• 0 Posts
  • 8 Comments
Joined 2 years ago
cake
Cake day: June 9th, 2023

help-circle
  • I know Lemmy hates AI, but this actually would be a perfect use for it. The problem is the idea of what an ad is. Yes, you could try to use secondary characteristics like image color or sound normalized volume (WhyTF do youtube ads still sound 3x louder than content? are we living in cable era again?), but they would be error-prone for any content more visually intense than a podcast. They would also not capture sponsorblock content like “I love showing you all these foreign countries but what I love even more is having my internet connection secure” that match the video flow. A crowdsourced lookup table of all known ad clip fingerprints would go a long way, until ad videos themselves start being AI-generated on the fly for that sweet personalization revenue.

    No, what I really want is to distill the idea of what I want to see into an AI and have it filter out what I don’t want to see for me. I know an ad when I see one, so AI can too. Pre-roll/mid-roll ads? Gone. Sponsorblock content? Gone. Like and subscribe? Skipped as if it didn’t exist. Virtual billboards on the sidelines of sporting events? Overlayed with kittens. Idiocracy banners squeezing the video from either side? Cropped and rescaled. Watermarks? Excised and content-aware-filled.

    The last frontier is when the content itself is secretly an ad, imprinting upon you some idea or point of view. You’ll have to watch out for that one on your own.


  • In the ultimate, you’d need to do something like run a headless browser in a virtual machine, have it play out and record the entire video, then use something like AI to splice out the ad segments and distracting elements (a souped-up sponsorblock will work for a while, but eventually ads will be injected into the raw video stream at random intervals), and present the pristine finished content to you. Basically we are going to re-invent TiVo all over again xD.

    In worst case, you can’t start watching until the pre-roll ad timers expire. This is how adblocking works on Twitch streams currently - you can only see a purple screen even if you block the ads.

    And yes, the headless browser will need to use AI for human-like mouse movement and to solve captchas - basically whatever state-of-the-art technologies spammers and scrapers are already currently using.

    Google is anticipating this future and is trying to implement and force hardware-based DRM for web video before then.



  • It’s a tarpit. If they simply displayed a blocked “no vids for u” message, you’d get outraged, go complain online, look for workarounds, and eventually find a bypass. If everything still works but poorly, you get annoyed, turn off your adblocker to troubleshoot, possibly blame the adblocker for being “buggy” and keep it off. Their help page solution implies they are hoping for just that. There is no “smoking gun” blocked message to go complain online about, even though it is indeed their servers that are degrading your connection on purpose in secret. Or maybe you give up and leave their ecosystem entirely, which is no big loss for them.

    The proper solution is to develop an adblock that they cannot detect is blocking ads. This may require actually downloading the ad video in background, and then lying that the video has played.



  • Then you’d be surprised when you calculate the numbers!

    A Falcon 9 delivers 13100kg to LEO and has 395,700kg propellant in 1st stage and 92,670kg in 2nd stage. Propellant in both is LOX/RP-1. RP-1 is basically long chains of CH2, so together they burn as:

    3 O2 (3x32) + 2 CH2 (2x14) -> 2 CO2 (2x44) + 2 H2O (2x18)
    

    Which is 2*44/(2*44+2*18) = 71% CO2. Meaning each launch makes (395700+92670)*.71 = 347 tons CO2 or 347/13.1 = 26.5 tons of CO2 per ton to orbit. A lot of it is burned in space, but I’m guessing the exhaust gases don’t reach escape velocity so they all end up in the atmosphere anyway.

    As for how much a compute satellite weighs, there is a wider range of possibilities, since they don’t exist yet. This is China launching a test version of one, but it’s not yet an artifact optimized for compute per watt per kilogram that we’d imagine a supercomputer to be.

    I like to imagine something like a gaming PC strapped to a portable solar panel, a true cubesat :). On online shopping I currently see a fancy gaming PC at 12.7kg with 650W, and a 600W solar panel at 12.5kg. Strap them together with duct tape, and it’s 1000/(12.7+12.5)*600 = 24kW of compute power per ton to orbit.

    Something more real life is the ISS support truss. STS-119 delivered and installed S6 truss on the ISS. The 14,088kg payload included solar panels, batteries, and truss superstructure, supplying last 25% of station’s power, or 30kW. Say, double that to strap server-grade hardware and cooling on it. That’s 1000*30/(2*14088) = 1.1kW of compute per ton to orbit. A 500kg 1kW server is overkilling it, but we are being conservative here.

    In my past post I’ve calculated that fossil fuel electricity on Earth makes 296g CO2 per 1 kilowatthour (using gas turbine at 60% efficiency burning 891kJ/mol methane into 1 mol CO2: 1kJ/s * 3600s / 0.6 eff / (891kJ/mol) * 44g/mol = 296g, as is the case where I live).

    The CO2 payback time for a ton of duct taped gamer PC is 1000kg * 26.5kg CO2/kg / ( 24kW * 0.296kg/kW/hour) / (24*365) = 0.43 years. The CO2 payback time for a steel truss monstrosity is `1000kg * 26.5kg/kg / (1.1kW * 0.296kg/kW/hour) / (24*365) = 9.3 years.

    Hey, I was pretty close!