• 1 Post
  • 39 Comments
Joined 3 years ago
cake
Cake day: July 17th, 2023

help-circle








  • Digging more than a few meters for a building is not only expensive, it can be very difficult and dangerous for the buildings a few hundred meters away, depending on the nature of the soil.

    There is a certain cursed railstation project in germany, where the nee railstation is build underground. Though the soil is a specific type, which sucks up any water it gets in contact with and then expands. If there is a significant leakage, we are talking about half a meter difference at ground level for the complete neighbourhood, probably very inconsistent. Building typically don’t like the ground moving that much. So you start investing billions more into the project to make it water tight and still fail to do so.

    And after many years you are still not finished and the project seems to be a coup by the car lobby to discredit travel by rail.


  • Ok, then I misunderstood your arguments. I understood “Away from people” an argument for increasing the distance to wind turbines. They are already normally build 500 to 1500m away from people. I also saw it in my local context. In germany we had the conservatives argue for increasing the mandated distances further, triggered by the far rights hate for renewables, which would make it impossible to build them in large parts of the country.

    When we talk about wind turbines vs datacenters, I think a big argument is the actual direct value for the community. Providing clean energy gives more and direct value to a community, while the AI datacenter are mainly bringing the benefit of slop. Both are done in the capitalism way, but Datacenters (especially AI DCs) are way farther away from the consumer.







  • The company is not paying you for the amount of expertise. It is paying for how much money you will make them. If you limit your work to only what you can do yourself (without leading others), then you have deminishing returns. Only because I can now solve very complex problems because of my expertise, I cannot do 10x the work. Time is limited, expertise won’t bring you more of it. Leading will help other do their work better and faster, so you will be more valuable with that.


  • The higher you get up in a company, the more it will be about running the company instead of what the company does in detail. While we all have our gripes with middle and upper management, such a structure will come naturally with a growing company. Really small companies often have owners and management, who are themselves experts in the field. For bigger companies that is not really achievable and not even wanted, since management has a lot to manage.

    When you become an expert and want to climb the latter, then some management will automatically come to you. You will be asked to lead colleagues more junior than you. You will be asked to manage strategy for your field of expertise. You will be asked to assess the effort needed to handle projects and the risk assiciated with them. This already means quite some management work. The reward is, that (if you do a good job) those under you will be able to do a better job and that they have time to themselves become experts by doing the technical work.

    Thats my current situation. In my IT job I have to do all of the above to guide the project into success while giving some of the technical work to those more junior. For some this is good. Though I personally probably won’t go much further into management positions, because I don’t like that work enough.


  • I support this change. Hardware can also be relevant apart from thr typical servers/mini PCs. Recently I asked about possibilities for a self hosted video/audio doorbell and was met with really helpful answers. While this question could also have fitted into communities for smart homes, I still think it is a good fit here. Self hosting already is a rather broad term. We should not limit ourselves too much in its definition.



  • Most buzz about quantum computers is how they might be able to break traditional encryption algorithms by fast defactorisation of very big numbers. Though they still owe us proof, that this actually works.

    There is this paper, which compares the “big” achievements in quantum defactorisation with a (not really) trained dog. Basically every of these achievements cheated with the prior knowledge of the factors or have chosen convenient numbers, while still being worlds away from common key sizes (like 2048 or 4096 bit).

    Real usage for quantum cpmputers will probably still take quite a while to manifest.