Business models, busines schmodels said Mandrake and made a magical gesture.
“A business model or company model, English business model, is in business administration a theoretical description of how a company, or a business operation, is intended to function. It is a conceptual tool that contains a set of components and describes their interrelationships in such a way that the business logic for a particular activity can be concretely described.” https://sv.wikipedia.org/wiki/Aff%C3%A4rsmodell
The above is hardly anything remarkable. Clearly, all companies should have the right to decide how they earn their money. This little rant is instead about the frustration of feeling deceived or at least seduced by a business model. Probably because it was never presented as just a business model but as the only obvious way to do something.
One example of these models is the chatbot that we chat with in a dialogue window. At least most of us do. Well, now I obediently have to sit and copy and paste what the AI tells me to write, do, or say. But as many have already figured out, this is not the best way to use language models. Letting them write directly into the documents you want to edit is, of course, immensely more logical and efficient. Or at least that’s how I felt when Claude co-work filled in a series of forms for my tax declaration. But why has this feature taken so long? Of course, because it’s important that we continue to pay for the service that includes the chat window and then a cohesive experience on the AI provider’s site. The risk is that if you leave the site where you chat and are lulled into the notion that the AI knows me. The risk is that you discover that there are only a few use cases where a dialogue box with an obliging chatbot is the best way to solve this.
So-called memory is another feature that is also part of a hidden business model. What we call memory in a chat GPT simulates human behavior with the aim of making us feel warmer feelings for the service and, in the worst case, anthropomorphize it. It’s manipulative brand magic. Yes, but there are certainly purely functional advantages to memory. But it doesn’t work at all as we imagine. The fact is that every time you send a chat message, you are talking to a language model with a completely blank memory. It has no idea who you are. It’s not even certain that you are talking to the same model between two exchanges. That’s why the chatbot has to send along everything you’ve said before, all information about you, and all relevant memories every time you naively press send.
The last example is the various strategies of social media to succeed in breaking through. You should post x number of times per day, you should post at certain times, you should be easily digestible and predictable. Why? It’s certainly not so that we users get a better experience. You might understand where I’m going with this by now. It’s because frequent posting benefits social media itself. They want us to believe that this behavior is a necessity for success, but in reality, it’s part of a clever business model that generates content for the platform and entices with promises of going viral. Like an American dream, but for content creation.
In summary: Business models should be transparent. If they are hidden under layers of manipulative user design and packaged as something you must adapt to, there is an imminent risk that you have been deceived.