Agents negotiating toward truth machines

By Iskander Smit in weeknotes — Feb 4, 2025

The latest notions from the developments in human-AI partnerships (yes also DeepSeek), and thoughts triggered by lessons from the global south and blockchain reflections.

Midjourney interpretation

Weeknotes 325

Thanks for landing here and reading my weekly newsletter. If you are new here, have a more extended bio on targetisnew.com. This newsletter is my personal weekly reflection on the news of the past week, with a lens of understanding the unpredictable futures of human-ai co-performances in a context of full immersive connectedness and the impact on society, organizations, and design. Don’t hesitate to reach out if you want to know more or more specifically.

Hi all!

In the first term, I followed an email newsletter that did an overview every day of actions by Trump. I am not sure if it is restarted, but I don’t think I can handle it. What a terrible week again. But hopefully he will get push back…

Let’s focus on other things… I am still tweaking the format. As the first month was quite busy, I always ran into the problem of having too little time while typing this. Last week I introduced the links to the references directly with the text. Today I use a slightly different layout for the headings and links. Hope you like it. Let me know what you think after a month in this new setup.

What did happen last week?

Every week, I try to keep this short. Not always successful. Last week, I worked a lot on the last sprint (iteration) of the research report on civic protocol economies (this project). I visited one movie on VHS culture that is recommended if you are nostalgic (my Letterboxd review). And the Safe.SelfSense installation/movie in WORM UBIK was very nice too (IFFR link). I was at a packed evening on democracies in AI; also, see this week’s Triggered Thought (link event page). And I got some time to start planning the next iteration of exhibitions of “Generative Things” and RIOT (more later this week).

What did I notice last week?

Last week’s DeepSeek saga continued with numerous more reflecting articles, wonder what it actually means. And downplaying the quality a bit, and the cheapness a lot. As also mentioned last week, it mainly tells a lot about the unstable market of AI on the one hand, and is it real nice hint towards the next iteration in our interactions with AI. The design angle is nicely described by Platformer and LukeW.>.>.>.>.>.

The performance of DeepSeek was also tested in different forms. The Chinese roots became clear in how it censors the answers, like informing about Ai Weiwei.>.>.>

Perplexity announced on Monday that DeepSeek is part of their engine on EU servers and without censorship.>.

OpenAI released the o3 mini as expected to counter the DeepSeek sentiment and remains introducing new functionality.>.>.

Comparing AI assistants, this will definitely become common.>.>.

Ethan Mollick is framing it as the end of the search and the beginning of the research. Reasoning models will generalize Nathan Lambert explains.>.>.

AI is an excuse to create a distance between those with power and those their power touches. Is the human in the loop a need or a distraction?.>.>.>.

What will the impact of synthetic agents on social platforms like Reddit? Is the Myspace moment near? Same for research papers.>.>.

In other news, an interesting report on two decades of digital commons policies in the EU.>.

Hands-on research; add AIs to robots to let them understand the world.>.

Jevons paradox is a hot reference nowadays. And copyright issues: can generated data be copyrighted?.>.

More on robotics, immersive connectedness, and tech societies, below.

This week’s Triggered Thought

Sometimes, you feel that some notions should be combined, even if they do not seem directly related. This week, I had this with two things I attended, and both triggered thoughts and things to chew on. But I think they might have a connection somewhere too. So let me first briefly explain them separately.

Last Wednesday, I attended a night on Democracies in the age of AI. Super relevant, and I was happy to see the room was packed with engaged people. The first speaker was Payal Arora, and she had a slightly different take on the role of AI for humans. She presented a perspective on AI's role in the Global South. Arora, known for her work on "the next billion users," highlighted how young people in Global South might value AI differently compared to typical Western norms and fears. Her optimistic outlook, “From Pessimism to Promise,” suggests that AI could serve as an equalizer, empowering individuals who are historically marginalized or constrained by tradition. She divided it into three elements: (1) breaking down of digital divide; the leapfrogging skipping into mobile technology as a primary computing tool; are they more digital native? (2) algorithms of aspiration, not for oppression. (3) surveillance of care. Generative AI is used to create pictures that are not traceable to real people but express other things.

Then, this Monday, I attended the PhD defense of Inte Gloerich. Her topic: Is Blockchain a Truth Machine, or with the full title: Reimagining the Truth Machine - Blockchain Imaginaries between the Rational and the More-than-Rational. (PDF) “The dissertation examines how blockchain is often presented as a technology capable of producing a new, objective, and reliable record of truth, a "neutral" system free from human corruption. However, Gloerich argues that this view obscures the social, cultural, and political influences shaping technology's use and impact.”

On the surface you would read the two as contraries; the optimistic look of Arora is countered with the case of problematic tech worshipping that Gloerich is making. At the same time are both focusing on a more care-oriented relational-driven approach.

Chasing a truth machine is blocking relational-based practice. If we are looking for mutual understanding, then we need negotiation; we need to compromise, not absolute truth. We need to make machines for mutual understanding. What role can agents play? How do we preserve human agency? Agents of aspiration and care to connect it to Arora.

This might be an insightful paper this week

Gu, E. H., Chandrasegaran, S., & Lloyd, P. (2025). Synthetic users: insights from designers’ interactions with persona-based chatbots. Artificial Intelligence for Engineering Design, Analysis and Manufacturing, 39. https://doi.org/10.1017/S0890060424000283

In this paper, we introduce the term “Synthetic User”to describe such representations of personas that are informed by traditional data and augmented by synthetic data. We study the effect of one example of such a Synthetic User – embodied as a chatbot – on the designers’ process, outcome, and their perception of the persona using a between-subjects study comparing it to a traditional persona summary.

What are the plans for the coming week?

The International Film Festival Rotterdam is always in the agenda. Will be seeing five movies this week at least.

Students of The Hague University of Applied Sciences Master Next Level Engineering have been working on a concept for the Wijkbot and presenting this at their end expo in Delft on Thursday. At the same time, there is an interesting little event in Utrecht (and online) on “Reimagining Techno-Futures Through Creative Practice,” organized by Inte Gloerich, who just has had her PhD defense this Monday. And General Seminar is back; these events, organized by Julian Bleecker of Near Future Laboratory, are a lovely form to dive a bit deeper into topics with new people from around the world. I have to say that the tickets have gained a lot in price over the years…

And if you are into Key Value Indicators, this a possible (online) workshop for you.