| ▲ | santhoshr 5 hours ago |
| Pelican riding a bicycle: https://pasteboard.co/CjJ7Xxftljzp.png |
|
| ▲ | xnx 5 hours ago | parent | next [-] |
| 2D SVG is old news. Next frontier is animated 3D. One shot shows there's still progress to be made: https://aistudio.google.com/apps/drive/1XA4HdqQK5ixqi1jD9uMg... |
| |
|
| ▲ | mohsen1 5 hours ago | parent | prev | next [-] |
| Some time I think I should spend $50 on Upwork to get a real human artist to do it first to know what is that we're going for. What a good pelican riding a bicycle SVG is actually looking like? |
| |
| ▲ | AstroBen 5 hours ago | parent [-] | | IMO it's not about art, but a completely different path than all these images are going down. The pelican needs tools to ride the bike, or a modified bike. Maybe a recumbent? |
|
|
| ▲ | robterrell 5 hours ago | parent | prev | next [-] |
| At this point I'm surprised they haven't been training on thousands of professionally-created SVGs of pelicans on bicycles. |
| |
| ▲ | notatoad 5 hours ago | parent [-] | | i think anything that makes it clear they've done that would be a lot worse PR than failing the pelican test would ever be. | | |
| ▲ | imiric 4 hours ago | parent [-] | | It would be next to impossible for anyone without insider knowledge to prove that to be the case. Secondly, benchmarks are public data, and these models are trained on such large amounts of it that it would be impractical to ensure that some benchmark data is not part of the training set. And even if it's not, it would be safe to assume that engineers building these models would test their performance on all kinds of benchmarks, and tweak them accordingly. This happens all the time in other industries as well. So the pelican riding a bicycle test is interesting, but it's not a performance indicator at this point. |
|
|
|
| ▲ | bn-l 5 hours ago | parent | prev [-] |
| It’s a good pelican. Not great but good. |
| |
