http://arxiv.org/abs/2509.15221v1
Abstract:
Vision-Language Models (VLMs) have enabled computer use agents (CUAs) that operate GUIs autonomously, showing great potential, yet progress is limited by the lack of lar...
2025-09-2023 min
The NZIIA PodcastA changing international (dis)order: implications for New Zealand & the regionThis discussion occurred at the NZIIA Conference in Wellington in June 2025. For decades, New Zealand’s approach to international relations has been driven by a pursuit of multilateralism, collective security and free trade. Since the end of the Cold War, this has been enabled to some extent by the pre-eminence of US power. However, the international order and the institutions underpinning it, are changing. The rise of China and other Asian powers is altering international affairs in profound ways, sparking competition and instability. Erosion of democratic norms and authoritarianism is also making indelible marks on world affairs. Lik...
2025-09-021h 20Daily Paper CastInternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
🤗 Upvotes: 120 | cs.CV
Authors:
Weiyun Wang, Zhangwei Gao, Lixin Gu, Hengjun Pu, Long Cui, Xingguang Wei, Zhaoyang Liu, Linglin Jing, Shenglong Ye, Jie Shao, Zhaokai Wang, Zhe Chen, Hongjie Zhang, Ganlin Yang, Haomin Wang, Qi Wei, Jinhui Yin, Wenhao Li, Erfei Cui, Guanzhou Chen, Zichen Ding, Changyao Tian, Zhenyu Wu, Jingjing Xie, Zehao Li, Bowen Yang, Yuchen Duan, Xuehui Wang, Songze Li, Xiangyu Zhao, Haodong Duan, Nianchen Deng, Bin Fu, Yinan He, Yi Wang, Conghui He, Botian Shi, Junjun He, Yingtong Xiong, Han Lv, Lijun Wu, Wenqi Shao, Kaipeng Zhang, Huipeng Deng, Biqing Qi, Jiaye Ge, Qip...
2025-08-2723 minDaily Paper CastGLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models
🤗 Upvotes: 79 | cs.CL
Authors:
GLM-4. 5 Team, :, Aohan Zeng, Xin Lv, Qinkai Zheng, Zhenyu Hou, Bin Chen, Chengxing Xie, Cunxiang Wang, Da Yin, Hao Zeng, Jiajie Zhang, Kedong Wang, Lucen Zhong, Mingdao Liu, Rui Lu, Shulin Cao, Xiaohan Zhang, Xuancheng Huang, Yao Wei, Yean Cheng, Yifan An, Yilin Niu, Yuanhao Wen, Yushi Bai, Zhengxiao Du, Zihan Wang, Zilin Zhu, Bohan Zhang, Bosi Wen, Bowen Wu, Bowen Xu, Can Huang, Casey Zhao, Changpeng Cai, Chao Yu, Chen Li, Chendi Ge, Chenghua Huang, Chenhui Zhang, Chenxi Xu, Chenzheng Zhu, Chuang Li, Congfeng Yin, Daoyan Lin, Dayong Yang, Dazhi Jiang, Din...
2025-08-1221 minDaily Paper CastChemDFM-R: An Chemical Reasoner LLM Enhanced with Atomized Chemical Knowledge
🤗 Upvotes: 21 | cs.CE, cs.AI
Authors:
Zihan Zhao, Bo Chen, Ziping Wan, Lu Chen, Xuanze Lin, Shiyang Yu, Situo Zhang, Da Ma, Zichen Zhu, Danyang Zhang, Huayang Wang, Zhongyang Dai, Liyang Wen, Xin Chen, Kai Yu
Title:
ChemDFM-R: An Chemical Reasoner LLM Enhanced with Atomized Chemical Knowledge
Arxiv:
http://arxiv.org/abs/2507.21990v2
Abstract:
While large language models (LLMs) have achieved impressive progress, their application in scientific domains such as chemistry remains hindered by shallow domain understanding and limited reasoning capabilities. In this work, we focus on the...
2025-07-3122 minDaily Paper CastThe Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs
🤗 Upvotes: 45 | cs.CL
Authors:
Zichen Wen, Jiashu Qu, Dongrui Liu, Zhiyuan Liu, Ruixi Wu, Yicun Yang, Xiangqi Jin, Haoyun Xu, Xuyang Liu, Weijia Li, Chaochao Lu, Jing Shao, Conghui He, Linfeng Zhang
Title:
The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs
Arxiv:
http://arxiv.org/abs/2507.11097v1
Abstract:
Diffusion-based large language models (dLLMs) have recently emerged as a powerful alternative to autoregressive LLMs, offering faster inference and greater interactivity via parallel decoding and bidirectional modeling. However, despite strong performance in code generation and tex...
