These Researchers Just Shrunk an AI Model and Somehow Made It Smarter
Related
my country has made a breakthrough in physical world modeling.
According to Mars Finance, Fysics AI has recently achieved a breakthrough in its next-generation physical world model, Fysiverse, built upon the Fysics differentiable physics engine. This breakthrough has pioneered a new technical approach to world models, effectively addressing common problems faced by current data-driven world models, such as "physical illusion, inference failure, and collapse of non-standard scenes." Fysiverse is an innovative next-generation physical world model that adheres to the laws of real-world physics. The disruptive innovation of Fysiverse lies in its proposed system path of "explicit physics simulator × generative renderer": using computable physical states as the intermediate layer, a differentiable physics engine as the world evolution mechanism, and a renderer as the visual expression layer. It doesn't simply add motion trajectories to the world model; instead, it reconstructs the computational chain of the world model into "observation—state—physical evolution—visual presentation." (Science and Technology Daily)
Anthropic: The Claude model contains elements similar to "conscious human thought."
BlockBeats reported on July 7th that Anthropic released a new research report discovering a spontaneously generated internal "global working space" called J-space (Jacobian space) within the Claude model. This is a dedicated set of neural activation patterns for the model to engage in silent thinking, allowing it to process concepts without writing them down, similar to conscious thought that humans can report. The research used Jacobian technology to identify neural activation patterns in the J-space, enabling it to read unspoken concepts and modify these patterns. Experiments showed that disabling the J-space weakens multi-step reasoning abilities but does not affect basic tasks or factual recall. Click the original link below to join the Beating · Lark AI news channel for 24/7 monitoring of global AI hotspots and news.
Myanmar's AI-driven telecom fraud industry exposed: Starlink becomes key infrastructure, encrypted payments and OpenAI/Google models incorporated into toolchains.
A leaked investigative report from a Myanmar scam Odaily park reveals that global telecom fraud is rapidly evolving towards an "AI industrialization + cross-border encrypted payment" system. These fraud networks use cryptocurrencies to transfer funds and employ automated tools based on large models for multilingual script generation, identity spoofing, and emotional manipulation. The investigation shows that these systems heavily utilize OpenAI's ChatGPT and Google's Gemini to support "large-scale social media fraud," while funds are rapidly laundered and transferred through on-chain payments and cross-border channels, forming a two-tiered structure of "AI customer acquisition + encrypted settlement," enabling the fraud industry to achieve high automation and transnational expansion capabilities. Furthermore, Elon Musk's Starlink has become the leading network service provider in the Myanmar scam industrial park, with US ISPs handling nearly one-fifth of the park's traffic. In response to the allegations, OpenAI stated that fraudsters using ChatGPT behave in a manner highly similar to ordinary users, making identification difficult. However, they have been using behavioral pattern recognition and risk control systems to ban approximately 100,000 suspicious accounts monthly. Google stated that its AI models have security safeguards in place and emphasized its commitment to "responsible AI development" to limit the use of tools for fraudulent and other illegal purposes. (Red Star News)
Meituan open-sourced its trillion-parameter large-scale model LongCat-2.0, and simultaneously released the inference code for domestically developed Chinese card processors.
According to Beating's monitoring, Meituan has officially open-sourced its trillion-parameter large-scale model, LongCat-2.0, with a total of 1.6T parameters and an average activation of approximately 48B, designed specifically for real-world agentic coding tasks. Architecturally, it innovatively introduces LongCat sparse attention and N-gram embedding. The former reduces fragmented memory access through flow-aware indexing and hierarchical indexing, accelerating training and inference with millions of contexts; the latter, while achieving nearly 97% sparsity in MoE, invests 135B parameters into the embedding layer, balancing parameter gains and structural stability. Post-training employs multi-teacher online distillation, categorizing experts into Agent, Inference, and Interaction types, seamlessly integrating them on a domestic computing power cluster through the MOPD architecture. As the industry's first trillion-parameter model to complete inference on a 50,000-card domestic computing power cluster, LongCat-2.0 validates the mature capability of domestic chips to handle complex large-scale model tasks. To address the multiple limitations of domestically produced Chinese chips in terms of memory, bandwidth, and interconnects, Meituan has made breakthroughs in three areas: model, chip adaptation, and deployment. At the model level, ScMoE leverages the core control capabilities of domestically produced chips to achieve physical core-level parallelism for Dense and MoE branches, combined with KV-cache partitioning to alleviate the pressure on ultra-long context memory. At the chip adaptation level, Super Kernel reduces operator startup overhead, and Weight Prefetch hides I/O latency, maximizing hardware utilization under constrained conditions. At the deployment level, PD separation is adopted to balance TTFT and TPOT, along with asynchronous Expert-Parallel load balancing to solve load unevenness under high EP (efficiency level). This open-source release simultaneously provides multiple precision versions, including BF16, FP8, and INT8, and fully opens up inference results optimized for domestic computing power, aiming to enable existing domestically produced cards and even older cards to smoothly deploy trillion-model inference services. --------------------------------- Click the original link below to join the Beating · Lark AI news channel and monitor global AI hot topics and news 24/7.
Myanmar's telecom fraud AI industrialization exposed: Starlink becomes key infrastructure, encrypted payments and OpenAI/Google models are incorporated into the toolchain.
According to a report by Red Star News, citing Mars Finance, an investigative report leaked from a scam industrial park in Myanmar reveals that global telecom fraud is rapidly evolving towards an "AI industrialization + cross-border encrypted payment" system. These fraud networks use cryptocurrencies to transfer funds and employ automated tools based on large models for multilingual script generation, identity spoofing, and emotional manipulation. The investigation shows that these systems heavily utilize OpenAI's ChatGPT and Google's Gemini to support "large-scale social fraud," while funds are rapidly laundered and transferred through on-chain payments and cross-border channels, forming a two-tiered structure of "AI customer acquisition + encrypted settlement," enabling the fraud industry to achieve a high degree of automation and transnational expansion. Furthermore, Elon Musk's Starlink has become the leading network service provider in the Myanmar scam industrial park, with US ISPs handling nearly one-fifth of the park's traffic. In response to the allegations, OpenAI stated that the behavior of fraudsters using ChatGPT is highly similar to that of ordinary users, making identification difficult, but they have already blocked approximately 100,000 suspicious accounts monthly through behavioral pattern recognition and risk control systems. Google stated that its AI models have safety barriers in place and emphasized its commitment to "responsible AI development" to limit the tools from being used for illegal purposes such as fraud.
The US plans to release industry standards for AI models, which could restrict the businesses of companies like OpenAI and Anthropic.
According to a report by the Financial Times, the US government is in talks with several AI companies to release voluntary industry standards for cutting-edge AI models as early as next week, aiming to prevent the misuse of advanced technologies by other countries. The standards will clearly define performance benchmarks, release dates, and domestic and international access permissions for these models. Due to recent tightening regulations, several leading AI companies have already adjusted their operations. OpenAI has postponed the full release of GPT-5.6 at the government's request, only making it available to a limited number of qualified partners; Anthropic's two top models were just released from export restrictions this week after nearly three weeks of restrictions; and Google is also in close communication with the government while preparing its next-generation code models. The report also mentions that both OpenAI and Anthropic, currently under regulatory scrutiny, are actively preparing for initial public offerings (IPOs).