Bridging Resourcefulness and Semantic Understanding

Weeknotes 331 - New LCMs, semantic AI vaporware, and triggering human resourcefulness. What is happening in our relation with AI? This and a look back at the news of last week.

Bridging Resourcefulness and Semantic Understanding
Interpretation by Midjourney

Hi all!

What did happen last week?

Let me see, what is worth mentioning here? I made draft proposals to share with consortium partners, finalized the Civic Protocol Economy report, had meetings with nice people, and worked on the Wijkbot Kit (aka Robot perspective). Also, we sent out the responses to the proposals for the ThingsCon RIOT publication after evaluating, and reach out to all exhibitors for the next Salon and exhibition on Generative Things. This week, I will update the website with all the details! It was nice to discover the mention in Handboek Publiekwijs. And the op-ed of Laurens mentioned someone that could be me :)

Yesterday I attended a session on futuring in systemic co-design. What are the models to jump out reality? Embedding it in organisational interplays with the future, is one of the conclusions. First questions were asked, needs for deeper exploring.

What did I notice from the news last week?

The long list was quite long this week, longer than usual.

  • There was a message earlier about the delay of real intelligent features in Apple Intelligence. John Gruber wrote an opinion piece that got traction. Hope Apple will get things right on time.
  • Google was impressed with Gemma 3 and got a lot of attention for its robot AI. And backlash on the fact Gemini can see your search history.
  • Amazon is just capturing everything you say. Echoing becoming real.
  • SXSW Interactive has ended, and less than in earlier years, I felt an overall vibe. The lack of political critiques was remarkable and worrisome. Is it indifference or fear?
  • Community notes are now at Meta. A sign of the times.
  • There seems to be a fork in the road regarding whether to focus AI on consumer-facing media or vibe coding for personal computing. I expect vibe coding to stay hot in the coming weeks.
  • Manus still got attention as people started to use it more and more.
  • How will this become Public Intelligence?
  • And much more, see below.

What triggered my thoughts?

In the rapidly evolving landscape of artificial intelligence, we find ourselves at a crossroads where human ingenuity meets machine capability. It was the topic here before, but I was triggered by a piece on the next generation of LCM (Large Concept Models), and the mentioning of Semantic AI in a piece on Apple Intelligence vaporware. It made me think about another project way back, on Resourceful Aging.

Beyond Language Models: The Promise of Semantic Understanding

The AI community appears to be shifting from Large Language Models (LLMs) toward what some are calling Large Concept Models. This transition represents signals a fundamental change in approach. While LLMs excel at generating creative language patterns, LCMs promise something more profound—an understanding of the world's semantic structure rather than merely manipulating words and phrases.

This evolution reminds me of Apple's unfulfilled promise of "Semantic AI" at last year's developer conference. Positioned as a cornerstone of Apple Intelligence, this semantic capability has largely failed to materialize. This echoes earlier conversations around Web3, back in the years before web3 was about tokens and DAOs, which was envisioned as a "semantic web"—a network that understood meaning rather than just connecting data points.

Human Resourcefulness: Our Unique Advantage

Parallel to these technological developments, I've been reflecting on human resourcefulness—our remarkable ability to repurpose, reimagine, and adapt our surroundings to serve new functions. Seven years ago, the project "Resourceful Aging" explored this quality, specifically among elderly individuals with limited physical capabilities. The researchers observed how seniors ingeniously repurposed everyday objects to overcome challenges, then attempted to capture these insights using AI to inform design solutions.

What makes this approach particularly compelling is that it didn't attempt to replace human resourcefulness but rather to learn from it. The AI served as a bridge between ingenious human solutions and scalable design implementation.

Machine Resourcefulness as a Catalyst

What if we reconsidered our understanding of resourcefulness itself? While we typically view resourcefulness as an inherently human trait, emerging AI systems display their own forms of adaptive problem-solving that, while different from human approaches, can be remarkably effective. These machine-generated solutions often follow unexpected patterns—combinations and applications that human minds might never consider due to our inherent biases and cognitive limitations.

This machine resourcefulness, when presented to humans, can serve as a catalyst for our own creative thinking. When an AI suggests an unconventional approach or makes connections between seemingly unrelated domains, it can jolt us out of established patterns and trigger new pathways of thought. The machine doesn't need to perfectly mimic human understanding to be valuable; its very otherness becomes an asset.

In our complex, constantly changing world, the most exciting frontier may not be machines that understand us perfectly, but machines that can surprise us productively. Large Concept Models might eventually offer both semantic understanding and a unique form of machine resourcefulness that complements and stimulates our human creativity. Rather than merely serving as tools that extend our capabilities, these systems could become genuine thought partners that challenge us to see problems differently.

The future of human-AI interaction may thus be less about partnership in the conventional sense and more about creative provocation—AI systems that don't just solve our problems but help us reimagine them entirely. As machines develop their own brand of resourcefulness, they might become not just amplifiers of human thinking but transformative influences that expand the boundaries of what we consider possible. This is in the end what vibe coding is about.

What inspiring paper to share?

Felienne Hermans had a nice piece in a Dutch newspaper on techbro-culture. She referenced an earlier paper of hers on the topic.

In this paper, we present feminism as a philosophical lens for analyzing the programming languages field in order to help us understand and answer the motivating questions above. By using a feminist lens, we are able to explore how the dominant intellectual and cultural norms have both shaped and constrained programming languages.

Felienne Hermans and Ari Schlesinger. 2024. A Case for Feminism in Programming Language Design. In Proceedings of the 2024 ACM SIGPLAN International Symposium on New Ideas, New Paradigms, and Reflections on Programming and Software (Onward! '24). Association for Computing Machinery, New York, NY, USA, 205–222. https://doi.org/10.1145/3689492.3689809

What are the plans for the coming week?