2025-07-2219 minDaily Paper CastSPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learning
🤗 Upvotes: 27 | cs.AI, cs.CL, cs.LG
Authors:
Bo Liu, Leon Guertler, Simon Yu, Zichen Liu, Penghui Qi, Daniel Balcells, Mickel Liu, Cheston Tan, Weiyan Shi, Min Lin, Wee Sun Lee, Natasha Jaques
Title:
SPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learning
Arxiv:
http://arxiv.org/abs/2506.24119v2
Abstract:
Recent advances in reinforcement learning have shown that language models can develop sophisticated reasoning through training on tasks with verifiable rewards, but these approaches depend on human-curated problem-answer pairs and domain-specific reward engineering. We int...
2025-07-0220 minDaily Paper CastCalligrapher: Freestyle Text Image Customization
🤗 Upvotes: 24 | cs.CV
Authors:
Yue Ma, Qingyan Bai, Hao Ouyang, Ka Leong Cheng, Qiuyu Wang, Hongyu Liu, Zichen Liu, Haofan Wang, Jingye Chen, Yujun Shen, Qifeng Chen
Title:
Calligrapher: Freestyle Text Image Customization
Arxiv:
http://arxiv.org/abs/2506.24123v1
Abstract:
We introduce Calligrapher, a novel diffusion-based framework that innovatively integrates advanced text customization with artistic typography for digital calligraphy and design applications. Addressing the challenges of precise style control and data dependency in typographic customization, our framework incorporates three key technical contributions. First, we develop a self-distillation mec...
2025-07-0222 min
Odd LotsZichen Wang's Exit Interview From AmericaZichen Wang is the writer of the Pekingnology newsletter, which translates important speeches and articles from China into English, and contextualizes them for Western readers. Over the past year, he's been a master's degree student at Princeton University, although he's recently returned home to resume his career at a think tank in China. His stint in the US obviously came at a very interesting time, both due to the rising US-China tensions, and also the growing restrictions on Chinese students in the US. So before making his trip back home he joined us for another episode of Odd Lots. We d...
2025-06-2046 min
Odd LotsZichen Wang's Exit Interview From AmericaZichen Wang is the writer of the Pekingnology newsletter, which translates important speeches and articles from China into English, and contextualizes them for Western readers. Over the past year, he's been a master's degree student at Princeton University, although he's recently returned home to resume his career at a think tank in China. His stint in the US obviously came at a very interesting time, both due to the rising US-China tensions, and also the growing restrictions on Chinese students in the US. So before making his trip back home he joined us for another episode of Odd Lots. W...
2025-06-2046 minDaily Paper CastSynthRL: Scaling Visual Reasoning with Verifiable Data Synthesis
🤗 Upvotes: 46 | cs.LG, cs.CL, cs.CV
Authors:
Zijian Wu, Jinjie Ni, Xiangyan Liu, Zichen Liu, Hang Yan, Michael Qizhe Shieh
Title:
SynthRL: Scaling Visual Reasoning with Verifiable Data Synthesis
Arxiv:
http://arxiv.org/abs/2506.02096v1
Abstract:
Vision-language models (VLMs) trained via reinforcement learning with verifiable reward (RLVR) have shown notable progress in scaling test-time compute effectively. In this work, we investigate how synthesized RL data can further improve RLVR. To this end, we propose \textbf{SynthRL}-a scalable and guaranteed pipeline for automatic data scaling in rea...
2025-06-0518 min
The TakeWhy is Trump cracking down on Chinese students?The US will begin revoking visas for Chinese students, including those in “critical fields” or with connections to the Chinese Communist Party. The US State Department has said this is to stop the exploitation of US universities and protect national security. Who is losing out in this latest development in US-China tensions? In this episode: Zichen Wang (@ZichenWanghere), Research Fellow & Director for International Communications at the Center for China and Globalization Episode credits: This episode was produced by Sarí el-Khalili, Sonia Bhagat, and Chloe K. Ki, with Phillip Lanos, Spencer Cline, Remas Alhawari, Maria...
2025-06-0423 minDaily Paper CastScienceBoard: Evaluating Multimodal Autonomous Agents in Realistic Scientific Workflows
🤗 Upvotes: 85 | cs.AI, cs.CL, cs.CV, cs.HC
Authors:
Qiushi Sun, Zhoumianze Liu, Chang Ma, Zichen Ding, Fangzhi Xu, Zhangyue Yin, Haiteng Zhao, Zhenyu Wu, Kanzhi Cheng, Zhaoyang Liu, Jianing Wang, Qintong Li, Xiangru Tang, Tianbao Xie, Xiachong Feng, Xiang Li, Ben Kao, Wenhai Wang, Biqing Qi, Lingpeng Kong, Zhiyong Wu
Title:
ScienceBoard: Evaluating Multimodal Autonomous Agents in Realistic Scientific Workflows
Arxiv:
http://arxiv.org/abs/2505.19897v1
Abstract:
Large Language Models (LLMs) have extended their impact beyond Natural Language Processing, substantially fostering the development of interdisciplinary res...
