Artificial intelligence (AI)-powered computers and chips with specialized Neural Processing Units (NPUs) have become increasingly popular. However, Google software developer Osvaldo Doederlein believes that many users do not need such computers.
AI-powered computers are not necessary, according to Google
The reason the Google software developer finds AI-powered computers unnecessary is that these technologies are driven by manufacturers’ desire to bring new products to market rather than genuine user demand.
AI computers may be essential for tech enthusiasts who need to run local AI applications such as large language models or image generators. However, most AI processing is cloud-based and operates on the servers of major companies like Microsoft and Google.
Therefore, for many users, local AI-powered computers are not needed. Doederlein suggests that the release of new-generation chips with NPUs without a real need could lead to disadvantages for regular consumers.
These chips, having fewer cores or cache and focusing on NPUs at the expense of other useful features, could have a negative impact. Additionally, such chips might not meet expectations in terms of security and privacy.
Most AI applications run in the cloud, allowing AI technology to be used even on less powerful devices. Local AI applications could provide more privacy than cloud-based alternatives, but Doederlein highlights the lack of interest in these privacy benefits.
Furthermore, when the disadvantages of local AI applications are combined with hardware-related issues, the need for these technologies may be further reduced. What do you think about this topic? Feel free to share your thoughts in the comments section below.