top of page


When "Avengers: Endgame" was came out this year, I saw a piece of news about OpenAI's text generator GPT-2. It can recognize all kinds of input, from news, lyrics, poetry, recipes, code, and even "The Lord of The Rings" and "The Avengers" themes, GPT-2 can perfectly connect to the context as long as you typed a beginning. Many netizens reported that the script for the "Avengers: Endgame" was more moving than the script written by the real screenwriter. Although it is still some way from the engineer's dream of a universal AI, GPT-2 is excellent at this stage, whether it's generating fake news or writing spam. It is conceivable that the proliferation of robot accounts on social networks will only become more illegible in the future. Algorithm can not only tell stories, but also may tell better than human beings in the future.


I want to start with the field of photography that I am most familiar with. Technical images are difficult to crack. Obviously, it doesn't need to be decoded, because its meaning is automatically reflected on the surface, just like a fingerprint, its meaning (finger) is the cause, and the image (imprint) is the result. (Flusser, V. 2018:15) Like all images, it is not only symbolic, but also a more abstract symbolic complex than traditional images. It is the meta-code of the text, it means the text, only indirectly means the outside world. The imagination in producing images is the ability to convert concepts from text (conversion codes) to images; when we look at images, we see re-coded concepts about the outside world.


But it seems that the symbolic complex of the camera and the photographer does not break the chain between the image and its meaning. Meaning seems to enter the complex from one side (that is, Input), flows out from the other side (that is, the process of Output), takes place within the complex. But it's always hidden: in fact, the complex is a "Black Box”. (Ashby, W. R.1956:86) However, the coding process of the technical image is carried out within the black box, so any criticism of the technical image must be clear about the internal operation of the black box. As long as we cannot criticize the technical image like this, then for the technical image, we are always illiterate.


Technical images absorb the whole history and form a kind of social memory that circulates forever. Nothing can resist the attraction of technical images: there is no artistic, scientific or political activity that is not directed at technical images, and there is no daily activity that does not desire to be taken into a photograph, film or video. Because everything is eager to enter such a permanent memory, can always be copied. 


Benjamin* has evaluated many works of the era of mechanical reproduction, but he has also said: “The long exposure process makes the photographer is not “live" out of the moment, but 'live' into which, in the long exposure process He seems to have settled in the image.” I prefer to keep some of our "aura" in the picture. Even after we leave this world, the "aura" preserved through photography will still be able to watch the world and talk with our viewers.


Technical developments are changing the world-algorithms are changing our world. Digitization has changed the way people think, and the development of algorithms has changed the tools we use to think. Today, the density of information is unbelievably high. Therefore, in fact, only fragments information can be processed. We may have gained more absolute knowledge than ever before, but everything has been shredded-just like a "shard" file on a hard drive. The holistic view or integrity no longer exists. I have tried to make a mini-program which can automatically overlap maximum 9 pictures you have added and make a new picture eventually. When we take pictures, we are conscious, but in this program is unconscious. Its act of generating images is like a fragmentation in subjective photography.


Then I begin to wonder whether the machine can do what people do. The program has no thoughts, but the photos generated by it could have psychological interaction with the viewers. People are used to focusing on meaningful things and are equally good at building a meaningful world. For the pictures, meaning corresponds to people's life, dreams, childhood, desires, etc. However, the image produced by the program language itself has become a neglected island away from the meaning-building continent. In an interview in 1987, photographer Friedlander said: "Other people may move a little for staying away from those obstacles, namely poles or trees, but for me, I will move for recording them in my photos. This is what I usually do, which is sometimes an unconscious act. " (Phillip,G. 2009:21) Meaningless things created by machines have more possibilities and can deepen such perception and insights. "Those things shown in pictures are not common experiences of a small group. " but are the experiences with more possibilities. Those things can be anything and anywhere. They become part of ourselves throughout the process.







Reference list:

Flusser, V. (2018). Towards a philosophy of photography. London: Reaktion Books, p.15.


Ashby, W. R. (1956) ‘Black-box theory’ Introduction to Cybernetics. London: Methuen Books, p.86.


Benjamin, W. (1998) Essais, the most important chapter of The Work of Art in the Age of Mechanical Reproduction has always worn the word “aura”. Available at: [accessed19 September 2017]


Phillip,G. (2009).Photography After Frank.New York: Aperture. ISBN 978-1597110952, said Friedlander could find out the texture of nature from the world that we lived in but we turned a blind eye to before he mixed all the factors together and presented them. The wisdom including perception, logic and insight have been shown in his observation of daily life through vision language.

  • 白色的Instagram圖標
  • 邮箱
  • Instagram
  • email
bottom of page