2025-05-2922 minDaily Paper CastShifting AI Efficiency From Model-Centric to Data-Centric Compression
🤗 Upvotes: 124 | cs.CL, cs.AI, cs.CV
Authors:
Xuyang Liu, Zichen Wen, Shaobo Wang, Junjie Chen, Zhishan Tao, Yubo Wang, Xiangqi Jin, Chang Zou, Yiyu Wang, Chenfei Liao, Xu Zheng, Honggang Chen, Weijia Li, Xuming Hu, Conghui He, Linfeng Zhang
Title:
Shifting AI Efficiency From Model-Centric to Data-Centric Compression
Arxiv:
http://arxiv.org/abs/2505.19147v1
Abstract:
The rapid advancement of large language models (LLMs) and multi-modal LLMs (MLLMs) has historically relied on model-centric scaling through increasing parameter counts from millions to hundreds of billions to drive performance gai...
2025-05-2822 minDaily Paper CastOptimizing Anytime Reasoning via Budget Relative Policy Optimization
🤗 Upvotes: 30 | cs.LG, cs.AI, cs.CL
Authors:
Penghui Qi, Zichen Liu, Tianyu Pang, Chao Du, Wee Sun Lee, Min Lin
Title:
Optimizing Anytime Reasoning via Budget Relative Policy Optimization
Arxiv:
http://arxiv.org/abs/2505.13438v1
Abstract:
Scaling test-time compute is crucial for enhancing the reasoning capabilities of large language models (LLMs). Existing approaches typically employ reinforcement learning (RL) to maximize a verifiable reward obtained at the end of reasoning traces. However, such methods optimize only the final performance under a large and fixed token budget, which hin...
2025-05-2221 min
Asia Inside OutZichen Wang on the Evolving U.S.-China RelationsZichen Wang, founder and editor of the Pekingnology newsletter and research fellow and director for International Relations at the Center for China and Globalization (CCG), speaks with Rorry Daniels, managing director of Asia Society Policy Institute, on our latest episode of Asia Inside Out. Wang and Daniels discuss the state of U.S.-China relations one month after “Liberation Day”, the future of the relationship, and the importance of maintaining people-to-people connections. Asia Inside Out brings together our team and special guests to take you beyond the latest policy headlines and provide an insider’s view on...
2025-05-0556 min
China Biz Talk - Baiguan RadioBaiguan Radio #28: Trade war, Boeing, Jensen Huang, high-sea fishing, Xi's Southeast Asia tripWelcome to a new collaboration episode between Baiguan, Zichen Wang of Pekingnology and Johnny Zou of East8’s Newsletter. It has been several weeks now since we shared our thoughts on the important events of the past week. In this episode, we cover: [00:41] Robert’s experience of publishing an opinion piece on the New York Times discussing China’s resolve in the trade war.[06:32] Significance of China halting Boeing deliveries[8:08] Jensen Huang’s China visit: Is it a slap on Trump’s face, or is Jensen actually the “go-between”?[17:13] Bloomberg’s st...
2025-04-2036 min
Peking PlaybookTariff, preconditions to talk, Jensen Huang's China visit, Xi in Southeast AsiaHello! Welcome to the 9th episode of the podcast “Peking Playbook,” featuring me (Zichen Wang), Robert Wu of Baiguan and China Translated substacks, and Johnny Zou of East8. [00:41] Robert’s experience of publishing a guest essay on the New York Times discussing China’s resolve in the trade war.[06:32] Significance of China halting Boeing deliveries[8:08] Jensen Huang’s China visit: Is it a slap on Trump’s face, or could the Nvidia be some sort of “go-between”?[17:13] Bloomberg’s scoop about China’s willingness to talk, with certain preconditions. How to take it?[19:1...
2025-04-2036 min
Peking PlaybookHu Bo on South China Sea and U.S. military activities in 2024Many China-focused English-language podcasts are online, mostly run by Western institutions and individuals. Some have tried hard to include Chinese mainland guests in their conversations. I applaud their efforts. For various reasons, their success is uneven. I guess—or I’d like to believe—I may be in a better position to bring some Chinese voices to a global audience, so let me have a try—starting with today’s guest, Hu Bo, Research Professor and Director of the Centre for Maritime Strategy Studies of Peking University. Professor Hu is perhaps better known as the Director o...
