• 6 Posts
  • 85 Comments
Joined 3 years ago
cake
Cake day: June 23rd, 2023

help-circle

  • My employer uses Akamai as a CDN (and security) provider for their websites. Akamai has an optional tool called Image and Video Manager that, in a nutshell, optimizes images for better site performance.

    Basically you put a high quality image on your website, and when Akamai retrieves it the first time it optimizes its color palette, compression, size, and format for multiple devices. Each derivative image is cached by Akamai for 30 days at a minimum.

    So while you may put a 1000x1000 JPG with a quality of 95%, an Android user may be served a 200x200 image with 80% quality, and it may be a WEBP or PNG format instead of JPG. Exactly which derivative image is served is based on criteria like the web browser being used, the viewport size, etc.
















  • This is one of the many edge cases that I’ve been convinced will keep self driving cars from becoming mainstream unless/until true AGI is achieved.

    A few years ago I stopped at a red light next to a construction site. I was watching the traffic light, so at first I didn’t notice a cop at the construction site trying to wave me through the red light. He finally took a few steps towards me and yelled to get my attention. Only then did I realize he was waving me through, so I did just that. I seriously doubt any current self driving car would recognize a police officer (and not just a random pedestrian) that’s overriding the traffic signal like that.

    Another edge case, coincidentally at the same intersection a few years earlier was when there was a car fully engulfed in flames as I drove up. I could hear sirens in the distance, and the cars in every direction were making sure to safely get out of the way of the approaching fire trucks. At least one or two cars cautiously crossed on the red to get out of the way. Again, I doubt any current self driving car would have navigated that situation anywhere nearly as well as a human.



  • This is one of the many edge cases that I’ve been convinced will keep self driving cars from becoming mainstream unless/until true AGI is achieved.

    A few years ago I stopped at a red light next to a construction site. I was watching the traffic light, so at first I didn’t notice a cop at the construction site trying to wave me through the red light. He finally took a few steps towards me and yelled to get my attention. Only then did I realize he was waving me through, so I did just that. I seriously doubt any current self driving car would recognize a police officer (and not just a random pedestrian) that’s overriding the traffic signal like that.

    Another edge case, coincidentally at the same intersection a few years earlier was when there was a car fully engulfed in flames as I drove up. I could hear sirens in the distance, and the cars in every direction were making sure to safely get out of the way of the approaching fire trucks. At least one or two cars cautiously crossed on the red to get out of the way. Again, I doubt any current self driving car would have navigated that situation anywhere nearly as well as a human.


  • Not exactly 911, but somewhat similar. A few years ago my wife & I were in a rental SUV while on vacation. It was a fairly new car with only something like 2000 miles on it. We were in the third lane of a 4 lane highway when a drunk driver hit us from behind with almost no warning. It caused our car to spin 360 degrees across 3 lanes before coming to a stop in the breakdown lane.

    Within about 5 seconds of the car coming to a stop we heard a voice asking if we’d been in an accident and were we ok. It turns out the rental car had one of those OnStar types of services. We were so pumped full of adrenaline that it was all just a blur as we tried to remember what highway we were on, near what exit, etc. We were so panicked… Luckily a state trooper on a routine patrol stopped maybe a minute later so we didn’t have to keep trying to figure out how to tell the OnStar person where we were.