andersource

8 ms·

1. 181.▲
   by andersource 5y ago

   I would say at latest when people convicted because of it had their names cleared - https://www.bbc.com/news/business-56859357
2. 182.▲
   by andersource 5y ago

   Most of the time, but it can , and I think that's the issue a lot of people have with it. https://twitter.com/mitsuhiko/status/1410886329924194309 https://news.ycombinator.com/item?id=2771028
3. 183.▲
   by andersource 5y ago

   You can see the trailer for the game they developed here: https://www.youtube.com/watch?v=F2J11YFTsAE > the first games were played on machines that only had text mode This is also mentioned in the video, the author&#x27
4. 184.▲
   by andersource 5y ago

   There's a beautiful visualization of prime factorization using UMAP: https://johnhw.github.io/umap_primes/index.md.html . I recommend watching the video!
5. 185.▲
   by andersource 5y ago

   Gladly! > So is "T" and "N" conventionally used in numerical optimization as the transform "variables" (like X and Y are in calculus)? Not really, numerical optimization offers a general toolbox for various
6. 186.▲
   by andersource 5y ago

   Hey! I is the flattened image (i.e., a matrix where each row represents the RGB values of a single pixel) T and N are some matrices, basically of appropriate dimensions such that the transformation you quoted results in an image of the same
7. 187.▲
   by andersource 5y ago

   Thanks! The optimal transport is really cool!
8. 188.▲
   by andersource 5y ago

   > "Flowers replaced to red" looks like it also tinted the whole sky red. That's true, to me at least it looks like in the original image the sky has a yellow undertone so I think it makes sense
9. 189.▲
   by andersource 5y ago

   Thanks! Yeah, I think one thing the simplistic optimization is missing is preserving the "texture" of the image. There are various numeric ways to represent textures so this is something that potentially could be added to the obje
10. 190.▲
    Image color replacement with numerical optimization(andersource.dev)

    78 points by andersource 5y ago|17 comments
11. 191.▲
    by andersource 5y ago

    Those are incredible, thank you!
12. 192.▲
    by andersource 5y ago

    OT but would you mind sharing some of it? That sounds really interesting.
13. 193.▲
    by andersource 5y ago

    That's really nice!
14. 194.▲
    by andersource 5y ago

    Not very neatly organized but it's almost all here: https://github.com/andersource/andersource.github.io/blob/ma...
15. 195.▲
    by andersource 5y ago

    Hey, what a surprise to see this here! This came up in a discussion a few weeks ago and there were some interesting suggestions, I'm itching to follow up on the swarm with "boid" behavior one.
16. 196.▲
    Image color replacement with numerical optimization(andersource.dev)

    2 points by andersource 5y ago|0 comments
17. 197.▲
    by andersource 5y ago

    Haven't read thoroughly but it seems they are investigating ResNet models [0] trained for image classification. > We apply CKA to a family of ResNets of varying depths and widths, trained on common benchmark datasets (CIFAR-10, CIFA
18. 198.▲
    by andersource 5y ago

    I didn't know there was a standard nomenclature for that! Though of course there is :) And yes, another commenter noted on that, it would make for an interesting addition to the problem.
19. 199.▲
    by andersource 5y ago

    Good catch! The technique needs modification to handle uncertainty (i.e. if we don't know which variation of the digit this clock uses), by relaxing the constraints around the uncertain LED segments for the digits with variations. My g
20. 200.▲
    The faulty digital clock problem(andersource.dev)

    91 points by andersource 5y ago|16 comments
21. 201.▲
    by andersource 5y ago

    That's a great idea!
22. 202.▲
    by andersource 5y ago

    Thanks so much! Yeah that idea is really itchy now :) I guess some people don't see (at least immediately) the artistic value in such things, and then it seems rather pointless.
23. 203.▲
    by andersource 5y ago

    > I think to a large extent coding can be like making art. That resonates very strongly with me. > I enjoyed doing it. People don't seem to think that's good enough. Yeah, I've found that a lot of times when you want to
24. 204.▲
    by andersource 5y ago

    Thanks! And that's really cool, added to the TODO list!
25. 205.▲
    by andersource 5y ago

    Thanks, and that's great! Due to experiences like the above I would probably not even think of bringing this up in a regular interview, unless I get the vibe that the interviewer might actually find it interesting for its own sake (whi
26. 206.▲
    by andersource 5y ago

    Thank you, I thought of writing one, this might give me the motivation to finally do it!
27. 207.▲
    by andersource 5y ago

    Thanks! First I'm generating the general geometry of the wings - the most basic shape is a "teardrop" / half of the infinity symbol (see this[0]). Then they are scaled and rotated randomly (within a range to match the ot
28. 208.▲
    by andersource 5y ago

    I strongly relate. I spent a couple of weekends working on some procedural art [0] and was pretty happy with the end result. When I showed it to some of my co-workers / manager, they were baffled . "Uhh.. OK. Why did you make thi
29. 209.▲
    by andersource 5y ago

    Thanks! Yeah that's my idea, I guess what surprised me is the ranges involved. In the few times I had to negotiate (always as a "seller" as in salary negotiation), I would state around 110% of my minimum, and I was really pre
30. 210.▲
    by andersource 5y ago

    Thanks! So basically (typically, of course depending on circumstances) any number explicitly stated by the buyer is less than the true maximum and you can expect to converge on a higher price?