2025-04-1945 min
Peking PlaybookTrump's Tariff on China 20250413Hello! Welcome to the 8th episode of the podcast “Peking Playbook,” featuring me (Zichen Wang), Robert Wu of Baiguan and China Translated substacks, and Johnny Zou of East8. [3:30] Did Trump plan for all of this? [12:42] Optimism about China-EU relationship [21:06] Pessimism about China-EU relationship [23:10] Will China see Trump’s sexemption for consumer electronics as a chance to soften its stance? [27:13] Who in the Trump team can actually call the Chinese side now? [30:31] The hope for China’s domestic demand stimulation[35:11] What’s the right policy to boost dome...
2025-04-1348 min
China Biz Talk - Baiguan RadioTariff war - Week 2 - Baiguan Radio #27Welcome to another collaboration episode between Baiguan, Pekingnology’s Zichen Wang and East8’s Johnny Zou. Highlights[3:30] Did Trump plan for all of this?[12:42] Optimism about China-EU relationship[21:06] Pessimism about China-EU relationship[23:10] Will China see Trump’s “electronic exemption” as a chance to soften its stance?[27:13] Who in the Trump team can actually call the Chinese side now?[30:31] The hope for China’s domestic demand policies[35:11] What’s the right policy to boost domestic demand?[39:58] Where does the money to boost domestic deman...
2025-04-1348 min
Peking PlaybookBeijing will not back down, I told Foreign Policy Research Institute audienceOn Monday, April 7, I was invited to the Foreign Policy Research Institute (FPRI), a 70-year-old nonpartisan think tank, to talk in front of an audience with Neysun A. Mahboubi, a Non-Resident Senior Fellow with the FPRI Asia Program as well as Director of the Penn Project on the Future of US-China Relations at the University of Pennsylvania.The title of the program A Grand Bargain Between the US and China? was finalized less than a month ago, and I was called in at the 11th hour to take the stage. In the hour-long program, which also includes...
2025-04-101h 08
Follow the RabbitFrom Davos to your Inbox: Narratives, Newsletters, and Nuances with Adrian MonckHow the former Davos communications chief navigates the tension between algorithmic demands and journalistic integrity in his quest to make sense of a world that resists coherenceIn this episode of Follow the Rabbit, we explore the changing landscape of cultural impact with Adrian Monck, whose newsletter "Seven Things" has become an intriguing case study in post-institutional thought leadership. After working in television journalism, academia, and finally as the World Economic Forum's communications architect for 13 years, Monck now creates a weekly geopolitical digest that reaches over 160,000 readers—all while earning "nothing" and negotiating holidays with himself....
2025-04-0943 min
Peking PlaybookPeking Playbook Unscripted 20250406Hello! Welcome to the 7th episode of the podcast “Peking Playbook,” mainly featuring me (Zichen Wang) and Robert Wu of Baiguan and China Translated substacks, just talking about recent news. Today, we have Johnny Zou of East8 join us as well. Johnny was with us on the January 28 episode.In the following 43-minute conversation recorded on Sunday morning ET, April 6, 2025, we talked about:[00:46] Tariff, first impressions[09:46] Is the US underestimating China?[15:00] How would countries from Japan to Vietnam respond?[22:44] Will China have closer integration with the rest of the worl...
2025-04-0743 min
China Biz Talk - Baiguan RadioEP26. TARIFF!!! - With Zichen Wang and Johnny ZouThis week, Zichen Wang of Pekingnology and I are joined by Johnny Zou, a leading Hong Kong-based venture capitalist who also writes at East8’s Newsletter (#97 rising business Substack right now, previously known as MJ Finance). We discussed last week's most important news item: Trump’s tariff war and its implications for China.Highlights[00:46] Tariff, first impressions[09:46] Is the US underestimating China?[15:00] How would middle-powers from Japan to Vietnam respond?[22:44] Will China have closer integration with the rest of the world or more trade friction?[29:27] Invest...
2025-04-0643 minDaily Paper CastUnderstanding R1-Zero-Like Training: A Critical Perspective
🤗 Upvotes: 25 | cs.LG, cs.AI, cs.CL
Authors:
Zichen Liu, Changyu Chen, Wenjun Li, Penghui Qi, Tianyu Pang, Chao Du, Wee Sun Lee, Min Lin
Title:
Understanding R1-Zero-Like Training: A Critical Perspective
Arxiv:
http://arxiv.org/abs/2503.20783v1
Abstract:
DeepSeek-R1-Zero has shown that reinforcement learning (RL) at scale can directly enhance the reasoning capabilities of LLMs without supervised fine-tuning. In this work, we critically examine R1-Zero-like training by analyzing its two core components: base models and RL. We investigate a wide range of base mod...
