SoftwareStudies

When did photography really become “automated”? Not just with AI or digital cameras, but across two centuries of technical changes in capture, editing, and distribution.

I just finished a long interactive timeline that maps this process from early 19th‑century photography to AI image generation today, building on ideas from my dissertation and later articles.

https://levmanovich.github.io/manovich-interactive/photo/

Why do most artists struggle today? And why it won't be fixed?

I offer only one narrative among many — and many will not agree. However, I think that tracing the arc from the pre-modern "golden age" through modernism, its eventual acceptance, and the social turn of the 21st century helps explain a crisis that isn't going away. Too many artists, too few buyers, and a widening gap between what institutions celebrate and what most people outside of the art world actually care about.

Swipe for the argument →

If you are teaching any university classes right now - can you help me? I am collecting global responses for a simple (but large-scale) survey about people's lifestyles & cultural likes. It takes < 10 min to fill out the form. If you can send the link to your students and encourage them to fill it out, this will be great. And of course, feel free to fill it yourself:

https://forms.gle/PNfwDu5S3hn3Sh4P8

The aggregated results will be used in a new cultural analytics & visualization project I am working on — and of course, all results and visualizations will be published online.

Lab's earlier projects:

https://lab.culturalanalytics.info/p/projects.html

***

Image in this post - from our project Phototrails (2013) - first to visualize large collections of Instagram photos:

http://phototrails.info/instagram-cities/

Join me for an online seminar " Medium That Thinks: Generative AI, Media Cognition, and Artistic Creation " on May 13th, 10:00 UTC/GMT (The Department of Film, Theatre & Television, The University of Reading, UK).

Please RSVP to receive a link:
https://www.eventbrite.co.uk/e/a-medium-that-thinks-generative-ai-media-cognition-and-artistic-creation-tickets-1987383797437?aff=oddtdtcreator

Today I made this dashboard to keep track of all professional invitations I am getting. First, I designed it using Perplexity AI and then finished the design in Claude.

I am releasing a new version of ImagePlot 2. This free software runs in your browser and it visualizes image collections. It can use existing image metadata - or 48 visual features the tool extracts. It can visualize using PCA, t-SNE, do animation, and more.

Download ImagePlot 2.1, code (HTML), its documentation, and sample image collections:

https://drive.google.com/drive/folders/1F-IShu5Uu_l0xY1wZTKfRc3keJJNtWQV?usp=sharing

This version only fixes bugs - but the next one I hope to release later this week adds a new feature - two visualizations side by side - to compare two different image collections or two parts of the same collection. So far, I have run the tool with 1000 images and have not seen any slowdown in performance.

***

If you are new to ImagePlot, you can find many examples (which use earlier versions) with texts explaining what, how, and why of each of them:

https://lab.culturalanalytics.info/p/projects.html

My talk today at Paik After Paik symposium today (Seoul). For me, Paik is an inspiration for thinking about new and different interfaces for our times - which for me is characterized by these features:

1) hundreds of millions are "broadcasting" on social media platform (creating media at scale impossible for us to follow in detail); 2) a large proportion of all these messages is the copy of much smaller corpus of original information;
3) Al enables very rapid experimentation and apps and Uls development;
4) we are stuck with older modes such as linear one column posts presentation, "summaries" of information, extraction of "trends" etc.

Yes they are working - but I wonder what other different modes we can invent? And what Al can do in this respect which we are not thinking about yet ?

This visualization is a pilot for a larger study of diffusion of new concepts across different languages. The pilot tracks how eight new concepts spread between 2000 and 2025, using creation dates of new Wikipedia articles in six language editions (English, Chinese, Spanish, German, French, and Japanese). I treat the creation of each new Wikipedia article as a marker of when a concept enters a language community — allowing me to trace how the gaps between adoption of new terms across languages shift over time.

For most of this period, English Wikipedia leads. New concepts tend to appear there first, with other languages following later. Some concepts—like “startup” or “coworking”—enter several languages within a few years, while others, such as “gig economy” and “platform economy,” took much longer to spread beyond English.
Overall, these lags tend to shorten after the mid‑2010s — suggesting that the diffusion of new concepts across languages has sped up significantly. For concepts that first appeared in English in the 2000s, the average lag for other languages to follow was 5.5 years. For concepts that entered English in the 2010s, this dropped to 2.3 years.

Around 2020, one particular relationship starts to change — between English and Chinese Wikipedia articles. Some terms enter the Chinese edition very quickly, and in a few cases even before English — suggesting that new concepts may now sometimes travel from Chinese into English, rather than always in the opposite direction.

This kind of analysis is an example of what I call cultural analytics: studying cultural patterns and trends using data. Today, AI allows me to run such pilot studies in 20–30 minutes — compared to the months or years that grant applications, data collection, and processing once required.