<div dir="ltr">Thank you Russel. That was a better answer than I have been able to find in any of the documentation. <br>Is a copy left activist the opposite of a copyright lawyer?<br><div><br></div><div><br clear="all"><div><div dir="ltr" data-smartmail="gmail_signature"><div dir="ltr">Cody Smith</div></div></div><br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Sun, Apr 10, 2022 at 7:14 PM Russell Standish <<a href="mailto:lists@hpcoders.com.au" target="_blank">lists@hpcoders.com.au</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On Fri, Apr 08, 2022 at 08:19:16AM -0600, cody dooderson wrote:<br>
> That soundscape thing is nice. What a nifty idea. <br>
> Speaking of AI, has anyone used GitHub copilot. It is AI for writing code. It<br>
> is spectacular. It writes decent code with very few prompts. It makes a few<br>
> mistakes but don't we all. <br>
> I haven't been able to find a good writeup on how it works. Does anyone know<br>
> about it? Does it run locally? Could it create itself? <br>
<br>
Yes - I've been reading some of the articles, and listened to some<br>
talks presented at NVidia's GTC conference.<br>
<br>
Copilot is a model based on GPT3, which is a general language parsing<br>
and generation model. Basically, it takes input tokens, and outputs<br>
most like response tokens based on that. It is a recurrent neural<br>
network with a few 100 billion parameters (ie essentially synaptic<br>
weights), with a complexity approaching that of a human brain (human<br>
brains contain around a trillion synapses). GPT3 has been pretrained<br>
on corpora sourced from the world-wide web, and interestingly it has<br>
been found that the pre-trained model can be fairly rapidly retrained<br>
on different tasks. In the case of Copilot, it has been trained on the<br>
contents of Github, which is the world largest open source<br>
repository. This has caused concern amongst copy left activists, as<br>
quite a bit of GPL licensed code was involved, and this is arguably<br>
not a case of "fair use" of copyright code.<br>
<br>
No - it does not run locally. It requires a massive, massive cluster<br>
of NVidia GPUs to run training, and the inference part requires a not<br>
quite so massive cluster of GPUs as well. Hence the need for it to run<br>
in the cloud.<br>
<br>
> <br>
> On Wed, Apr 6, 2022, 1:24 PM Marcus Daniels <<a href="mailto:marcus@snoutfarm.com" target="_blank">marcus@snoutfarm.com</a>> wrote:<br>
> <br>
> <br>
> Change the sentence to be “Put Ukrainian soldier nearby a civilian on<br>
> street in Bucha, shooting her.”<br>
> <br>
> <br>
> <br>
> From: Friam <<a href="mailto:friam-bounces@redfish.com" target="_blank">friam-bounces@redfish.com</a>> On Behalf Of Gillian Densmore<br>
> Sent: Wednesday, April 6, 2022 11:09 AM<br>
> To: The Friday Morning Applied Complexity Coffee Group <<a href="mailto:friam@redfish.com" target="_blank">friam@redfish.com</a>><br>
> Subject: Re: [FRIAM] This is scary, and yet very cool...Ai neural networks<br>
> making pictures, look really good<br>
> <br>
> <br>
> <br>
> Oh shit! that's trippy! I though I stumbled over an equally trippy screen<br>
> saver, and (separately) some kind white noise, or background sounds<br>
> generator that used Ai somehow. <br>
> <br>
> Found one of the screen savers that uses Ai. It's free if you let them use<br>
> your GPU and CPU otherwise it's just a few bucks. <a href="https://electricsheep.org" rel="noreferrer" target="_blank">https://electricsheep.org</a><br>
> / <br>
> <br>
> I'm not finding the specific whitenoise maker I tried to help with<br>
> insomnia. Just for that side, worked out great. I don't think it was as<br>
> sophisticated as dall-e2. IIRC I had to give it somehelp with some stuff.<br>
> Maybe so it knows where to start? I didn't think about it till now. Because<br>
> if I like backgrounds that are warm and wholesome, it'd need to know what<br>
> mix together and kinds of tones or something? I have no idea. I'm just<br>
> guessing<br>
