@cosmos4u@scicomm.xyzA tiny hymenopteran - I think - extremely interested in sage plants on our balcony in Bochum, Germany. The size of a common fly and very hectic so the pics didn't come out too well, but perhaps good enough to ID this #Insect.
@Mediagazer@mstdn.socialWhen asked what it means to be a good news consumer, 20% of US adults mention being discerning, 17% following the news, 13% getting news from quality sources (Elisa Shearer/Pew Research Center)
https://www.pewresearch.org/short-reads/20
@rasterweb@mastodon.social➡️ Common Dreams - News for the Progressive Community
#bookmarks
@cosmos4u@scicomm.xyzGeologists on the silver screen—the sequel: #movies, too ...
@detondev@social.linux.pizzaThe most common three wise monkeys variations intrigue me. The positively framed versions are cute, but subdued ominous aura is part of the appeal. The ones with "do no evil" where the monkey covers genitals have too much ominous aura for the mainstream. This could be a Goldilocks situation
@steve@s.yelvington.comWatching the Cape Fear miniseries. Good move keeping the Bernard Hermann score. Not so good on the accents. The actors are struggling. A common problem.
@yetiinabox@todon.nlLovely comparative research on human cooperation by Andre, Boneva, Chopra, and Falk.
They studied 101,123 people from 125 countries - actual people, not models.
"Across the globe, cooperation is strongly associated with beliefs about others’ cooperation, approval of social norms, and preferences for the common good; altruism, patience, and risk attitudes also positively predict cooperative behavior."
People cooperated even though they doubted that others would: 69…
@davidaugust@mastodon.online
@arXiv_eessAS_bot@mastoxiv.pageToo Good to Be True: A Study on Modern Automatic Speech Recognition for the Evaluation of Speech Enhancement
Danilo de Oliveira, Tal Peer, Timo Gerkmann
https://arxiv.org/abs/2605.12107 https://arxiv.org/pdf/2605.12107 https://arxiv.org/html/2605.12107
arXiv:2605.12107v1 Announce Type: new
Abstract: Speech enhancement (SE) systems are typically evaluated using a variety of instrumental metrics. The use of automatic speech recognition (ASR) systems to evaluate SE performance is common in literature, usually in terms of word error rate (WER). However, WER scores depend heavily on the choice of ASR system and text normalization pipeline. In this paper, we investigate how modern ASR models correlate with human recognition of enhanced speech. A listening experiment reveals that modern ASR models with large-scale noisy training and embedded language models correlate more with human WER than simpler ones, with a transducer model providing the most reliable transcriptions. Nevertheless, we also show that these models' robustness to noise and use of context can be uninformative to an acoustics-focused evaluation of enhancement performance.
toXiv_bot_toot
@hex@kolektiva.socialI tried to write something up to talk about an idea, but it didn't quite work. I have a lot more I need to put into it. But I want to get an idea out, and, after talking with a person who pointed out some of the flaws in what I wrote, I think I can maybe write down the kernel of the idea here.
An acquaintance of mine did a deep dive on Operational Art and wrote his thesis (which prompted an earlier set of posts and an article I wrote for my professional-ish blog) on the intersection of the OODA loop and critical philosophy. I've been spending a lot of time thinking about Kilcullen's Three Pillars model (after watching Andrewism's wonderful video) and Beer's VSM. The TL;DR of it is that there's a much better insurgency model. Of course, the insurgency model also works for a bunch of other things, because cybernetics lets you do all kinds of cool abstraction like that.
So as I was reading the essay of a comrade the other day, that model popped back into my head and I'm going to try to share what I can of it.
When colonizers came to the Salish region, they saw what they believed to be an untouched wilderness. They failed to see the ways in which Salish people tended the land. Indigenous fire practices were common on the northwest coast, and the suppression of those practices remains a problem. There is an interrelationship between an environment and the systems within it. Systems, like people, animals, and cultures, adapt to the environment. In doing so, those systems will also change the environment.
Social technology was invisible, so colonizers defaulted to either some kind of Rousseauvian or Malthusian model of these people. They were not, for the colonizers, people who had developed advanced social technologies to live in harmony with their world. They were, rather, people in "a state of nature."
The European influenced left continues to draw this Rousseauvian model, which continues through a lot of Anarchist revolutionary thought. European anarchists were heavily influenced by observations and theories around the behavior of indigenous people. The remnants of this thought still exist in the idea that the system must only be destroyed for us to be free.
