LOL!! 
9 Likes
Are you sure this is correct? The link doesn’t contain a 2000 but says 2014. And when I use the slider on the right as reference, then it’s 2001.
8 Likes
2014 is the unique id of the message, but it doesn’t reflects its position since the list no longer contains the messages that were deleted.
As for being 2000 or 2001, it was 2000 but now is 2001… I smell a ISD joke somewhere. 
8 Likes
Did he perhaps add a comment somewhere among the other 2000? Are we supposed to find it??
9 Likes
Maybe he undeleted a deleted comment… but, go catch that rabbit! Could be anywhere! 
9 Likes
Uruguay and Portugal play in 30 minutes. Don’t forget to tune in!
9 Likes
Wanted to add a thank you to CCP for the recently added “look outside” option on the station screen in Elgoi. It’s helpful and a logical addition to the game. o7.
10 Likes
Big bucks for silicon compliers.
DARPA Unveils $100M EDA Project Two programs aim to craft ‘silicon compiler’
10 Likes
not bad.
10 Likes
Beddy time for me… nighties lovelies!
9 Likes
I wonder if they could just take AI and use it to build working compilers, just like a game of life builds sertain schematics.
Timezone everyone 
9 Likes
9 Likes
No, you can’t. Hardware is build to the needs of the industry. Compilers are build to make use of these features. They are pretty close to what is theoretically possible, because you have many thinking brains working on them and they know in advance what the hardware can and cannot do.
An AI these days is mainly useful to deal with completely new data, where it would take a person a long time only to find a structure in the data. AIs are good at detecting such patterns and helping us to find sense in it, but they won’t replace the human minds for a while, when we already have a pretty good knowledge of what we’re dealing with.
AIs do get used in hardware however, or rather it’s in the very early stages. So does branch prediction in the latest generation of AMD CPUs make use of neural nets.
Deep learning is the current cutting edge of AI today. I’m not sure what exactly is being done with deep learning, but its not so much about learning of a few things of large data sets, which is what we humans do with our brains, but deep learning is about learning as much as possible about small data sets and constantly adapting to changes. An example of this is to have an algorithm trace the outline of a shape and let it learn about the path as it travels along the outline. Thus the algorithm can learn about shapes in imagines independent of position, size or rotation of the imagine. It’s even possible to allow for some distortion and still be able to detect similar shapes.
One can express this for example by saying a circle is a shape where one always turns left by a small degree and always by the same amount. A triangle is then like a long road, with a sharp turn, another long road, another sharp turn into the same direction as before, and finally long road, which gets one back to the origin.
A deep learning algorithm learns about a very long sentence of events. It’s not how we learn shapes. The human brain learns by looking at the entire image at once and uses countless neurons to process an image as a whole.
Current AIs are still very far away from what a human brain can do.
9 Likes
With time everything may change, we can start work on it just now. Just like there was no Dreamcatcher in Autodest, until someone though its possible and worked on it. Its future, it starts now. 
8 Likes
Did you get flogged?
9 Likes
Not flogged and not fragged, because they stay there. Even after 24 hours. 
8 Likes
All mother floggers will get caught by ISD.
8 Likes
Guess that I am doing it wrong. 
8 Likes
Them eyebrows… in some frames it looks as if he’s wearing sunglasses!
8 Likes
Also, a musical interlude…
Because rocket science is a difficult business after all…
Never give up!
9 Likes