This week I will have a similar week as last week, with time to update some communications on Cities of Things, writing proposals and catching up with people. I might check a meetup on sustainability on Service Design on Wednesday if the waiting list is shortened. The talk on Probing Futures - co-designing change for tomorrow at the same timeslot is interesting too, but for me similar to the Esc event on Monday.

I intend to check the demo against fascism/racism in Amsterdam this Saturday.

See you next week!

References with the notions

Human-AI partnerships

AI, as in Apple Intelligence, is talk of town. Delays on the integrated AI promises.

Daring Fireball: Archive
Apple innovation and execution — Benedict Evans
It matters that Apple’s new Siri will be late, and it matters more that Apple didn’t realise. Is it more than that?

Google has new things with Gemini visual tools and smaller models. Deep Search has a different meaning, looking into your history. Still, AI search is unreliable.

Google calls Gemma 3 the most powerful AI model you can run on one GPU
If you only have one Nvidia H100 available.
Google’s Gemini AI can now see your search history
Gemini can do more with your data if you let it.
AI search engines cite incorrect sources at an alarming 60% rate, study says
CJR study shows AI search services misinform users and ignore publisher exclusion requests.
Google is officially dumping Assistant for Gemini
RIP.

Agentic and other AI upgrades of existing services with Zoom, OpenAI, Snap,

No one knows what the hell an AI agent is | TechCrunch
AI agents are all the rage. But no one knows exactly what an agent is, partly because companies define them radically differently.
Coming into focus: the State of the AI Agent Ecosystem in March 2025
OpenAI dropped an agent release today, Claude’s MCP is going nuts, and I figured it was time for an overview of players in the AI agent ecosystem—who’s hot, who’s not, and what do we build with them?

Kevin Kelly used to have a world brain as model, and adds now the concept of Public Intelligence.

Public Intelligence
Imagine 50 years from now a Public Intelligence that was a distributed, open-source, non-commercial artificial intelligence, operated like the internet, and available to the whole world. This public AI would be a federated system, not owned by any one entity, … Continue reading →

Is our mental capacity eroding? A nice topic for the next bd party.

Experts question if our mental capacity is eroding
Researchers are increasingly worried that humans may have reached their cognitive peak.

Vibe coding this week. Is that contributing to the Jevons Paradox after all?

Speaking things into existence
Expertise in a vibe-filled world of work
Jevons Paradox: A personal perspective
What if the most game-optimal play in the new system is actually to become relentlessly, unapologetically you?
The Vibe Coding Bible: How to Build Useful Things with Short Prompts
Vibe Coding is all the rage since Andrej coined the term a month ago, but what does it really mean? How can you build useful things with it? What are the principles and pitfalls? Is it just a gimmick?
I Tried AI Coding Tools. Now I Want to Learn to Code.
Here’s what they don’t tell you about vibe-coding tools: They’re gateway drugs.

Robotic performances

Googles has a Robotic AI. Is a ChatGPT moment for robotics near?

When will we get the ChatGPT of robotics? The future of embodied AI is bright - The Robot Report
Generative and embodied AI may lead to general-purpose robots, but developers must avoid speed bumps, says Converge.
Google’s new robot AI can fold delicate origami, close zipper bags without damage
Google Gemini Robotics AI model gives robots fine motor skills and adaptability for general world use.

Issues with the home robotics OG

These new Roombas aren’t the robots I know and love
These may no longer be the robots you are looking for

New moves.

Driverless ‘bus of the future’ is tested in Barcelona
Renault is testing a new driverless mini-bus in Barcelona this week. The autonomous vehicle is running on a circular route of just under 1.5 miles, with four stops in the center of the Spanish city.

Immersive connectedness

Starlink is triggering competition

Alphabet spins off Starlink competitor Taara
Taking the ‘fiber’ out of ‘fiber optics’.

Immersive connectedness should be extended with immersive intelligence.

China’s AI frenzy: DeepSeek is already everywhere — cars, phones, even hospitals
Chinese companies, as a signal of patriotism, are racing to build services on top of the homegrown AI model that has taken the world by storm since January.

New futures, new applications

Virtual reality goggles helped me on the path to physical recovery
An unexpected bit of tech opened up a whole new world of possibilities for a frustrated, injured writer

This will be the immersive aesthetics

refik anadol turns frank gehry’s forms into AI-generated landscapes at guggenheim bilbao
refik anadol presents living architecture: gehry, an audiovisual installation at the guggenheim museum bilbao until october 19th, 2025.

Tech societies

The new black is agentic and reasoning models combined.

🔮 The age of reason(ing models)
Learning, agents and the $ trillion opportunity

Community notes are the excuses for taking responsibility

How Meta’s take on Community Notes misses the mark
Crowdsourced moderation is better than nothing — but how much better remains an open question

In the end it comes down to physics.

The tiny chips behind Amazon’s big AI investment
Anthropic prepares to put the company’s Project Rainier, potentially the world’s most powerful computer, to the test.

Bigger ownership politics. Web3 is far away.

OpenAI and Google ask the government to let them train AI on content they don’t own
OpenAI says it’s a “matter of national security.”
Everything you say to your Echo will be sent to Amazon starting on March 28
Amazon is killing a privacy feature to bolster Alexa+, the new subscription assistant.
OpenAI and Google’s dark new campaign to dismantle artists’ protections
OpenAI and Google have cozied up to Trump’s imperialist approach to AI for a shot at tearing down artists’ copyright protections.
Buy Me a Coffee at ko-fi.com