Raw LLM Responses

We IT users which is all of us need to STOP sending unimportant photos, videos etc to Cloud storage in these data centres. They store not just one copy but multiple copies of your photos etc. They also run these data centre for your photos etc 24/7 just in case you want to access them. Do you access your photos taken in a non-important shop that happened to have some recent Xmas decorations at 2am? You do not. This 24/7 multiple copy storage access is the incorrect approach. We should just buy external SSD (storage disc) and save a couple of copies (one copy in each of 2 such SSD) for photos etc at home. Yes do save important documents etc to Cloud storage just in case. But ordinary stuffs, just do the storage at home. This would free up data centres and AI centres for more important functions. Yes, we consumers have ourselves to blame for not diligent enough to be wise about our own memory storage. Second observation. Bulk of these data centres and AI centres use x86 machines. X86 chips are essentially Intel and NVIDIA chip types. x86 chips are NEVER used for smart phones or tablets at all. They generate high heat and hence use more electricity for the same tasks. They are not ideal for phones and tablets. But bulk of data centres and AI centres use these x86 pc and processors etc. Automatically they use more electricity and generate more heat. From this, uses of water and more electricity are inevitably in much higher amount. ARM based chips like those used by Apple products are less complicated so use less electricity and generate less heat for the same tasks vs x86 chips. NVIDIA several years ago wanted to buy ARM, the company that holds the intellectual property rights and the design philosophy of these ARM chips used in all mobile devices. But the British authority knocked back on NVIDIA attempt to take over ARM cutting it would be anti-competitive. That is, NVIDIA, arguably the main hardware supplier for AI these days, were and are aware that it’s x86 based chips are NOT good enough in terms of of energy-water uses for AI tasks longer term. This is irony of highest orders of course. It also says that unless there’s major shift in our own private storage approach for our photos etc, we will continue to pay the costs of higher bills and environmental impacts at much higher rates. With AI, the physical technology (chip) is obviously NOT good enough YET for AI learning etc. NOT good enough in that its energy snd water uses are just not acceptable even tufts let alone next many years. There might be a need for different AI approach that uses less chip power, or less energy and water uses. Sadly from what I have seen, we are stuck with this wasteful AI approach. It’s not driving cars but our stupid pseudo needs to storage rubbish photos etc in the cloud and our own tech world not having the best AI approach that will continue to damage our environment.