> <br>
> electricsheep on the other hand can make some stuff that on the entire<br>
> spectrum from trippy, to that's just cheating: a beautiful scenic town kind<br>
> of things.<br>
> <br>
> <br>
> <br>
> Do you know if they needed to train the Neural Networks so it knows whats<br>
> what? like popart from Lichtenstein and andy worhole or what we might find<br>
> at the indian market. Such that later on you say ahah I want a cow print<br>
> slowcooker picture Dall-e can do that?<br>
> <br>
> <br>
> <br>
> On Wed, Apr 6, 2022 at 11:04 AM Marcus Daniels <<a href="mailto:marcus@snoutfarm.com" target="_blank">marcus@snoutfarm.com</a>><br>
> wrote:<br>
> <br>
> Speaking of which, <a href="https://openai.com/dall-e-2/" rel="noreferrer" target="_blank">https://openai.com/dall-e-2/</a><br>
> <br>
> <br>
> <br>
> From: Friam <<a href="mailto:friam-bounces@redfish.com" target="_blank">friam-bounces@redfish.com</a>> On Behalf Of Gillian Densmore<br>
> Sent: Monday, April 4, 2022 10:01 PM<br>
> To: The Friday Morning Applied Complexity Coffee Group <<br>
> <a href="mailto:friam@redfish.com" target="_blank">friam@redfish.com</a>><br>
> Subject: Re: [FRIAM] This is scary, and yet very cool...Ai neural<br>
> networks making pictures, look really good<br>
> <br>
> <br>
> <br>
> 👍<br>
> <br>
> <br>
> <br>
> On Mon, Apr 4, 2022 at 8:37 PM Marcus Daniels <<a href="mailto:marcus@snoutfarm.com" target="_blank">marcus@snoutfarm.com</a>><br>
> wrote:<br>
> <br>
> I am not a photographer but I have been startled by how recent<br>
> iPhone photos sometimes look better than what I saw when I took it.<br>
> This article explains.. <br>
> <br>
> <br>
> <br>
> <a href="https://www.newyorker.com/culture/infinite-scroll/" rel="noreferrer" target="_blank">https://www.newyorker.com/culture/infinite-scroll/</a><br>
> have-iphone-cameras-become-too-smart<br>
> <br>
> <br>
> <br>
> Filmy-ness sounds like crude interpolation more than structure<br>
> being imposed.<br>
> <br>
> <br>
> <br>
> Suppose I had 1000 pictures of my dog in many lighting conditions<br>
> and from different angles. Using photogrammetry reconstruction<br>
> techniques these could be used to prepare a 3-d textured model of<br>
> her. My iPhone could determine that the 1001st photo also<br>
> included her. It could then reference this model to enhance her<br>
> image in the new context. Maybe inferring the light sources and<br>
> ray tracing her at the required orientation and scale, but at a<br>
> resolution far beyond what was in the photo. That would be more<br>
> art than a photo, but who cares about the truth anymore? Photos<br>
> are to be staged!<br>
> <br>
> <br>
> <br>
> <br>
> <br>
> On Apr 4, 2022, at 5:04 PM, Gillian Densmore <<br>
> <a href="mailto:gil.densmore@gmail.com" target="_blank">gil.densmore@gmail.com</a>> wrote:<br>
> <br>
> Is that also why some of the older software for upscaling<br>
> tricks the new pictures have a kind of saturated or filmy<br>
> thing over them? or is that just from the particular Neural<br>
> Networks or Ai models used? Still very impressive.<br>
> <br>
> <br>
> <br>
> On Mon, Apr 4, 2022 at 5:40 PM Gillian Densmore <<br>
> <a href="mailto:gil.densmore@gmail.com" target="_blank">gil.densmore@gmail.com</a>> wrote:<br>
> <br>
> aaaaaaaaaaaaaaah! coool!! thanks!<br>
> <br>
> <br>
> <br>
> On Mon, Apr 4, 2022 at 5:30 PM Marcus Daniels <<br>
> <a href="mailto:marcus@snoutfarm.com" target="_blank">marcus@snoutfarm.com</a>> wrote:<br>
> <br>
> The idea [1] is that they learn the distribution<br>
> function of different kinds of distortion using a<br>
> machine learning algorithm.<br>
> <br>
> Then that algorithm can invert that distribution<br>
> function. Kind of like a lens can correct for<br>
> nearsightedness.<br>
> <br>
> <br>
> <br>
> [1] <a href="https://arxiv.org/pdf/2107.10833.pdf" rel="noreferrer" target="_blank">https://arxiv.org/pdf/2107.10833.pdf</a><br>
> <br>
> From: Friam <<a href="mailto:friam-bounces@redfish.com" target="_blank">friam-bounces@redfish.com</a>> On Behalf Of <br>
