Do you know how you feel when you read something clearly written by an AI?
This is happening more and more and it saddens me to see real human communication usurped, even if it means fewer typing errors. It is clear where we are headed and untold billions will continue to be created but it is haphazard, unverifiable and often confidently wrong.
If this image is any indication, the same problems are coming in the real world.
An OpenAI employee posted two images of the ‘OpenAI Library’ this week, presumably at their San Francisco headquarters: the one above and this one.
If you’ve done any image encoding with AI, it’s obvious what it is. Of course it was conceived by ChatGPT and then some expert craftsmen were brought in to transform the computer vision into real life.
In principle, I like this kind of thing. This is a great app for improving your own home and for interior designers at disastrous events.
But seeing it in real life is disturbing. Aesthetics are the disturbance of some kind of machine’s sense of harmony, not the normal use of space by man. This is a sign of ‘balanced timelessness’, with the soft lighting being very cinematic and typical of AI images.
But I’m afraid it doesn’t workIt’s as if Sam Altman created the image, gave it to the workers and said ‘do it exactly like this’ without anyone pushing back. For one thing, the lighting doesn’t work for the library. There is modest natural light coming in from the window but at night it won’t be bright enough to read anything in most of the room, especially those cozy looking black sofas. Secondly, the books look like props. It doesn’t look like a real library where the shelves are full and people actually read books and instead it feels like they were chosen for the sizes and colors of their bindings.
And seriously, pay attention to the plants on the top shelves. Now that beauty may look good in photos but in real life plants need light. There are no plants on Earth that can survive in those places. Furthermore, the person who posted the image said that they were “mostly cacti but it is true that they looked a little wilted last night”. So while the cactus may have less need to climb that ladder for water, it will still need hours of direct sunlight every day. Even a floor snake plant (a notoriously indestructible houseplant) probably won’t be able to do this.
Why didn’t anyone look at those cacti and say, ‘Wait, won’t they die?’
This is a preview of the world to come. It looks beautiful but the information given in it is hollow. The image is beautiful, but it will not sustain the life for which it was designed. In short, it’s a big sales pitch, not something that fixes a real-world problem.
The machine will tell us what is best and we will have to live with the results. It amazes and scares me that the people at OpenAI are so enslaved to these machines that they created them without even noticing the obvious flaws.
The person who posted this image works in ‘Human Data’ at OpenAI. These are the people who should be most aware of its disadvantages, such as its tendency to be “confidently wrong” or to optimize for looks rather than function, yet they replicate those exact disadvantages in the real world. What does this say about their ability to control or direct it?
This alignment is not what it looks like.
