Hi all! Hope you have a good week.
I had my official goodbye moment at INFO. Thanks Jann for the nice words and all former colleagues of course too. As tradition it is expected to say some words as leaving employee. So I wrote some things and thought I would have Lex, the generative AI, to complete a part, or, better, give some suggestions. I did not expect it to make a complete description of INFO’s core promises. I did like this framing: Squeezing that in one sentence would be like this: INFO opens up clients, to create the potential for growth, the INFO way…
It triggered me in the rest of the speech writing as intended by the tool. Especially the “INFO opens up clients” I liked as it connects to the mutual relationship INFO like to have with her clients; not only serving the best but also aiming for real partnership.
The next day on Friday, I visited a little symposium in Eindhoven on “AI for All; From the Dark Side to the Light”. Bruce Sterling did a good job showing how the current generative AI is thinking and what its capabilities are. Using a hand as a prompting case, it became clear that these tools do not really understand the concept of a hand but try to create images based on text similarities. Resulting in strange representations of hands. It was not only funny, it made clear that we should not trust too much in these tools, or better said: value them for what they are: “automated statistics”. An excellent way to explore more is to feed the tools with concepts like “the unimaginable, the undreamed-of, etc.
Marleen Stikker addresses some counter AI arguments, or better said, some in awareness of how to appreciate the AI as a decision maker; the AI is not alive, is living without consequences, keep that in mind. “Quantification of nature is a false representation as humans are already contextualising “reality””
In everything, the talks made clear that designing the right collaboration between humans and AI is key (the co-performance). In the latest blog of Tobias Revell the showcases of Dcode project were mentioned, which are a good inspiration for these types of considerations.
For the coming week, these are some possible events on the agenda
- 9 December, Rotterdam: ThingsCon Winter Unconference
- 1 December, Amsterdam: Public Spaces Mastodon Fediverse Proeverij
- 30 November, online: The Hmm in Gaming Spaces
- 29 November, Amsterdam: Lichtkogel experience
- 28 Nov – 2 Dec, Rotterdam: Immersive Tech Week
And more news bits from last week:
|A simpler path to better computer vision|
REALITY MAPPING – “Research finds using a large collection of simple, un-curated synthetic image generation programs to pretrain a computer vision model for image classification yields greater accuracy than employing other pretraining methods that are more costly and time-consuming, and less scalable.”
|Stable Diffusion 2.0 Release |
GENERATIVE AI – The open source text to image generator has a new release, with some upscaling and some new features like depth2img.“Depth2img infers the depth of an input image (using an existing model), and then generates new images using both the text and depth information.”
|Meta researchers create AI that masters Diplomacy, tricking human players |
HUMAN AI – Games are a thankful measurement tool for the human likeness of AI: “Meta AI announced the development of Cicero, which it claims is the first AI to achieve human-level performance in the strategic board game Diplomacy. It’s a notable achievement because the game requires deep interpersonal negotiation skills, which implies that Cicero has obtained a certain mastery of language necessary to win the game.”
|New ultra-thin solar cells could be the future of space power|
SOLARPUNK – “Ultra-thin solar cells could extend the operational lifetimes of satellites while also making missions less costly and more efficient.”
|Japan to open roads to autonomous vehicles in 2023|
AUTONOMOUS – “Japan’s National Policy Agency announced that it would lift a ban preventing SAE Level 4 autonomous vehicles from operating on Japanese roads.”
|Drones on strings could puppeteer people in virtual reality|
VREALITY – “Having drones on strings attached to a person could provide a more realistic simulation of interacting with physical objects in virtual reality”
|CityStream Live – The Real-Time Mapping Platform|
TWINNING – Waze became famous as the mapping route planning app, and other added driving-tracking traffic engines. Now there is real-time mapping too: “Only with Real-Time Mapping can vehicles with advanced driver assistance know what’s coming their way, react to varying speed limits, find parking, avoid work zones and one day – drive themselves.”
|Honda’s new humanoid-robot push aims for practical use|
HUMANOIDS – “A decade from now, a man suddenly collapses on the street. An ambulance is called but will take time to arrive. In the meantime, a robot stationed nearby rushes to the scene to provide emergency care, controlled by an expert operator sitting dozens of kilometers away.”The vision of Honda on humanoids in our daily life. I had to think that we have in a certain way the first iteration with the heart machines that guide you through the process step-by-step…
|RoboSense launches flash solid-state LiDAR |
AUTONOMOUS – Components are important for development. In this case, making LiDAR more applicable can influence the sensemaking of autonomous vehicles. “(…) E1 greatly streamlines the circuit design and production processes, creating the performance and cost advantages necessary for the durability and reliability requirements of blind spot LiDARs in the automotive market.”
|CES 2023 robotics Innovation Award winners announced |
ROBOTICS – These winners of the 2023 CES Innovation award give an impression of the current (possible) application of robots in our life.
|6 New Theories About AI – Napkin Math|
AI FUTURES – “(1) Just as the internet pushed distribution costs to zero, AI will push creation costs toward zero. (2) The economic value from AI will not be distributed linearly along the value chain but will instead be subject to rapid consolidation and power law outcomes among infrastructure players and end-point applications. I still hold these frameworks to be true—now I’d like to apply them to six “micro-theories” on how AI will play out.“
|Netflix Presents ‘Kaleidoscope’ Non-Linear Streaming Experience |
NEW NARRATIVES – “Netflix’s Heist Series ‘Kaleidoscope’ Can Be Streamed in Any Order: A non-sequential puzzle.”Non-linear and interactive storytelling has been around for a while. I can remember one from the famous dystopian series Black Mirror a couple of years ago, and there were many more in art. I am curious if this delivers a new type; having different angles to a story might be another take, like the Spotify docu “The Playlist” did in a scripted version.
|At a China Covid Protest, a Mix of Giddy Elation and Anxiety|
GEO – This news on the increasing riots in China made me think back to the Ai WeiWei documentary on Hong Kong protests from 2020 Cockroach, which is not a promising scene…
Paper for this week
In a sense, this paper from 2018 connects to the thoughts on generative AI tools: Situating methods in the magic of Big Data and AI
“Big Data” and “artificial intelligence” have captured the public imagination and are profoundly shaping social, economic, and political spheres. Through an interrogation of the histories, perceptions, and practices that shape these technologies, we problematize the myths that animate the supposed “magic” of these systems.
Elish, M. C., & Boyd, D. (2018). Situating methods in the magic of Big Data and AI. Communication monographs, 85(1), 57-80.
See you next week!