Donât often click on China Uncensored, given its parent, but fur some reason clicked on it.
Very interesting inside stuff from the advisor to Pompeo when he was Sec. of State under Trump.
They changed a lot of state department thinking toward China, that any Taiwanese or supporters here would like. Even changed state departmentâs handling of Taiwan in a positive way for Taiwan. Anyway, if youâve got time to spend, try it out. May even change some peopleâs mind on how far Trump did go in regards to fighting China, and not just on trade war.
More should be done against the bullshit interpretation of that resolution and the imaginary Nine-dash line.
and ofc those fuckers of 5 star and most of the left (italians) voted AGAINST, not just abstained, AGAINST.
What a bunch of paid grifters.
Ah yes the time-honoured phrase in journalism: âin a move sure to anger China.â
In 2024, Taiwan exists . . . âin a move sure to anger China.â
In 2024, many people still enjoy drinking bubble tea . . . âin a move sure to anger China.â
Expressed visually, just for @Mr_PBH because he has told us how much he loves such visuals:
Guy
Feeling needy, are you? The echo chamber boring you?
Sorry I missed that. I was out getting some bubble tea.
Guy
Some insights from former Australian PM Kevin Rudd on what the PRC is up toâand what many nations including Australia and Canada can continue to expect:
An except from this paywalled article: Why Xi Jinpingâs plans for Australia arenât working - according to Kevin Rudd
Guy
TSMC to suspend production of advanced AI chips for China from Monday, FT reports
TSMC has told Chinese customers that it will no longer manufacture AI chips at advanced process nodes of seven nanometres or smaller, the report said.
So, whatâs that chips, Precious?
From chaptgpt:
1. AI and Machine Learning (ML) Accelerators
- Training Deep Learning Models: AI chips at these process nodes are highly optimized for training large deep neural networks (DNNs). They excel in parallel processing tasks, which is essential for handling the massive datasets and complex models used in AI training. Chips like NVIDIAâs A100 and AMDâs MI200 are built on 7nm or smaller processes and are widely used in data centers for AI workloads.
- Inference Acceleration: Once models are trained, they must be deployed for inference, which requires quick decision-making in real-time. AI chips at advanced process nodes provide the speed and low latency needed for tasks like image recognition, natural language processing (NLP), and autonomous vehicle navigation.
2. Graphics Processing Units (GPUs)
- Graphics and Gaming: While GPUs are traditionally used for graphics rendering, modern GPUs (like those from NVIDIA or AMD) built on 7nm or smaller processes are also optimized for AI and machine learning tasks. For example, the NVIDIA RTX 30-series GPUs and AMD RDNA 2 architecture are fabricated on advanced process nodes and are used for both gaming and AI-related tasks like real-time ray tracing and ML model inference.
3. Specialized AI Processors (TPUs, FPGAs)
- Googleâs Tensor Processing Units (TPUs): TPUs are custom AI accelerators that Google developed for its data centers and cloud services. Built on cutting-edge process nodes, these chips are optimized for tensor computations, which are central to many AI tasks like training deep learning models.
- Field-Programmable Gate Arrays (FPGAs): FPGAs, such as those produced by Xilinx (now part of AMD) and Intelâs Altera, are also built using advanced process nodes and are used in AI-specific applications like image recognition, data analytics, and network acceleration. These chips can be reprogrammed for specific AI algorithms, offering flexibility and efficiency.
4. Edge AI Chips
- Mobile and IoT Devices: Advanced AI chips are increasingly used in edge devices where AI workloads need to be performed locally rather than in the cloud. Chips like Appleâs A-series and M-series chips, Qualcommâs Snapdragon processors, and MediaTekâs Dimensity series are built on advanced nodes and feature integrated AI acceleration to handle tasks like facial recognition, natural language processing, and real-time data analysis on smartphones, wearables, and other IoT devices.
- Autonomous Vehicles: Advanced AI chips are critical in autonomous driving systems, where large amounts of sensor data (e.g., from LIDAR, cameras, and radar) need to be processed in real time for decision-making. Companies like Tesla, NVIDIA, and Intel (through its Mobileye division) produce AI chips based on advanced nodes that power autonomous driving and advanced driver-assistance systems (ADAS).
