tobychev @tobychev@qoto.org
Not sure what this account will be about, mostly boosting things I find interesting.
Joined Nov 2022
tobychev
boosted
Fun new game for art gallery visits: find all the portraits that could be valid UK passport photos
@petersuber
Isn't the big problem with the BigAI systems that they produce text with a form that implies there has been thorough vetting, but they just don't have algorithms to produce reliability scores in the first place? The most common symptom of this being the "hallucination" event, where the system extrapolates from its data without giving any warnings it is replying beyond the known edges of its knowledge.
Meanwhile humans can be fooled indeed, and might occasionally bluff even in their area of expertise, but very rarely will the expert be both fooled by a paper and willing to bluff about its contents at the same time.
@nikolar
En del är sant, men att ärkefienden kunnat döda den lokala påven och nästan helt fritt flugit runt och bombat hela landet för de ville det är inte direkt en enorm propaganda seger.
Att regimen skulle falla av ena bombningarna var aldrig troligt, även om det förstås går att sälja det som en bedrift att nationen överlevt: i stort har folket ingen koll på hur lite bombningar brukar åstadkomma så även det mest inkompetenta försvaret kan hävda stor framgång bara för att det är kvar.
tobychev
boosted
France is transitioning government desktops to Linux, with each ministry required to formalize its implementation plan by autumn 2026.
https://linuxiac.com/france-launches-government-linux-desktop-plan-as-windows-exit-begins/
tobychev
boosted
tobychev
boosted
OpenAI Backs Bill That Would Limit Liability for AI-Enabled Mass Deaths or Financial Disasters
https://www.wired.com/story/openai-backs-bill-exempt-ai-firms-model-harm-lawsuits/?utm_source=flipboard&utm_medium=activitypub
Posted into Business @business-WIRED
tobychev
boosted
The number of #elderly inmates in #Japan has quadrupled over the past 20 years. Around a fifth of people over 65 live below the #poverty line, struggling to make ends meet. Some go to extraordinary lengths to improve their circumstances, deliberately getting caught shoplifting in the hope of landing in prison. There, they can access a secure place to live, proper meals, and medical care – basic necessities often denied to them outside.
tobychev
boosted
Every euro spent by national governments is treated by commission as generating just €0.60 of economic activity no matter where it goes. Grid investment, education, tax cuts for the wealthy: all are logged as a net loss.
https://euobserver.com/210119/eu-fiscal-models-degrade-strong-case-for-clean-energy/
tobychev
boosted
For no reason at all I watched the film Threads the other day.
It's a good movie. Definitely not hopeful.
https://www.youtube.com/watch?v=IUmUz8ol9Ow
tobychev
boosted
At the same time, the “data-driven" nature of quantitative methods tends to be disjoint from morphological theory.
We took concepts known from 80s and re-conceptualised them for urban morphometrics. This leads to identification of "morphotopes" (as biotopes), the smallest morphologically homogeneous regions.
And because there's a lot of them, we have organised them into a taxonomy.
2/n
tobychev
boosted
In the first days of December, as it became clear that the ICE invasion was a real thing that was really happening to us, as groups of us gathered swapping rumors about the kidnappings and clearly inadequate tips about phone security, we had no idea what to expect, no idea what would happen, no idea what we were going to do. As much as we’d planned, heard from other cities, tried to be ready, we had no idea.
Only one thing was crystal clear: nobody, absolutely nobody, was coming to save us.
2/
tobychev
boosted
And because we believed that nobody was coming to save us, miraculously, the world shifted around us. Help came pouring in from everywhere. Suddenly, we were heroes?? That part still doesn’t sit well with me, the whole Nobel Prize thing, all of it. But one thing does sit well, very well: when the work fell to each of us, we all started doing it. All of us.
I’ve felt a lot of things during the ICE siege, but one thing I’ve never felt is alone.
9/
tobychev
boosted
to be clear - what does this one extremely simple feature reflect for the user?
In the "traditional way," the user can go to a website and see an example of how to do this themselves with information that's derived from the actual values used in computation. cool.
In the "LLM way," the user has no agency, can't see why the LLM might fail because they can't see the system prompt is directly feeding in the metadata about the available fields and thus has no idea that the model is capable of being wrong about its own fucking code and so they are shown wrong fucking values.
So the cost of transforming something to the "just prompt it" modality is "it being completely fucking wrong" even when that thing is literally just a feature that refers to the program state that is entirely owned by the fucking program" - to say nothing about how that pattern of development being recursively applied to the develpoment of the tool causes it to be fucking wrong as a matter of practice.
tobychev
boosted
California is "bending the duck curve" with batteries:
Grid batteries reach stunning new peak of 44 pct of evening demand in world’s fourth biggest economy
tobychev
boosted
reminder that anthropic ran (and is still running) an ENTIRE AD CAMPAIGN around "Claude code is written with claude code" and after the source was leaked that has got to be the funniest self-own in the history of advertising because OH BOY IT SHOWS.
it's hard to get across in microblogging format just how big of a dumpster fire this thing is, because what it "looks like" is "everything is done a dozen times in a dozen different ways, and everything is just sort of jammed in anywhere. to the degree there is any kind of coherent structure like 'tools' and 'agents' and whatnot, it's entirely undercut by how the entire rest of the code might have written in some special condition that completely changes how any such thing might work." I have read a lot of unrefined, straight from the LLM code, and Claude code is a masterclass in exactly what you get when you do that - an incomprehensible mess.
tobychev
boosted
Anthropic Issues Copyright Takedown Requests To Remove 8,000 Copies of Claude Code Source Code https://developers.slashdot.org/story/26/04/01/158240/anthropic-issues-copyright-takedown-requests-to-remove-8000-copies-of-claude-code-source-code?utm_source=rss1.0mainlinkanon
tobychev
boosted
A failed military police investigation?
The Afghan war crimes inquiry continues to throw light on alleged crimes by UK special forces and their possible cover-up
By me at Prospect
tobychev
boosted
I assumed most guys masturbated to Miss Piggy growing up in the late 1970s and early 1980s; having mentioned this to friends down the pub; it turns out I was very very mortifyingly wrong.
@grimalkina
Interesting episode Cat, but unfortunately I was quite confused most of the way through about Ashley's book situation. By the end I'm pretty confident she's already written one book and is working on a second, is that correct?
Also, did I hear it correctly that there was four different people speaking in this episode? Besides you two and the named man (whose name I've forgotten and gender I've assumed) at some point it sounded like a third woman spoke?
Mostly it's easy, but sometimes it's hard for me to distinguish if it's Cat or Ashley speaking and I wonder if it's something with the audio? I mean, during the first podcast episodes I listened to it was just hard because you didn't introduce yourselves by name in that episode, but since then I've listened to several hours of Cat's YouTube presentations so I hope I've gotten better at it. But maybe you two just sound similar sometimes?
@analog_ashley
Not sure what this account will be about, mostly boosting things I find interesting.
Joined Nov 2022