2025-04-0419 min
Peking PlaybookPeking Playbook Unscripted 20250330Hello! Welcome to the 6th episode of the podcast “Peking Playbook,” mainly featuring me (Zichen Wang) and Robert Wu, of Baiguan and China Translated substacks, just talking about recent news.In the following 39-minute conversation recorded on Saturday evening ET, March 29, 2025, we talked about:[00:52] Xi met global CEOs [04:25] What’s more serious for foreign businesses in China, organic domestic competition or the alleged policy discrimination? [10:00] Some Western firms (such as top consulting firms) are facing pressure at home to do business with China [12:06] There is also discrimination against local busine...
2025-03-3039 min
The US-China PodcastHow Can People in the United States and China Better Understand Each Other? The decline in U.S.-China people-to-people exchange opportunities between the United States and China is widening the gap of cross-cultural understanding, leaving room for misconceptions in spite of many shared concerns and aspirations. How can people in the two countries learn about their differences and similarities, and what are the key things they need to know to better understand one another? Zichen Wang joins the National Committee in an interview recorded on Feb 28, 2025, to explain the mixed signals both sides are sending to each other, and how both the United States and C...
2025-03-2823 min
Peking PlaybookPeking Playbook Unscripted 20250325Hello! Welcome to the 5th episode of the podcast “Peking Playbook,” mainly featuring me and Robert Wu, of Baiguan and China Translated, just talking about recent news.In the following half-hour conversation recorded on Sunday evening ET, March 23, 2025, we talked about[01:32] Some observations during Zichen’s ongoing study at the mid-career Master in Public Policy (MPP) program at Princeton School of Public and International Affairs[04:40] The ongoing controversy regarding CK Hutchison’s coerced Panama ports sale [20:16] New top-level policy document dedicated to boosting domestic demand[27:03] China Development Forum[29:16...
2025-03-2534 min
China Biz Talk - Baiguan RadioEP24. Li Ka Shing, domestic consumption, Xi-Trump meeting rumors - Baiguan X PekingnologyHi everyone, welcome to another collaboration episode between Pekingnology and Baiguan. Quick notice: we heard some complaints about the audio quality of our episodes recently, and so we have made special efforts to improve the experience this time. We hope you enjoy it better! In this episode, you will hear about:[01:32] Zichen Wang’s observations during his study at Princeton[04:40] The ongoing controversy regarding CK Hutchison’s Panama ports sale [20:16] New top-level policy on boosting domestic demand[27:03] China Development Forum[29:16] Will Xi and Trump meet soon...
2025-03-2434 min
Podcast di Palazzo Ducale di GenovaFestival Limes "Fra Genova e Trieste: vie della seta contro via dell’oro"9 febbraio 2025 - Carlos Roa, Laris Gaiser, Zichen Wang, Massimo Deandreis e Luca Sisto
2025-03-121h 25
Peking PlaybookPeking Playbook 20250309Hello! Welcome to the 4th episode of the on-and-off podcast that has been tentatively named “Peking Playbook,” mainly featuring me and Robert Wu, of Baiguan and China Translated, just talking about recent news.In the following half-hour conversation recorded on Sunday morning ET, March 9, 2025, we touched on[00:56] Two Sessions?[03:31] Birth subsidies[06:33] 1-trillion-yuan state VC fund?[11:33] China’s retaliatory tariff on US agricultural products[17:32] Trump’s impact on stock market confidence[19:47] TSMC’s $100b investment in the USAt the time of the recording, I was in P...
2025-03-1029 min
Podcast di Palazzo Ducale di GenovaFestival Limes "USA vs Cina"8 febbraio 2025 - Scott Smitson e Zichen Wang
2025-03-061h 26
Peking PlaybookPeking Playbook Unscripted 20250223Hello! Welcome to the 3rd episode of the on-and-off podcast that has been tentatively named “Peking Playbook,” mainly featuring me and Robert Wu, of Baiguan and China Translated, just talking about recent news.In the following half-hour conversation recorded on Sunday morning ET, February 23, 2025, we touched onTimeline:[01:14] Chinese leaders’ symposium with private entrepreneurs[15:49] Market rallies in Hong Kong and China ADRs[21:29] New White House memo on restricting Chinese investment[27:22] U-turn of U.S. policy on UkraineListen to Peking Playbook on Apple Podcast, Spotify, YouTub...
2025-02-2331 minDaily Paper CastSURGE: On the Potential of Large Language Models as General-Purpose Surrogate Code Executors
🤗 Upvotes: 11 | cs.LG, cs.CL
Authors:
Bohan Lyu, Siqiao Huang, Zichen Liang
Title:
SURGE: On the Potential of Large Language Models as General-Purpose Surrogate Code Executors
Arxiv:
http://arxiv.org/abs/2502.11167v1
Abstract:
Large language models (LLMs) have demonstrated remarkable capabilities in code-related tasks, such as code understanding and code generation. However, an equally important yet underexplored question is whether LLMs can serve as general-purpose surrogate code executors, to predict the output and behavior of a program without actually running it. To systematically investigate this capability, we int...
