Share.

    2 Kommentare

    1. AI is going to be crucial for the future. This means countries need to look at the amount of compute they have. It starts with data centers, filled with GPUs that can train different AI models. There is also storage along with inference.

      Whatever numbers are seen today is small compared to what will be needed in the future.

      AI will be the barometer of how healthy a country is. This is akin to being a manufacturing and financial center over the past 150 years. Both of those were able to drive particular countries.

      The United States, Japan, Great Britain, and China all held the lead in those areas at particular times.

      As always, things are changing. AI and robotics are the future. Of course, as most are aware, this comes with huge energy requirements.

    2. african_cheetah on

      I am very bullish on compute on edge vs large compute on datacenters. May be for training yes but not for everyday autonomous productivity.

      Our brains are compute on edge. 20 watt datacenter weighing 20kg, reliable for 80 years working non-stop.

      First giant mainframes is what current AI datacenters are. Now we have more compute on our wrists than those giant mainframes.

      The landaeur limit for energy per compute is a loooooot lower than what GPUs operate at.

      It will take decades, but physics says it’s possible to build computers smaller, faster and more efficient than brains.

      The ceiling of intelligence is muuuuuuch higher than what big tech tells us.

    Leave A Reply