消息称MacBook Neo 2拟搭载A19 Pro芯片和12GB内存
Related
The South Korean government is considering extending the sovereign AI core GPU guarantee project, with a budget of up to 4 trillion won next year.
According to Odaily Odaily, the South Korean government is considering extending its support project for GPUs, a core resource for AI. On July 7th, industry sources indicated that the Ministry of Science and ICT is investigating the capacity of major South Korean cloud service providers to build additional GPUs, aiming to extend the "AI Computing Resource Utilization Infrastructure Enhancement Project." The core of this project involves government-funded purchases of GPUs, deployment in cloud service provider data centers, and support for industry-academia-research collaboration. Last year, the South Korean government allocated a budget of 1 trillion won, and this year it allocated 2 trillion won, securing over 20,000 GPUs. The project was originally scheduled to end this year, but with the surge in demand for advanced resources such as NVIDIA's next-generation GPU Vera Rubin, the South Korean government is developing comprehensive countermeasures, including data center efficiency optimization and new space security plans. It is reportedly already applying to the Planning and Budget Office for a "limited additional budget" for GPU security. Industry estimates suggest that, considering this year's GPU security budget and the government's intentions, next year's budget could reach a maximum of approximately 4 trillion won. Industry experts say that compared to the GPU supply itself, "equipment installation space" will be the key to the success or failure of the project, because running thousands of high-performance GPUs simultaneously requires a large power supply and advanced cooling systems, and the equipment installation space held by South Korean companies is nearing saturation. (ETNews Electronics)
Meituan open-sources LongCat-2.0, simultaneously releasing inference code for domestically produced SIM cards.
Mars Finance reported on July 6th that Meituan officially open-sourced its trillion-parameter large-scale computing model, LongCat-2.0. LongCat-2.0 has a total of 1.6T parameters and an average activation of approximately 48B. It features deep collaborative optimization in model architecture, chip adaptation, and deployment strategies, specifically targeting domestically produced computing chips with limited GPU memory and bandwidth. (Wide Angle Observation)
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.
Polymarket launches Bitcoin Lightning Network instant deposits and integrates the Spark protocol.
BlockBeats reported on July 8th that prediction market platform Polymarket announced support for instant, self-custodied deposits via the Bitcoin Lightning Network, powered by the Spark protocol. Compared to the previous on-chain deposit method, which required 3 to 6 block confirmations and took approximately 10 to 60 minutes, the new solution achieves near-instantaneous deposits, lowering the deposit threshold and transaction costs. According to reports, Spark performs checks for double-spending risk, fees, and Replace-by-Fee (RBF) during transaction broadcasting, achieving "zero-confirmation" deposits. It also supports on-chain, Lightning Network, and stablecoin payment channels, eliminating the need for the platform to operate its own Lightning Network nodes. Polymarket stated that this move will further improve the efficiency of Bitcoin users' fund utilization and enhance its competitiveness against rival Kalshi.
The positive earnings forecast in the interim report signals a profit inflection point, and the humanoid robot sector is ushering in a new era of industry synergy.
On July 7th, the A-share humanoid robot sector experienced a correction after a rapid rise, with the Wind Humanoid Robot Concept Index falling 2.93% that day. Currently, the industry fundamentals are undergoing profound changes: multiple companies in the industrial chain have released preliminary earnings announcements for the first half of the year, showing improved profitability trends from core components to complete machine integration; simultaneously, leading global companies are accelerating mass production, with capacity construction and order verification entering a critical window period. Analysts believe that the humanoid robot industry is gradually shifting from the initial concept-driven phase to a new stage of capacity implementation and performance realization. Coupled with multiple factors such as the IPOs of leading companies and the accelerated commercialization of leading domestic and international manufacturers, the medium- to long-term investment logic of the sector is receiving strong support from fundamentals. The short-term correction may be a normal digestion of the previous gains by the market. With the profit inflection point approaching in each link of the industrial chain, the humanoid robot sector is expected to usher in a new development stage of resonance both domestically and internationally. (China Securities Journal)
TAC, a project already listed on Binance Alpha and its contracts, plummeted 80% in one hour.
PANews reported on July 8th that, according to Binance Alpha data, TAC Protocol (TAC) plummeted by approximately 80% in the past hour, currently trading at $0.006. TAC will begin trading on Binance Alpha on July 15, 2025, simultaneously launching the TACUSDT perpetual contract. TAC is a blockchain connecting EVM decentralized applications with the TON and Telegram ecosystems, aiming to allow Ethereum applications to seamlessly integrate with the TON network. As of now, the official explanation for this significant price fluctuation has not been provided.