> Gillian Densmore<br>
> Sent: Monday, April 4, 2022 3:25 PM<br>
> To: The Friday Morning Applied Complexity Coffee Group<br>
> <<a href="mailto:friam@redfish.com" target="_blank">friam@redfish.com</a>><br>
> Subject: [FRIAM] This is scary, and yet very cool...Ai<br>
> neural networks making pictures, look really good<br>
> <br>
> <br>
> <br>
> <a href="https://github.com/xinntao/ESRGAN" rel="noreferrer" target="_blank">https://github.com/xinntao/ESRGAN</a><br>
> <br>
> <br>
> <br>
> Stumbled across this looking for a way to gently adjust<br>
> some old pictures of mine without watermarks<br>
> (gigapixel), photoshop wasn't cutting it because not<br>
> enough pixels or data in the originals.<br>
> <br>
> <br>
> <br>
> I am beyond fascinated how do they do it? just guess<br>
> based on colors and add more pixels with that color?<br>
> <br>
> <br>
> .-- .- -. - / .- -.-. - .. --- -. ..--.. / -.-. --- -.<br>
> .--- ..- --. .- - .<br>
> FRIAM Applied Complexity Group listserv<br>
> Zoom Fridays 9:30a-12p Mtn UTC-6 <a href="http://bit.ly/virtualfriam" rel="noreferrer" target="_blank">bit.ly/virtualfriam</a><br>
> un/subscribe <a href="http://redfish.com/mailman/listinfo/" rel="noreferrer" target="_blank">http://redfish.com/mailman/listinfo/</a><br>
> <a href="http://friam_redfish.com" rel="noreferrer" target="_blank">friam_redfish.com</a><br>
> FRIAM-COMIC <a href="http://friam-comic.blogspot.com/" rel="noreferrer" target="_blank">http://friam-comic.blogspot.com/</a><br>
> archives:<br>
> 5/2017 thru present <a href="https://redfish.com/pipermail/" rel="noreferrer" target="_blank">https://redfish.com/pipermail/</a><br>
> <a href="http://friam_redfish.com/" rel="noreferrer" target="_blank">friam_redfish.com/</a><br>
> 1/2003 thru 6/2021 <a href="http://friam.383.s1.nabble.com/" rel="noreferrer" target="_blank">http://friam.383.s1.nabble.com/</a><br>
> <br>
> <br>
> .-- .- -. - / .- -.-. - .. --- -. ..--.. / -.-. --- -. .--- ..-<br>
> --. .- - .<br>
> FRIAM Applied Complexity Group listserv<br>
> Zoom Fridays 9:30a-12p Mtn UTC-6 <a href="http://bit.ly/virtualfriam" rel="noreferrer" target="_blank">bit.ly/virtualfriam</a><br>
> un/subscribe <a href="http://redfish.com/mailman/listinfo/" rel="noreferrer" target="_blank">http://redfish.com/mailman/listinfo/</a><br>
> <a href="http://friam_redfish.com" rel="noreferrer" target="_blank">friam_redfish.com</a><br>
> FRIAM-COMIC <a href="http://friam-comic.blogspot.com/" rel="noreferrer" target="_blank">http://friam-comic.blogspot.com/</a><br>
> archives:<br>
> 5/2017 thru present <a href="https://redfish.com/pipermail/" rel="noreferrer" target="_blank">https://redfish.com/pipermail/</a><br>
> <a href="http://friam_redfish.com/" rel="noreferrer" target="_blank">friam_redfish.com/</a><br>
> 1/2003 thru 6/2021 <a href="http://friam.383.s1.nabble.com/" rel="noreferrer" target="_blank">http://friam.383.s1.nabble.com/</a><br>
> <br>
> <br>
> .-- .- -. - / .- -.-. - .. --- -. ..--.. / -.-. --- -. .--- ..- --.<br>
> .- - .<br>
> FRIAM Applied Complexity Group listserv<br>
> Zoom Fridays 9:30a-12p Mtn UTC-6 <a href="http://bit.ly/virtualfriam" rel="noreferrer" target="_blank">bit.ly/virtualfriam</a><br>
> un/subscribe <a href="http://redfish.com/mailman/listinfo/friam_redfish.com" rel="noreferrer" target="_blank">http://redfish.com/mailman/listinfo/friam_redfish.com</a><br>
> FRIAM-COMIC <a href="http://friam-comic.blogspot.com/" rel="noreferrer" target="_blank">http://friam-comic.blogspot.com/</a><br>
> archives:<br>
> 5/2017 thru present <a href="https://redfish.com/pipermail/" rel="noreferrer" target="_blank">https://redfish.com/pipermail/</a><br>
> <a href="http://friam_redfish.com/" rel="noreferrer" target="_blank">friam_redfish.com/</a><br>
> 1/2003 thru 6/2021 <a href="http://friam.383.s1.nabble.com/" rel="noreferrer" target="_blank">http://friam.383.s1.nabble.com/</a><br>
> <br>
> <br>
> .-- .- -. - / .- -.-. - .. --- -. ..--.. / -.-. --- -. .--- ..- --. .-<br>
