In the Midjourney subreddit, the post of the user reciperobotai is gaining popularity. The author posted two photos of one dish: baked goat cheese with tomato sauce and garlic bread. Here's just one photo taken by a person, and the second is a new version of the neural network. These are the pictures:
Almost 700 comments have gathered under the post, in which they discuss where what picture is and what the neural network gives out. Can you determine the work of AI?
"Most of you guessed it: the right photo was taken by Midjourney. I chose the option without the most obvious flaws, but you saw them here too. Depth of field, focus, slightly strange location and texture of bread..."
"In fact, the sprinkling of greens gave the AI away in a second. The color is too even, which is why the greenery seems to float from above and looks wrong. And I doubt that Midjourney would have been able to draw the asymmetrical chopped greens as well as in the photo on the left."
"The frying pan gives me the right answer: the left one looks like it's printed on a 3D printer."
"Usually things at the same distance from the camera should be approximately equally blurred. Midjourney often has problems with this. In the photo on the right, the bread and the sauce under it should be at the same distance from the camera, but the bread is in focus, and the sauce is not."
Were you able to guess the correct answer? If not, don't worry: The commentators above are members of the Midjourney community and are used to the work of the neural network and its typical problems — unlike the average user. However, at this rate, it will be difficult to distinguish real photos from generated ones, even to a trained eye.