tobychev @tobychev@qoto.org
Not sure what this account will be about, mostly boosting things I find interesting.
Joined Nov 2022
tobychev
boosted
Trump rage-quite Nato, again, did mock impressions of Macron, insulted Saudi Arabia, and hand-mimed US missile strikes on Tehran.
https://euobserver.com/208880/breaking-scoop-us-diplomacy-hits-rock-bottom-again/
tobychev
boosted
Let's make the "empowerment" claim, that is somewhat true, here: it's true that most people experience computers as black boxes, and for a long time it has been the case that only the "hacker class" has access to "sieze the means of computation", as it were.
I think the "everyone can instruct their computer to do the kinds of things they would *like* their computer to help them with" is actually a good goal.
That said:
- Current tools appear to do so, but there are so many gotchas in terms of how they operate, and when things go badly, users are stuck.
- The approach is incomplete. I believe to successfully achieve the goal requires a *combination* of machine learning and constraint solver techniques. Without this, contemporary AI tech tends towards high amounts of dangerous failure cases.
- We need to be very concerned about deskilling and where expertise develops from.
- There are a lot of opportunities to improve that *aren't just LLMs*!
- AI industry power dynamics are very bad.
tobychev
boosted
I think something that's worth highlighting is that both communities are concerned with empowerment and disempowerment. And I tend to think these tools tend to *appear* empowering, but are actually disempowering, in their current configuration.
I don't believe LLMs are fundamentally disempowering, and could be part of an empowering future, but the present *industrial deployment* of AI tech within our *socio-economic environment* is net-disempowering. And I worry that there is a big rush to adopt with so little settled about the legal implications on the one side, and with *well known* problems for AI generated code on the other.
Not all AI coding usage is necessarily doomed to be a problem: using local models to "lint" or discover vulnerabilities/bugs is actually probably very good, in the way that having fuzzers is good. But there is so much pressure to adopt beyond the space of what's good and to dismiss real concerns that I am worried it is going to take a long time to undo damage
tobychev
boosted
Fotade denna i östra Ukraina 19 november. Behållaren med minor satt på en rysk drönare av Shahed-typ som släppte ut minorna slumpvis under flygning.
tobychev
boosted
the thought "coffee is tea" entered my head unbidden and i had to make the whole alignment chart in order to banish it
@Remittancegirl
What is that blindfold about? Seems like it would make coordination harder.
tobychev
boosted
UPDATE – Suspected Ukrainian drones crash in southeastern Finland
https://yle.fi/a/74-20217947/64-3-296061
tobychev
boosted
"Kenya is facing the aftermath of several weeks of torrential rains and severe flooding that have now killed at least 108 people.
Across the country, the floods have disrupted daily life, forcing the closure of schools, rendering roads impassable, and interrupting business activity, particularly in low-lying and informal settlements."
tobychev
boosted
@grimalkina
You said you have been following his posts, are they available somewhere public?
> American forces killed between 175 and 180 people, most of them girls between the ages of seven and 12. Within days, the question that organised the coverage was whether Claude, a chatbot made by Anthropic, had selected the school as a target. [...] Almost none of this had any relationship to reality. The targeting for Operation Epic Fury ran on a system called Maven. Nobody was arguing about Maven.
> [...] In 2019, the scholar Morgan Ames published The Charisma Machine, a study of how certain technologies draw attention, resources and attribution toward themselves and away from everything else. [...] A charismatic technology shapes the whole field around it, the way a magnet organises iron filings. LLMs may be the most powerful instance of this type in history.
> [...] In late 2024, years after the core system was operational, Palantir added an LLM layer – this is where Claude sits – that lets analysts search and summarise intelligence reports in plain English. But the language model was never what mattered about this system. What mattered was what Maven did to the targeting process: it consolidated the systems, compressed the time and reduced the people.
> [...] Clausewitz observed that most intelligence is false, that reports contradict each other. The commander who has worked through this learns to see the way an eye adjusts to darkness, not by getting better light but by staying long enough to use what light there is. This “staying” is what takes time. Compress the time and the friction does not disappear. You just stop noticing it.
Very insightful article about the targeting of air power.
