SkyShell Studio (@SkyShell_studio) header
SkyShell Studio (@SkyShell_studio) profile picture

We strive to take a fresh perspective on what neural networks can provide and come up with innovative concepts. #NeuralNetworks https://teapawt.net/t.co/EwwAoxowvG

Lukomorye, Green Oak
Joined February 2014
Photo by SkyShell Studio (@SkyShell_studio)Photo by SkyShell Studio (@SkyShell_studio)Photo by SkyShell Studio (@SkyShell_studio)Photo by SkyShell Studio (@SkyShell_studio)Photo by SkyShell Studio (@SkyShell_studio)Photo by SkyShell Studio (@SkyShell_studio)Photo by SkyShell Studio (@SkyShell_studio)Photo by SkyShell Studio (@SkyShell_studio)Photo by SkyShell Studio (@SkyShell_studio)Photo by SkyShell Studio (@SkyShell_studio)
There have been several discussion threads about Reserves and Fees on the XRPL. On behalf of @XRPLLabs I just changed the configuration of our validator, to be enabled the next validator restart. # We will vote for: Transaction fee 馃敿 from 10 drops to 200 drops Account reserve 馃敾 from 10 XRP to 1 XRP Object reserve 馃敾 from 2 XRP to 0.2 XRP What we believe this vote will do is: - Make it slightly (still very cheap) more expensive to transact - Limit barrier to entry / participation # Side effects / trade-offs The number one concern would be increased activity on the ledger to the point where it becomes a challenge for infrastructure. At this point I feel this is a good problem to have. We have dealt with (and lost sleep over) infra challenges before, and we will again. In my opinion we face the choice between an easy going ledger that's hardly being used, and a ledger that gets more use with some challenges. Bring the challenges. True engineers love challenges. If increased use of the ledger tue to lower barrier to entry mean infra providers have to step up their game again (in whatever way: hardware, configuration, horizontal scaling, economics, monetisation, ...) then so be it: a good problem to have. We're doing all of these things anyway. Bring it. Thanks for starting the discussion and backing it up with numbers @Kirjakulov
65
126
34
462
130,855
Replying to @Hmprybackhome
We understood that our technical and financial resources are not sufficient for a full process neural network training, as we wanted. Our plans are the same, but now we are focused on the services and applications, which will the basis of our idea and easier to develop.
2
4
236
Replying to @ShiningMan_JP
Hi! We're in the process of resolving technical difficulties. Unfortunately, there are only 24 hours in a day. There is very little time and there is not enough time for many things.
5
7
13
Replying to @jekoy_777
During this weekend we update the information on our website and publish the #whitepaper . We are also moving forward in the development of our mobile application for data collection. Immediate plans - launch a working API. #XRPLToken #SKStoken #SkyShell
1
6
1
11
SkyShell Studio retweeted
Explaining in Style: Training a GAN to explain a classifier in StyleSpace https://teapawt.net/t.co/c1JJT1gve1 #AI #DeepLearning #DataScience #MachineLearning
1
8
1
11
Replying to @sesetra
Good question. We need blockchain to promote the project. The project itself aims to investigate the neural anomalies that mathematics and deep learning can give.
2
7
18
SkyShell Studio retweeted
The Sensory Neuron as a Transformer https://teapawt.net/t.co/wiZdIZvYvY #AI #DataScience #MachineLearning #DeepLearning
2
7
1
13
Replying to @SchnellvinS
At the initial stage, it is supposed to pay with the $SKS token for collecting statistical data. In the future, there will be a fee for the results of the neural network. In the coming days, we will describe in more detail the tokenomics and our goals.
7
10
54