)'-.,_)‘-.,)'-.,_)'-.,)'-.,_)’-.,_

  • 0 Posts
  • 14 Comments
Joined 9 months ago
cake
Cake day: May 19th, 2024

help-circle
  • Here’s the idea: Deprive YT of their revenue and maximize their expenses.

    Browse the site as usual, but never watch any videos. Instead, download them at maximum resolution. You’ll skip all the adds, and take the maximum amount of bandwidth.

    The business stays profitable as long as people watch ads and don’t use bandwidth anywhere near as much as you could. Download a bunch of very long videos just to check if they’re worth watching. Ignore anything shorter than 10 minutes.

    Make a channel and upload a bunch of long videos. A 4K dash cam should be a pretty good source of data. Make sure your videos are so boring that even bots don’t like them. These videos are essentially just dead weight in the servers.

    You could also try reporting random videos and comments here and there. I’m pretty sure no human ever sees those inputs, but some machine learning model might. Messy data isn’t going to break the system, but you could make that speed bump a little higher.


  • That would be possible, but seasonal production has some serious drawbacks.

    Let’s say you have a steel mill with several production lines, solar powered arc furnaces, and enough batteries to keep production running through the night. During the summer you can continue production 24/7, but in the winter you’ll have to shut down completely, because there’s not enough energy to keep even a single production line running. This means that there will be wild fluctuations in a variety of things:

    • number of employees on site
    • rate of steel produced
    • demand for storage space for raw materials and steel products
    • demand for logistics
    • demand for maintenance

    This means, that in order to deal with the fluctuations, you would need to have lots of spare capacity in pretty much everything: More machines, more people, more money. If you could keep the production steady throughout the year, you could do so with less. Also, what will the employees do during the winter? The skiing resorts can’t possibly employ all of them.

    In the winter you’ll have plenty of time to fix anything that’s broken, but if there’s an unscheduled shutdown during the summer, you’re suddenly going to need lots of maintenance personnel and materials. Incidentally, those would be in short supply in the summer, because all the other factories would have the same problem. You would need to have lots of spare capacity in maintenance as well.

    The AI industry should be fine, since you could train models when energy is cheap. Oh, but what if the summer isn’t long enough for you to update all your models? Simply just buy more computers so you have more spare capa… Oh, it’s the steel mill problem all over again. Oh, but what about the people who use the models during the winter? Maybe you could charge your customers double the price during the winter so that the traffic would be reduced to a reasonable level. Fortunately though, wind power and other renewables could help with the winters, but having more grid energy storage would make things run smoother.











  • So far, so good. The software works. You’re able to watch videos, comment, follow people etc. There are minor glitches here and there, but there are also lots of updates all the time.

    It reminds me of something I’ve read about Mastodon. That place was designed to be less addictive than Twitter, so feeling a little bored after a while is a feature, not a bug. Loops gives me that same vibe, although it’s probably caused by the the small number of videos and people making them.


  • The marketing department seems to have highly inflated expectations, so I think this is mostly a marketing thing. They are painting a picture of something that doesn’t exist yet. The same thing happened with Siri and Google Assistant when they were released. People expected these tools to be capable of doing the kinds of things that modern LLMs are now beginning to approach, so the amount of disappointment was pretty brutal.

    But is it a solution looking for a problem? If you need to generate pictures for a horror themed book, image AIs can get the job done properly. You don’t even need to request mutated abominations, because the AI gravitates towards those anyway. If you need to write a pointless self help book that meets the page count without conveying any actual information, LLMs are the way to go.

    These are all pretty niche examples, but people should understand that all of the AIs we have today are narrow AIs. The scope is about as narrow as the tip of a torx screwdriver. Try it on a different screw and you’ll see what I mean. Use it as a hammer, and you’ll gain full understanding of the nature of this problem.

    If you have exceedingly low expectations, LLMs can actually do some trivial tasks very well. If you have hardly any experience in writing or programming, an LLM will prove to be an impressive tool. If you are a professional though, you’ll notice pretty quickly that the LLM is only good for the simplest tasks, while you still need to do everything that is even a little bit demanding.