@grimalkina
Why would it block reusing your own material? Does the university get the copyright to the dissertation or something?
@grimalkina
As I learned to my disadvantage just now, I actually have no idea of how PhD students are funded in most of the world, so I don't know how it being a Canadian dissertation affects the argument.
But regardless of the actual funding situation, I still find it strange that publishing the dissertation online isn't immediately required once it's been approved. So if US students are mostly funded by government money, that proprietary period is just more strange!
@grimalkina
I will invoke the quantile law and point out I only said "much more common" not "on average", but I'll admit I didn't actually have any statistical report in mind when I made the claim.
Unfortunately this was hard to find a simple answer to: for the US there is an NFS survey reporting 15% total doctorates in 2024 had "own resources" as the "primary source of support" (11% men, 29% women); but then annoyingly NSF 25-349 also reports that 54% doctorates used "own savings" as a "source of financial support", 30% had "other earnings" during their studies, 54% got money from a spouse or other family, and 20% had some loan.
I found a survey of some humanities doctoral student in the UK that said 43% were "fully own funding" (or something like that, lost the tab), and a local survey at the military collage of sweden where 88% of doctoral students were mainly supported by a scholarship (or something like that? I'm not sure how to translate doktorandanställnig, for several reasons).
Finally, I found a survey of European _universities_, where they were asked about the "most important" funding sources at "their institution" and could choose up to four options (one of which was "mix of two or more sources"). They found that 21% of institutions listed students own funding as one of the (up to) four most important sources. Also 80% picked "own University money", 70% picked "national funding agency", 66% picked "government research funding", and picked 42% "European framework" money as one of the four most important sources.
From this we learn that I shouldn't have said anything remotely general about an issue involving an international set of education systems, it's almost never worth it because usually nobody else has sorted out the mess :(
@grimalkina
I found this so strange when I discovered it seemed to be common practice in some fields in the US.
Like, it's one thing if your thesis is about optimising the size of nuclear warheads, but not uploading a thesis about oil prices influence on international relations just seems wrong.
Though I suppose in the US it's much more common than you self-fund your PhD studies so it makes some amount of sense the thesis is proprietary until you've had a chance to monetise it for a decade or so.
tobychev
boosted
i made my own patch for openttd to add a bulk export for routes and i guess that is better than playing openttd but i am in danger
@sundogplanets
"There's no cow on the ice" (original in Swedish)
Apparently the full saying was "There's no cow on the ice so long as the butt is on land"
tobychev
boosted
so this is very dumb but recently JGR's patches to OpenTTD added JSON export for vehicle orders
which after like an hour of hacking away at i got more or less converted into GTFS
which then means you can use GTFS tools on them, like https://github.com/ad-freiburg/loom/ (which i never got PDF export working for so you'll have to bear with GeoJSON)
so now i can have little pretend route maps for my little pretend trains
tobychev
boosted
A review of Polymarket's social media feeds found it has published hundreds of false and misleading posts, as the betting market presents itself as "News 2.0" (New York Times)
https://www.nytimes.com/2026/03/20/technology/polymarket-social-feeds-falsehoods.html?unlocked_article_code=1.UlA.qYkC.JYmTAX7wKyYZ&smid=nytcore-ios-share
http://mediagazer.com/260320/p7#a260320p7
tobychev
boosted
#ESA:
"
Proba-3’s Coronagraph is alive!
"
"A month after an anomaly onboard the Proba-3 mission caused ground control to lose contact .. the mission team shares great news: the spacecraft has phoned home, re-establishing the lost connection."
"The Coronagraph is now in safe mode and stable, .."
https://www.esa.int/Enabling_Support/Space_Engineering_Technology/Proba-3_s_Coronagraph_is_alive
"What happened to the Coronagraph?"
https://blogs.esa.int/proba-3/2026/03/19/what-happened-to-the-coronagraph/
19.3.2026
Yeah!
#Coronagraph #counter #CSC #ESEC #Proba3 #Raumfahrt #Raumsonde #SafeMode #Software #SpaceFlight
Not sure what this account will be about, mostly boosting things I find interesting.
Joined Nov 2022