2025-02-1923 min
Asia In-DepthHow We Talk About China, with Kaiser KuoKaiser Kuo, host of the long-running Sinica podcast, joins us in Zurich to discuss the evolving narrative on China, and why the West is so often 'stunned' by Chinese breakthroughs in electric vehicles, green technology, or, most recently: AI, with DeepSeek.We explore the reasons behind the challenges of reporting on China, the importance of strategic empathy, and the shifting dynamics of global influence, particularly in relation to American exceptionalism.Recommendations mentioned at the end of the episode:Chartbook, by Adam ToozePekingnology and The East is Read, by Zichen...
2025-02-0442 min
State of AsiaHow We Talk About China, with Kaiser KuoKaiser Kuo, host of the long-running Sinica podcast, joins us in Zurich to discuss the evolving narrative on China, and why the West is so often 'stunned' by Chinese breakthroughs in electric vehicles, green technology, or, most recently: AI, with DeepSeek.We explore the reasons behind the challenges of reporting on China, the importance of strategic empathy, and the shifting dynamics of global influence, particularly in relation to American exceptionalism.Recommendations mentioned at the end of the episode:Chartbook, by Adam ToozePekingnology and The East is Read, by Zichen WangInterconnected, by K...
2025-02-0442 min
Peking PlaybookThe State of China, with Adam Tooze and Qing Wang — Moderated by Finbarr Bermingham of SCMPImmediately after the U.S. elections in November 2024, I spoke at a panel at the STATE OF ASIA 2024 in Zurich organized by Asia Society Switzerland, alongside Wang Qing (王罄), the host of the popular Chinese podcast "The Weirdo" (不合时宜) and Adam Tooze, economic historian and professor at Columbia University and author of Chartbook. The discussion was moderated by Finbarr Birmingham, Europe correspondent for the South China Morning Post.Asia Society Switzerland has recently published the discussion as a podcast, and I’m happily syndicating it here. This is a public episode. If you would like to discus...
2025-02-0246 min
Sinica PodcastThe State of China, with Adam Tooze, Qing Wang, and Zichen Wang — Moderated by Finbarr Bermingham of SCMPHappy Chinese New Year! This week, while I'm decompressing from 10 days in the Alps, my friends at the Asia Society of Switzerland have graciously offered to let me share a podcast recorded just after the U.S. presidential election in November at their annual State of Asia event. "The State of China" features three terrific guests: Wang Qing (王卿), the host of the popular Chinese podcast "The Weirdo" (不合时宜), Zichen Wang of the Center for China and Globalization, and Adam Tooze, one of the truly great public intellectuals of our time. It's all skillfully moderated by the South China Morning Post's Europe edito...
2025-01-3048 min
Peking PlaybookPeking Playbook Unscripted 20250128: DeepSeekHappy Lunar New Year to all those who celebrate it! The top news on the last day of the lunar year of Dragon is obviously DeepSeek, the Chinese AI start-up. You’ve probably heard Americans talk about it all day now. Here is me, Robert Wu of Baiguan and China Translated, and Johnny Zou of MJFinance talking about it after the close of the brutal trading day when Nvidia sheds nearly $600 billion.Listen to Peking Playbook on Apple Podcast, Spotify, YouTube Podcast, and elsewhere! Don’t forget to subscribe!Highlights[2:43] Is Deep...
2025-01-2835 min
China Biz Talk - Baiguan RadioEP21. On DeepSeek - Baiguan X Pekingnology, with Johnny ZouWe have to talk about DeepSeek. Today, I am joined by Zichen Wang of Pekingnology and Johnny Zou of MJFinance’s Newsletter.Johnny is a Hong Kong-based venture capitalist that I interviewed for Episode 14.Highlights[2:43] Is DeepSeek China’s intentional attack against the US?[10:51] How involved is Beijing in all of this?[19:44] Investment strategies in a post-DeepSeek world[24:43] Effectiveness of US export controls[31:02] Censorship on DeepSeek This is a public episode. If you'd like to discuss this with other subscribers or get acce...
2025-01-2835 min
China Biz Talk - Baiguan RadioEP20. Baiguan X Pekingnology podcast collab - pilot episodeThis is a pilot episode as a collaboration between Zichen Wang and Baiguan. Zichen is a former Xinhua journalist and is currently a Research Fellow & Director for Int'l Comms of the Center for China and Globalization. Zichen is also a top-tier China policy analyst and writes prolifically at Pekingnology and The East is Read.The Pekingnology X Baiguan collaboration envisions an unscripted, high-frequency, 20-minute-ish commentary on events of the day in China, combining Zichen’s deep knowledge of China’s policymaking with Baiguan’s data-driven and capital market-oriented takes. We are still testing this out, so we...