This is the same obliviousness to social technology, that social technology actually exists, often informs both early colonizers and modern radicals.
It is through this obliviousness that we fail to recognize how capitalism is a social technology that is managed into existence and maintained, and how changes in the environment can threaten institutions that have become over-adapted to a specific version of that environment.
We can extend Kilcullen's metaphor of a "conflict ecosystem" through cybernetics into a much more rich model, populated by viable systems. The ecosystem itself has a fitness function, which drives adaptation within the environment. But all actors in the environment also affect it. Some try to manage the environment. Revolutions are often over who manages a social ecosystem, over who controls the social technology and what it does.
Once we see this dynamic at play, calls of "riot" and "revolution" make a whole lot less sense. Rather, the question becomes, "how do we change the ecosystem in such a way that it cannot be 'managed' at all?"
Graeber/Wengrow talked about Turtle Island indigenous social technologies in Dawn of Everything, such as the system of moieties and clans described in the book. So I have a good reading list as I think through this model, but I hope the "ecosystem" model is helpful (if not completely fleshed out).
I'd be interested in any critiques or thoughts to help develop this idea more.
@mgorny@social.treehouse.systemsNew on #blog (this time with quotes from Fedi): "Why Gentoo?"
#Gentoo is perceived by the wider public, the non-users. What probably stands out most is compiling. Almost everyone who heard of Gentoo knows it has something to do with compiling everything. And why are we doing that? Well, besides being hardcore, the common sentiment goes for performance. So yeah, Gentoo users must be some kind of hardcore ricers who try to squeeze every last bit of their system performance.
To be honest, I don’t think that’s a good way to describe Gentoo. Yes, compiling is at the core of it. But performance? I don’t think so, at least not in the obvious, -O9999 -fzomg-fast way. The world has moved on, CPUs have gotten faster, optimizations have gotten smarter, and distributions have started optimizing more aggressively. Optimization-wise, I suspect your average Ubuntu package with generic optimizations may be no slower than the equivalent Gentoo package fine-tuned for your CPU. And if it’s not, then it probably won’t make a real difference anyway.
There’s much more to Gentoo than that. Yes, some of it comes from building from source: the flexibility. But a lot of it comes from the wider Gentoo philosophy, the philosophy that brought us all together. The idea that Gentoo is the distribution we’re making for ourselves and people who enjoy Gentoo. So if I were to make a few arguments for Gentoo, I’d focus on that. And this is what I’d like to do here.
"""
@arXiv_csGT_bot@mastoxiv.pageSimultaneous EF1 and approximate MMS allocations for submodular valuations
Uriel Feige, Assaf Fine
https://arxiv.org/abs/2606.06451 https://arxiv.org/pdf/2606.06451 https://arxiv.org/html/2606.06451
arXiv:2606.06451v1 Announce Type: new
Abstract: There are two common classes of fairness notions that are considered when allocating $m$ indivisible items to $n$ agents of equal entitlements. One is that of share-based fairness notions, with the maximin share (MMS) and its relaxations to $\rho$-MMS being prominent representatives of this class. The other is that of comparison-based fairness notions, with envy-freeness (EF) and its relaxations such as EF1 being prominent representatives of this class. In general, no class offers good guarantees for the other class. In this work, we design allocations that simultaneously satisfy notions from both classes, and specifically, are $\rho$-MMS for constant $\rho$ and EF1 (in fact, also EFL). Such results were previously known when agents have additive valuations, and we prove such results for the more general class of submodular valuations.
toXiv_bot_toot
@relcfp@mastodon.socialFaith and AI Postdoctoral Fellowship
https://ift.tt/psQNHWF
CALL FOR PAPERS: Modernizing Winter Sam White 05/20/2026 - 9:31pm Announcement Type …
via Input 4 RELCFP https://
@PaulWermer@sfba.social @paulwermer@sfba.social
@joe@toot.worksThat's good to hear but I am not sure why that wasn't already the case.
"The Milwaukee Common Council approved legislation that modifies Milwaukee Police Department standard operating procedures governing officers' duty to intervene in and report unreasonable uses of force, including force used by federal agents."
#Milwaukee #ICE #Police
@thomasrenkert@hcommons.socialAre there any good projects out there using #frugal #AI in #lowtech scenarios? I am interested in #solarpunk -esque examples where AI (local Models) is being used for the common good.
