The development of artificial

[shamimhasan07]

I think most industry participants were surprised to see Snaptrude at Autodesk University 2023. Was this due to Autodesk's serious interest in the young company, or was it a reflection of their new openness? In the past, trusted third-party developers have been kept out of AU because they duplicated the functionality of one of Autodesk's tools.

With efforts to automate the production of entire drawing sets or develop detailed BIM models using only AI, it is clear that BIM tools will change significantly in the coming years. We all need to become more data-centric, rather than application-centric. The structure of the technology ecosystems we build and the choice of tools and services we buy are currently driven by applications and data formats. If we want to achieve true openness and flexibility of data, we must break free from these ingrained limitations. With cloud-based phone number indonesia SaaS applications, API integration, open data, and more application choices, we have the opportunity to create a best-in-class technology stack.

Hardware
intelligence will also affect our computers. Processors in personal workstations and cloud servers are beginning to be equipped with special silicon neurons for AI work. And companies like AMD, Nvidia, and Intel are already releasing graphics processors with cores optimized for working with artificial intelligence.

There will also be a new generation of accelerated processing units (APUs), which will combine a CPU, GPU, and neural processing unit (NPU) on a single silicon chip. They will have incredible power and broad capabilities for offloading data streams to different dedicated cores, while consuming less power than previous generations. AMD and Intel currently have the technology to create APUs, and Nvidia has teamed up with ARM in this endeavor. So far, the company only supplies integrated CPU/GPU chips for data centers, but there are rumors of a chip for desktop computers that will be able to run Windows.

The big question is: where will most of the data processing take place? Since software and data will likely be in the cloud, will companies have to pay for hardware rental and constant connectivity? Renting GPUs in the cloud is expensive. So using the cloud is not always the right solution. However, the main processing for AI should take place in the same place where the data is stored.