AI, again

Martin:

Good Morning Tim! How are you? I am sorry to annoy you with a question about AI, again. Two years ago we sat on the couch of the Taaalks podcast (https://flexiblevisualsystems.info/resources/taalks/) and were quite critical of AI. We are still critical with AI, but we also started using it for a couple of tasks. Did you change your opinion on AI? What are you using AI for?

Tim:

You don’t annoy me anyway 🙂 I think it’s really important and valuable to be able to work without AI and not build dependencies. I design my systems so that I can use them without AI and avoid it wherever possible, especially when it comes to truly creative work. I think my limitations make my work human. My boundaries guide me the way.

I strictly separate two areas of my work. When it comes to creative work involving writing text or code, I avoid AI. I don’t generate content, but I do use it for translations. I never use image generators. And I never copy and paste generated code into my Processing Editor.

However, when it comes to purely technical tasks, such as developing a component for a website or wire up a user interface for a visual system, I am a little less strict. But I always ask myself one question: Could I build this system myself?

And to be honest, I almost always have a bad feeling about it.
How about you? How do you deal with it?
Is it maybe a new kind of temporal intelligence to work without AI?

Martin:

I basically use it in a similar way. If there is something technical that does not require creativity I automate. When I say “creative” I am not just referring to visual design work, I am referring to every thought process that isn’t purely logic. AI is great in predicting the usual logical route some path would lead you, but it doesn’t help if you want to challenge the “normal”. I need to write, draw and design on my own, because working with text or images is a reflective process. Sometimes even seemingly dull work gives me the time to think about what I am actually doing and how it might need to be different.

A creative thought process is also a learning process. You can’t learn if you automate. You can’t outsource learning. You need to do it for yourself. Learning is highly creative. To me learning is not to memorize existing knowledge, but making it your own. To weave the new into the old you have in you. It is not about finding logical patterns but about finding surprisingly new connections. This can happen at first on a purely intuitive layer. It is only later that logic can help you as a guidance why this could also make sense to others. It is a communication tool.

I can give you an example of what I outsource. For example, if I had a meeting, I use a transcript tool to sum up what has been talked about, and restructure what has been said into an action plan. Sometimes AI gets it wrong and misses the big picture, what really matters, but often it also surprises me because something I forgot, ignored or did not pay attention to is in the transcript and reminds me to include it. AI is great in capturing information, relevant and irrelevant, restructuring it with your prompts and quickly sharing it with others. It can be of great help as a reminder, but I would never use it for my output. I hate reading text written or co-written by AI. There is no overarching sense-making, it feels fragmented, redundant and empty of, for a lack of a better word, soul. If I don’t feel words are contextualised by human complexity, I don’t want to give it my valuable time. I don’t read it.

You said you feel bad using AI. Why is that?

Tim:

As you know I have built this application for us, the Coding Systems Pullquote Design app. To do this in a manageable timespan, I have renewed my subscription to Cursor, an IDE with AI built in. I needed quite a few prompts and at least two days of work to make it run the way I wanted it.

But every prompt came with these doubts – and working with AI takes a lot of energy, on many levels. Working with AI is the norm today, everywhere. The people within the extractive economy can mostly not afford to question or even reject this new way of working. It’s an arms race. People are trapped in this mode of acceleration and the force of efficiency.

I see how the internet gets flooded with terrible AI slop. I discover terribly vibe coded websites and tools and semi-human content, where it is not clear if its machine made or not. It feels like we are burning down the last forest to win the race.

“AI is the asbestos we are shovelling into the walls of our society, and our descendants will be digging it out for generations.”

Cory Doctorow:

Martin:

AI is even worse than asbestos because we know where the asbestos is hidden. AI undermines anything digital. Even science is not safe from it. The half-truths, hallucinations and bullshitting is everywhere. They are even becoming the source for new AI datasets. At a speed, that makes fact checking harder and harder.

I feel it would be a good moment to talk about the HI behind the AI. As humans we tend to see AI as a being. It is not. It is, in comparison with living beings, a very simple construct. Still, we are in awe of it. The very same thing that creates the mess, is the thing that we love, fear, worship, and curse it for. It’s (apparent) autonomy. It is more than just efficiency. It takes the weight of responsibility off our shoulders. It knows all the answers and if it doesn’t, it invents them. We become intimidated, lazy and impatient. We make ourselves dependent on automation of reasoning. We are stripped of our autonomy.

This is the perfect business model. Create a product nobody can live without and you will have demand for life. As everything we really need for a good life has already been created, we need to make the current version of life look like shit and sprinkle little dopamine and serotonin on our product to make it look like an improvement.

Coming back to design. I have had that sensation since I started studying. I have seen a lot of change, but I haven’t seen any real improvement in design by technology. If I look at the books by Aldus Manutius and his punchcutter Francesco Griffo I am amazed by the quality. This is from fucking half a millenial ago. What did we use that time for? I think we should have a good hard look at what really matters. The book, like the bicycle, is still one of the best inventions we came up with. Their positive effects are numerous and can’t be measured by output, but have to be evaluated systemically by their outcome.

I also have a quote for you, by Kate Raworth: “Don’t be an optimist if it makes you relax, don’t be a pessimist if it makes you give up. I say be an activator, activate in the spaces where you have influence”. I like that quote because it looks at what we activate by our actions. We went through a century of self (Adam Curtis) and it is time to see ourselves as part of something bigger again.

Tim:

I couldn’t agree more. With Coding Systems, we can be activators. We spoke about this some time ago and shared the conversation on our website. But maybe, building on this discussion, what do you think, how could we be activators in the context of technology and design?

My five cents: I think design education is an amazing opportunity for us, it’s one of the very few areas where design can still be relatively free from the forces and pitfalls of the market. Could we use this potential to design life-enobling technologies? How?