> - .<br>
> FRIAM Applied Complexity Group listserv<br>
> Zoom Fridays 9:30a-12p Mtn UTC-6 <a href="http://bit.ly/virtualfriam" rel="noreferrer" target="_blank">bit.ly/virtualfriam</a><br>
> un/subscribe <a href="http://redfish.com/mailman/listinfo/friam_redfish.com" rel="noreferrer" target="_blank">http://redfish.com/mailman/listinfo/friam_redfish.com</a><br>
> FRIAM-COMIC <a href="http://friam-comic.blogspot.com/" rel="noreferrer" target="_blank">http://friam-comic.blogspot.com/</a><br>
> archives:<br>
> 5/2017 thru present <a href="https://redfish.com/pipermail/friam_redfish.com/" rel="noreferrer" target="_blank">https://redfish.com/pipermail/friam_redfish.com/</a><br>
> 1/2003 thru 6/2021 <a href="http://friam.383.s1.nabble.com/" rel="noreferrer" target="_blank">http://friam.383.s1.nabble.com/</a><br>
> <br>
> <br>
> .-- .- -. - / .- -.-. - .. --- -. ..--.. / -.-. --- -. .--- ..- --. .- - .<br>
> FRIAM Applied Complexity Group listserv<br>
> Zoom Fridays 9:30a-12p Mtn UTC-6 <a href="http://bit.ly/virtualfriam" rel="noreferrer" target="_blank">bit.ly/virtualfriam</a><br>
> un/subscribe <a href="http://redfish.com/mailman/listinfo/friam_redfish.com" rel="noreferrer" target="_blank">http://redfish.com/mailman/listinfo/friam_redfish.com</a><br>
> FRIAM-COMIC <a href="http://friam-comic.blogspot.com/" rel="noreferrer" target="_blank">http://friam-comic.blogspot.com/</a><br>
> archives:<br>
> 5/2017 thru present <a href="https://redfish.com/pipermail/friam_redfish.com/" rel="noreferrer" target="_blank">https://redfish.com/pipermail/friam_redfish.com/</a><br>
> 1/2003 thru 6/2021 <a href="http://friam.383.s1.nabble.com/" rel="noreferrer" target="_blank">http://friam.383.s1.nabble.com/</a><br>
> <br>
<br>
> <br>
> .-- .- -. - / .- -.-. - .. --- -. ..--.. / -.-. --- -. .--- ..- --. .- - .<br>
> FRIAM Applied Complexity Group listserv<br>
> Zoom Fridays 9:30a-12p Mtn UTC-6 <a href="http://bit.ly/virtualfriam" rel="noreferrer" target="_blank">bit.ly/virtualfriam</a><br>
> un/subscribe <a href="http://redfish.com/mailman/listinfo/friam_redfish.com" rel="noreferrer" target="_blank">http://redfish.com/mailman/listinfo/friam_redfish.com</a><br>
> FRIAM-COMIC <a href="http://friam-comic.blogspot.com/" rel="noreferrer" target="_blank">http://friam-comic.blogspot.com/</a><br>
> archives:<br>
> 5/2017 thru present <a href="https://redfish.com/pipermail/friam_redfish.com/" rel="noreferrer" target="_blank">https://redfish.com/pipermail/friam_redfish.com/</a><br>
> 1/2003 thru 6/2021 <a href="http://friam.383.s1.nabble.com/" rel="noreferrer" target="_blank">http://friam.383.s1.nabble.com/</a><br>
<br>
<br>
-- <br>
<br>
----------------------------------------------------------------------------<br>
Dr Russell Standish Phone 0425 253119 (mobile)<br>
Principal, High Performance Coders <a href="mailto:hpcoder@hpcoders.com.au" target="_blank">hpcoder@hpcoders.com.au</a><br>
<a href="http://www.hpcoders.com.au" rel="noreferrer" target="_blank">http://www.hpcoders.com.au</a><br>
----------------------------------------------------------------------------<br>
<br>
<br>
.-- .- -. - / .- -.-. - .. --- -. ..--.. / -.-. --- -. .--- ..- --. .- - .<br>
FRIAM Applied Complexity Group listserv<br>
Zoom Fridays 9:30a-12p Mtn UTC-6 <a href="http://bit.ly/virtualfriam" rel="noreferrer" target="_blank">bit.ly/virtualfriam</a><br>
un/subscribe <a href="http://redfish.com/mailman/listinfo/friam_redfish.com" rel="noreferrer" target="_blank">http://redfish.com/mailman/listinfo/friam_redfish.com</a><br>
FRIAM-COMIC <a href="http://friam-comic.blogspot.com/" rel="noreferrer" target="_blank">http://friam-comic.blogspot.com/</a><br>
archives:<br>
5/2017 thru present <a href="https://redfish.com/pipermail/friam_redfish.com/" rel="noreferrer" target="_blank">https://redfish.com/pipermail/friam_redfish.com/</a><br>
1/2003 thru 6/2021 <a href="http://friam.383.s1.nabble.com/" rel="noreferrer" target="_blank">http://friam.383.s1.nabble.com/</a><br>
</blockquote></div>