Tootfinder

Got the latest ollama running with image generation models now.
Used models are x/flux2-klein:latest and x/z-image-turbo
took about 1min 20s to generate these images on a M2 Max CPU with 64GB RAM
#ollama #GenAI

This could be long. It's hard to express my thoughts on this.
I had a bit of a heart to heart with my son today. My son is a student at Minneapolis College of Art and Design. He's 26. I'm white, his mother is third generation Mexican-American. To me, my kids are just my kids, I see myself and my ex-wife both in them. But he has been pushing me to realize for some time that isn't how the rest of the people around us see them. They don't see me in them at all, they ju…

Zelensky meets Lithuania's president as country pledges nearly 100 generators for Ukraine amid Russian strikes: https://benborges.xyz/2026/01/25/zelensky-meets-lithuanias-president-as.html

Microsoft Open-Source Programs Office Director Stacy Haffner, and Developer Community VP Scott Hanselman announced today in the wake of Microsoft's acquisition of Activision, that they have just released the source code for the games Zork 1-3 under a MIT open-source license.
"Preserving code that shaped generations: Zork I, II, and III go Open Source"

I think it's time for some #FollowFriday action once again. Today, let's focus on computational/algorithmic/generative art/design practitioners (not AI art :)
An entirely incomplete list, in A-Z order:
@…

I was sent a link to slowinternet.info and it's interesting how off-putting the images are to me. I struggle to take in the message because of how angered I am by the AI-generated pictures, which are clearly mashed up using art from Simon Stålenhag.

The 2026 Ockham New Zealand Book Award disqualified two books citing AI-made covers; their publisher says AI's rise has put the industry in "uncharted waters" (Jin Yu Young/New York Times)
https://www.nytimes.com/2025/11/19/books/new-zealand-ocham…

Weighted Stochastic Differential Equation to Implement Wasserstein-Fisher-Rao Gradient Flow
Herlock Rahimi
https://arxiv.org/abs/2512.17878 https://arxiv.org/pdf/2512.17878 https://arxiv.org/html/2512.17878
arXiv:2512.17878v1 Announce Type: new
Abstract: Score-based diffusion models currently constitute the state of the art in continuous generative modeling. These methods are typically formulated via overdamped or underdamped Ornstein--Uhlenbeck-type stochastic differential equations, in which sampling is driven by a combination of deterministic drift and Brownian diffusion, resulting in continuous particle trajectories in the ambient space. While such dynamics enjoy exponential convergence guarantees for strongly log-concave target distributions, it is well known that their mixing rates deteriorate exponentially in the presence of nonconvex or multimodal landscapes, such as double-well potentials. Since many practical generative modeling tasks involve highly non-log-concave target distributions, considerable recent effort has been devoted to developing sampling schemes that improve exploration beyond classical diffusion dynamics.
A promising line of work leverages tools from information geometry to augment diffusion-based samplers with controlled mass reweighting mechanisms. This perspective leads naturally to Wasserstein--Fisher--Rao (WFR) geometries, which couple transport in the sample space with vertical (reaction) dynamics on the space of probability measures. In this work, we formulate such reweighting mechanisms through the introduction of explicit correction terms and show how they can be implemented via weighted stochastic differential equations using the Feynman--Kac representation. Our study provides a preliminary but rigorous investigation of WFR-based sampling dynamics, and aims to clarify their geometric and operator-theoretic structure as a foundation for future theoretical and algorithmic developments.
toXiv_bot_toot

Google says Gemini 3 Pro scores 1,501 on LMArena, above 2.5 Pro, and demonstrates PhD-level reasoning with top scores on Humanity's Last Exam and GPQA Diamond (Abner Li/9to5Google)
https://9to5google.com/2025/11/18/gemini-3-launch/

Imagine ChatGPT but instead of predicting text it just linked you to the to 3 documents most-influential on the probabilities that would have been used to predict that text.
Could even generate some info about which parts of each would have been combined how.
There would still be issues with how training data is sourced and filtered, but these could be solved by crawling normally respecting robots.txt and by paying filterers a fair wage with a more relaxed work schedule and mental health support.
The energy issues are mainly about wild future investment and wasteful query spam, not optimized present-day per-query usage.
Is this "just search?"
Yes, but it would have some advantages for a lot of use cases, mainly in synthesizing results across multiple documents and in leveraging a language model more fully to find relevant stuff.
When we talk about the harms of current corporate LLMs, the opportunity cost of NOT building things like this is part of that.
The equivalent for art would have been so amazing too! "Here are some artists that can do what you want, with examples pulled from their portfolios."
It would be a really cool coding assistant that I'd actually encourage my students to use (with some guidelines).
#AI #GenAI #LLMs