2025-01-2427 min
Peking PlaybookPeking Playbook Unscripted 20250124An unscripted experiment where Zichen Wang, with Robert Wu of Baiguan, talk on January 24, 2025, about top China news of the day, covering Xi Jinping’s visit to northeast China, the government’s effort to bring long-term investment into its beleaguered stock market, the emerging controversy in China on the effectiveness of generic drugs via its universal health insurance system, and recent speculation on TikTok - all in around 25 minutes with minimal editing. This is the very first episode. This is a public episode. If you would like to discuss this with other subscribers or get...
2025-01-2427 min
Middle Earth - China’s cultural industry podcast#106 Why RedNote (小红书) ripped success out of the "TikTok refugees" (Live Podcast Rec. on RedNote)Send us a textIt has been a crazy week following the non-TikTok ban and witnessing real 'positive energy' on both sides of the Pacific. How did over 1 million Americans flock to RedNote/小红书? What hurdles lie ahead for us, and are there any potential business opportunities for international content creators?Featuring- Jim Fields: Relay Video - CEO(小红书: Jim技术咨询)- Lucy Lv: Culture news reporter(IG-TT: nihaolucylv)- Wang Zichen: Policy Expert(www.pekingnology.com)To go further- Middle Earth episode #50 Xiaohongshu : Marryi...
2025-01-2259 minAsia In-DepthThe State of China, with Qing Wang, Zichen Wang, and Adam ToozeChina is undergoing unprecedented changes domestically, while also dealing with the challenges of a volatile world. Listen to Qing Wang, host of the popular Chinese-language podcast The Weirdo; Zichen Wang, Research Fellow at the Center for China and Globalization; and Adam Tooze, economic historian and professor at Columbia University discuss the state of China. The discussion is moderated by Finbarr Birmingham, Europe correspondent for the South China Morning Post, and was recorded at STATE OF ASIA, the flagship conference hosted by Asia Society Switzerland last November, in Zurich.
2025-01-1446 minState of AsiaThe State of China, with Qing Wang, Zichen Wang, and Adam ToozeChina is undergoing unprecedented changes domestically, while also dealing with the challenges of a volatile world.Listen to Qing Wang, host of the popular Chinese-language podcast The Weirdo; Zichen Wang, Research Fellow at the Center for China and Globalization; and Adam Tooze, economic historian and professor at Columbia University discuss the state of China.The discussion is moderated by Finbarr Birmingham, Europe correspondent for the South China Morning Post, and was recorded at STATE OF ASIA, the flagship conference hosted by Asia Society Switzerland last November, in Zurich.
2025-01-1446 min
A Thousand WhysChina’s goddess of war and her bronze owlsWhat did owls symbolize 3,000 years ago in China? How were they linked to an ancient woman known as the goddess of war and a female politician of vision at that time? In this episode, we uncover the story of Queen Fu Hao, from the Shang Dynasty, through the bronze owl vessels found in her tomb.Special Thanks:Henan MuseumWang ZichenLiu Yanfeng
2025-01-1011 minOdd LotsRichard Koo and Zichen Wang on What Just Happened in ChinaIn September, Chinese policymakers shocked the markets by unveiling a set of stimulus measures designed to boost the economy and bolster the real estate market. While it's too soon to know whether the announcements will be successful, the stock market took off like a rocket in the wake of the news. Since the initial unveiling, further efforts have been announced with promises of more to come. So how big of a deal is this really? Is this the start of a major turning point in China's economic trajectory? Or is this just business as usual? On this episode, we...
2024-10-171h 18Odd LotsRichard Koo and Zichen Wang on What Just Happened in ChinaIn September, Chinese policymakers shocked the markets by unveiling a set of stimulus measures designed to boost the economy and bolster the real estate market. While it's too soon to know whether the announcements will be successful, the stock market took off like a rocket in the wake of the news. Since the initial unveiling, further efforts have been announced with promises of more to come. So how big of a deal is this really? Is this the start of a major turning point in China's economic trajectory? Or is this just business as usual? On this episode, we speak...
2024-10-171h 18Sinica PodcastSinica Unscripted: Wang Zichen of CCG with a Third Plenum Preview and moreI'm trying something different: totally unscripted and very, very lightly edited recordings grabbed on the go where I happen to be. For the inaugural episode, I've got Wang Zichen, the author of the amazing Pekingnology newsletter on Substack, as well as the man behind the Center for China and Globalization's newsletter "The East is Read." Hear Zichen's origin story, his approach to publishing Pekingnology, the skinny on his new Got China show with Liu Yang and Jiang Jiang, as well as his take on what we can expect from the Third Plenum.See Privacy Policy at https...
