Microsoft has taken out deals to lease its data centers for additional training for Openi’s language model Chattgot. This news may seem surprising to the popularity of the model, but for many good reasons the AI technology field is a controversial. Microsoft, in addition to high-run cost, relatively low-lower returns, and growing competitiveness- in addition to its own sick AI-Med earthquake 2 demo, has proven enough reason for Microsoft for Microsoft that they can be out of two Gigwat worth of two Gigwat worth of two Gigwat worth of the entire United States and Europe.
Sometimes AI can refer to a smart machine learning that helps researchers find plastic food enzymes, or teaches robot dogs how to walk using their nervous system. However, AI often refers to the strength of language models and image generators, is usually trained on the stolen data on trained data so that they can provide the results. Chat GPT falls into the latter of these categories, and accordingly ReutersA large part of the deduction on Microsoft lease is due to its decision, especially its creator, do not support the burden of additional training work for the open.
Models like Chat GPT require major data centers, which are commonly leased to train and work through tech giants such as Microsoft. The use of these data centers is not cheaper with both financial and electricity, especially with the current boom. In addition, companies can lease these centers to other use, which can be more paid. While a large language model like Chat GPT Can When used properly, it is very useful, and is not much dependent, there are more hidden costs that the last consumers do not see.
Along with Microsoft, its investors have also witnessed a relatively slow payment with the rise of rival models like China’s Deep Sak. Compared, DiPsic is a very high cost -efficient AI, which is gaining more kindness, including the attention of a former Intel CEO’s Fath Tech company. So Microsoft is not surprised in the great scheme of things behind leasing its data centers.
Facebook’s meta in the United States has accelerated leasing its data centers to fill the gaps found by Microsoft, and Google’s parent company Alfbit has done the same in Europe. Reuters also reported that the alphabet pledged $ 75 billion on its AI bloodout this year, which is 29 % on Wall Street expectations. Meta, ready to spend up to $ 65 billion on AI. All three companies, including Microsoft, have defended AI’s expenses to maintain competitiveness with Deep Sak’s choice.
Despite withdrawing from hosting more data center training for Chat GPT, Microsoft says AI is still on track to meet its $ 80 billion investment in AI infrastructure for the year. The company says that although it “can speed up or adjust our infrastructure in some areas with a strategy, we will continue to develop firmly in all regions”.
