Weeknotes 324 - thinking-out-loud AIs as learning value

By Iskander Smit in weeknotes — Jan 28, 2025

Reflecting on the volatile market of GenAI and other aspects of DeepSeek-impact, and thinking about the impact on human value. With lots of links as always.

Interpretation by Midjourney, prompted by DeepSeek from the triggered thought

Hi all!

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.

What happened last week?____Another iteration in the research report on Civic Protocol Economy was the primary goal, but it was also busy with other things that are nice to mention. I visited student projects in Delft, both from Industrial Design Engineering and The Hague University of Applied Sciences, who are working on a Wijkbot project. Together with Lisa, I discussed the final work for the WijkbotKit towards a Robo-perspectives systemic co-design kit. I was also able to attend a workshop organized by Speculative Futures The Hague in Rotterdam, envisioning the future of fully automated digital services, and in the CoECI event, I attended a workshop called “Partnercipatie, conflicting interests, and collective ownership”. Two articles/papers related to Wijkbot methodology were submitted; thanks to the main authors, I could be a co-author.

What did I notice last week?____DeepSeek was the talk of the town last week with tech reviewers, setting new standards, so it seems in costs mainly, and in how it is sharing its thinking process with the human partner in crime. See below. It was only this week that the stock markets started to shake up, as all the chip manufacturing-related companies were falling, the projections needed reconsidering, and investors were getting nervous. It is a volatile market, that is for sure. GenAI markets could get wild… How much is this part of the geopolitical ‘war’ taking place? The well-timed introduction is a good instrument for making markets nervous. > > >

By the way, this is an old interview with the CEO of DeepSeek from some months ago that points out the complexity and promise of operating in the rapidly maturing data-intelligence sector that China is.

Next to DeepSeek, the new operator functions of OpenAI dominated the news. Although quite an early implementation, they signal our relationship with AI, just like other agent-like behavior. Below, see how to relate to this experience of more reasoning AI thoughts. > >

In the aftermath of the dark day of 20 January, the new US government is taking some shine for a new program for developing AGI on US soil: Stargate. In the light of DeepSeek this can be seen as overspending, and what are the real interests of Big Tech to fight China via the connection with the Trump presidency… Also; what does this deal mean for Microsoft and OpenAI? > >

It is also part of the discourse on the speed of developing AGI or ASI; Anthropic took the stage to predict as early as 2027… >

In other AI news, Google launched in partnership with Samsung, a new iteration of the mobile assistant. Perplexity is launching an Android-only assistant, too. Will Apple Intelligence speed up fast enough before assistants are influencing platform choice? If you read the experiences of John Gruber, it feels long away. > > > >

And the real change will be from quantum, MIT thinks. >

Robots might intrigue the skeptics this year, MIT predicts. >

A short but rich piece on autonomy, design, and AI by Kars: “The buyers and deployers of AI could and should be made more accountable to the people subjected to AI.” >

I used to like my Pebble back in the days, a pioneering smartwatch. It seems it might be making a comeback. >

Batteries based on sugar make sense. >

More links below, like new functions or fun things, robotics, and immersiveness. And struggling policymakers for AI. Are other AIs possible?

And a service part: Which AI to use now, an updated opinionated guide by Ethan Moillick >

Triggered thought ____No doubt developments go fast. With AI starting to work with us as smart colleagues, that is. And with AI taking over some typical jobs. Browsing the news (it should be possible for an assistant to make a preselection of the articles worth sharing and start a conversation about the ones it is not sure about. In the time I save, I can spend more time on this part of the newsletter, which is more about my personal connection with a topic and the links that I have. But, well, I am distracting myself from the key point that I like to make this week. Or maybe better said, two key points to combine.

Two podcasts of the last week(s) address a more fundamental shift initiated by AI; the change in human value. Indy Johar makes it very sharp in a podcast; we already see through the increase of people living on the street that certain types of functions make less sense to let perform by humans. In POKI (Dutch podcast), one of the hosts was also rather impressed by the speed of developments, among others, through the new DeepSeek introduction. He was especially impressed by the new form of reasoning that shows how AI thinks. Even though the agent functions of OpenAI (Operator) are not yet impressive, the direction is clear.

Agents are hot; that is already the returning belief. This thinking-out-loud AI that has been popping up in the last weeks is part of this development. Thinking AI, AI-operator, our relationship with AI is also constantly shifting. Do we want to have an AI taking over our computer? And if we delegate tasks to our AI agent, is there a need to mimic mouse-clicking? It feels rather silly to let two digital AI systems have a conversation by mouseclicking interfaces. Still, it might be necessary for the process of acceptance and feeling of agency to contribute to the rapid change that will be made. That is what the thinking-out-load AI is performing, and it delivers not only potential better results in the first order, but is also shaping a new order.

Paper for the week___Extra relevant in times of designing transformations In this article, I examine the foundations of design knowledge and how they have been disrupted as the design discipline moves progressively away from industrial production. (…) I propose a shift in the traditional principles of designing, moving away from the idea of perfect solutions and toward learning systems that are good enough for now.

Mortati, M. (2022). New design knowledge and the fifth order of design. Design issues, 38(4), 21-34.

What are the plans for the coming week?____Some interesting events this week, that I probably have to skip mostly. On Wednesday “Taking back society” in A-Lab Amsterdam, and DE/MO democracies in the age of AI at Pakhuis de Zwijger (in Amsterdam too). On Thursday Adam Greenfield is presenting his latest book.

See you next week!

References with the notions

Human-AI partnerships

Robotic performances

Immersive connectedness

Tech societies