2024-07-1256 min
Got ChinaEP1: Where Pottinger and Gallagher got it wrong on China (Part I)In the inaugural episode of Got China, Xinhua Insitute's Liu Yang and Jiang Jiang and CCG's Wang Zichen discuss where Pottinger and Gallagher got it wrong in their Foreign Affairs essay.In part I, the trio ● Observe how the article has shaped the discourse both in Washington and Beijing● Forecast what the article bodes for the future of China-U.S. relationship● Identify which arguments may not be based on sound facts or accurate understandingWhat would you like for the program to cover in the future? Leave your comments below!
2024-07-0325 minGot ChinaEP2: Where Pottinger and Gallagher got it wrong on China (Part II)In the inaugural episode of Got China, Xinhua Insitute's Liu Yang and Jiang Jiang and CCG's Wang Zichen discuss where Pottinger and Gallagher got it wrong in their Foreign Affairs essay.In part II, the trio continues the conversation and talks about● The worrying undercurrent reflected by the article● The differences in Biden and Trump's approach to China ● How other world countries see the China-U.S. competitionWhat would you like for the program to cover in the future? Leave your comments below!
2024-07-0321 min
The Sound of EconomicsKey take-aways from China’s Two SessionsIn this episode of The Sound of Economics, Yuyun Zhan sits down with Alicia García-Herrero and Zichen Wang to talk about China’s Two Sessions, the Chinese government's annual plenary sessions of the National People's Congress (NPC) and of the Chinese People's Political Consultative Conference (CPPCC), held from 4–11 March 2024. They discuss the growth targets which were given during the sessions and what they might mean for the Chinese economy.
This episode is part of the ZhōngHuá Mundus series of The Sound of Economics. ZhōngHuá Mundus is a newsletter by Bruegel, bringing you monthly analysis of China in...
2024-03-2026 min
PRIO's Peace in a PodWhat Does China Want?On November 15th, the world watched as a highly anticipated meeting between Chinese leader Xi Jinping and US President Joe Biden took place amid the scenic backdrop of Woodside, California. From a controversial visit by a US House Speaker to Taiwan, to a Chinese spy balloon crossing the United States – the two countries’ relations have sunk to their lowest point in decades. And from the war in Ukraine to the conflict opposing Israel and the Palestinians, and a general election in Taiwan that could decide the fate not just of the island...
2023-11-2835 min
China Field Notes – with Scott KennedyA New Generation of Chinese Intellectuals: Wang Zichen of the Center for China and GlobalizationJoin Scott for a wide-ranging conversation with Wang Zichen, a rising star amongst Chinese public intellectuals who is now with the Center for China and Globalization, based in Beijing. They discuss Wang’s first trip ever to the United States, the difficulties in U.S.-China relations, how Chinese think tanks try to influence public policy, and much more.Wang Zichen 王子辰 is Research Fellow and Director for International Communications at the Center for China and Globalization (CCG), a leading non-governmental thinktank in Beijing. Before joining CCG in October 2022, he worked, in China and Europe, for 11 yea...
2023-10-2446 min
Our Roots, Our Stories: High Schoolers Share Cultural ExperiencesEpisode 10: Francis' StoryIn this episode, senior Zichen (Francis) Wang shares his experience growing up in a “typical Chinese family.” He feels that the environment in the US is better suited for him as a place to explore his academic interests. Unfortunately, transitioning to life in the US was not a choice he could make himself due to Covid. Although living in the US has been difficult, he has made some of his closest friends here at WRA.
2023-05-0302 min
Inside the China Room with Jiang JiangAlibaba, TikTok, and the changing landscape of Chinese tech in the US – views from Ivy YangThis week on Ginger River Radio, Ivy Yang, who writes an op-ed column called “Discourse Power” or 话语权时代 for FT China and has written for The China Project, shares her insights into the recent restructuring of Alibaba and TikTok's dilemma in the US. Ivy has worked in strategic communications and in corporate comms for more than a decade, including two years at Alibaba Group. She follows topics from TikTok to cross border e-commerce to the effects of the CHIPS Act and much more.Highlights:On TikTok4:56 - How the TikTok hearing has been received in the U.S. an...
2023-04-1449 min
Il cielo sopra PechinoICSP0505 podcastTra i termini che segnano le linee guida strategiche della Cina di oggi, "prosperità condivisa" è probabilmente uno dei più oscuri. Proprio per questo Marco e Gabriele si sono appassionati al tema e ci stanno tornando più volte in questo inizio di quinta stagione. Oggi l'ospite è davvero speciale: Wang Zichen vive a Pechino, è uno dei giovani columnist più promettenti dell'agenzia stampa cinese Xinhua ed è anche il fondatore della newsletter #Pekingnology, che nel giro di pochi mesi è diventata un punto di riferimento per tutti coloro che in Occidente si occupano di Cina.
2021-12-0432 min