5. Data Centers and Cloud Computing
- Cloud-based AI Services: Advanced AI chips are also deployed in data centers for cloud-based services, such as Amazon Web Services (AWS), Google Cloud, and Microsoft Azure, where they power AI, machine learning, and data processing workloads for a variety of industries. These chips are often optimized for parallel processing, with thousands of cores working simultaneously to handle large-scale AI model training and real-time inferencing.
6. High-Performance Computing (HPC)
- Scientific Simulations: AI chips at advanced nodes are used in scientific research, where AI is applied to simulations, data analysis, and predictive modeling. These chips enable faster simulations for fields like climate modeling, genomics, and physics.
- Supercomputers: Supercomputing systems like the Fugaku in Japan and Frontier in the U.S. use AI-accelerated chips built on smaller process nodes (often 7nm or 5nm) to achieve unprecedented computational power. These supercomputers leverage AI for complex simulations and data analysis, including drug discovery and climate forecasting.
7. Artificial General Intelligence (AGI) Research
- Neuromorphic Computing: AI chips at smaller nodes are also being explored for neuromorphic computing, which attempts to mimic the brainâs structure and function. These chips could potentially pave the way for Artificial General Intelligence (AGI) by providing the necessary processing power for large-scale, brain-inspired models.
8. Cryptography and Security
- AI-Powered Security Solutions: AI chips are also used in security and cryptography, where advanced AI algorithms are applied to identify patterns of fraud, detect threats, and enhance encryption techniques. These chips benefit from the high processing capabilities offered by advanced nodes, ensuring low latency and real-time threat detection.
So, where are they made?
1. Main Manufacturing Facilities in Taiwan
TSMCâs primary manufacturing facilities for chips at 7nm and smaller process nodes are located in Taiwan, where the company has several fabs dedicated to cutting-edge production technologies:
- Fab 14 (also known as TSMC Fab 14B): Located in Hsinchu, Taiwan, Fab 14 is one of the key facilities where TSMC produces chips using its 7nm and 5nm process nodes. It is part of TSMCâs Advanced Technology Development hub.
- Fab 15: Also located in Hsinchu, Fab 15 is another critical facility for the production of 5nm and 7nm chips. This facility includes several production lines, including those dedicated to the high-volume manufacturing of leading-edge semiconductors.
- Fab 18: Located in Taichung, Taiwan, Fab 18 is TSMCâs cutting-edge fab designed to handle the most advanced process nodes, including 5nm and 3nm. Fab 18 is crucial for the production of the 5nm chips used in many modern processors, such as those used in Appleâs A-series and M-series chips. As TSMC moves toward 3nm, Fab 18 is expected to be key in this advanced manufacturing.
- Other fabs in Taiwan: TSMC has a number of other fabs spread across Taiwan, including Fab 6, Fab 8, and others that support advanced nodes and various process technologies for the production of 7nm, 5nm, and 3nm chips. TSMC has been constantly upgrading its Taiwanese fabs to keep up with the demands for smaller process nodes.
2. Other Locations Outside Taiwan
While the majority of TSMCâs advanced chips (including 7nm, 5nm, and 3nm) are produced in Taiwan, the company has been expanding its global footprint, and has begun constructing new fabs outside Taiwan:
- United States (Arizona): TSMC is building a 5nm-capable semiconductor fab in Phoenix, Arizona, with production slated to begin in 2024. This fab will help TSMC meet the growing demand for advanced chips in the U.S. and will produce 5nm chips for customers, including major U.S.-based companies like Apple, Qualcomm, and AMD.
- Japan: TSMC has announced plans to build a 3nm-capable fab in Japan, as part of a partnership with Sony. This facility, which will be a smaller-scale operation compared to TSMCâs Taiwanese fabs, will primarily focus on specialized semiconductor production for industries such as automotive and other localized applications.
- Other Regional Partnerships: TSMC is also in discussions with other countries and regions about the potential for establishing more advanced production facilities. This is part of a broader strategy to reduce geopolitical risks and increase its global presence.