Sign in

Michelle Shevin

@michebox.bsky.social
124 followers 177 following 32 posts

Future Preservation Society

PostsRepliesMedia
Reposted by Michelle Shevin
Deb Raji @rajiinio.bsky.social · 08/05/2025
This might be my villain origin story: www.wired.com/story/openai...
wired.com
OpenAI and the FDA Are Holding Talks About Using AI In Drug Evaluation
High-ranking OpenAI employees have met with the FDA multiple times in recent weeks to discuss AI and a project called cderGPT.
15112
Reposted by Michelle Shevin
Dr Francis Young @drfrancisyoung.bsky.social · 08/05/2025
People are always saying on social media that gen-AI is here to stay and that it’s retrograde and standing in the way of progress if we reject it. But there are plenty of things that were once the embodiment of progress that we collectively decided to reject, from cocaine in soft drinks to eugenics
1630173
Reposted by Michelle Shevin
Tressie McMillan Cottom @tressiemcphd.bsky.social · 07/05/2025
The hysteria when you say, “nah I’m good” on all the chat bots is how you know they’re in the grips of roulette fever. You know the guy at a casino, shaking at 2 am, telling you how dumb you are not to get in on this?
102142373
Reposted by Michelle Shevin
Julia Angwin @juliaangwin.com · 30/04/2025
DOGE is assembling a master database of personal data from across the government -- including tax records, social security files, immigration dossiers etc. This cross-agency master database is exactly what Congress tried to prevent when it passed the Federal Privacy Act of 1974.
10450163
Reposted by Michelle Shevin
Julia Angwin @juliaangwin.com · 30/04/2025
A centralized database will be "ill-equipped to correct errors, allow for extenuating circumstances, or bring facts up to date," warned Vance Packard. "The notion of the possibility of redemption is likely to be incomprehensible to a computer."
A centralized database will be "ill-equipped to correct errors, allow for extenuating circumstances, or bring facts up to date," warned Vance Packard. "The notion of the possibility of redemption is likely to be incomprehensible to a computer."
318249
Reposted by Michelle Shevin
Sean Martin McDonald @seanmartinmcdonald.bsky.social · 17/04/2025
This consolidation started long before now and was a product of a lot of American hubris. But this is definitely the “find out” phase and it’s going to be brutal.
042
Reposted by Michelle Shevin
Brian Merchant @bcmerchant.bsky.social · 12/04/2025
This is, the therapists say, very dangerous—the algorithm doesn't know how to detect nuance, or when to ask follow up questions, for instance, and may be turning away people in genuine crisis. Algorithmic triage, they say, has been a disaster, and they've fought it since its inception.
1879
Reposted by Michelle Shevin
ali alkhatib @ali-alkhatib.com · 11/04/2025
i'm ambivalent about litigating algorithmic fairness and accuracy; this is why. negotiating administrating this system fairly or accurately is nonsense. there's no running a colony ethically, or an empire in a human-centered way. you can make up bullshit and hit those targets, but they mean nothing.
0202
Reposted by Michelle Shevin
Brian Merchant @bcmerchant.bsky.social · 09/04/2025
After Trump was elected the first time, there was also a lot of talk about making iPhones in the US. At the time, I co-edited the sci-fi outlet Terraform, and asked @timmaughan.bsky.social to imagine what such a future might really look like. With his permission, I'm republishing the story in full:
bloodinthemachine.com
Flyover Country
What a future where iPhones are made in America would really look like.
23704328
Reposted by Michelle Shevin
ohpipedown.bsky.social @ohpipedown.bsky.social · 08/04/2025
Go slow and love intensely.
032
Michelle Shevin @michebox.bsky.social · 06/04/2025
“Is your heart beating and do you plan to live? Come this way…” —@adriennemareebrown.bsky.social
000
Reposted by Michelle Shevin
Anna Bower @annabower.bsky.social · 05/04/2025
Erez Reuveni argued on behalf of the Justice Department during Friday’s hearing in the Abrego Garcia case. Reuveni has reportedly been placed on leave. Here’s one exchange with the judge that may have contributed to his ouster:
THE COURT: Mr. Abrego Garcia should not have been removed, right?
MR. REUVENI: Yes.
THE COURT: Can you answer for me, then, on what authority was he seized? When he was --when he was taken off the street, taken out of his car, what authority did those law enforcement officers have to do that?
MR. REUVENI: So, Your Honor, my answer to a lot of these questions is going to be frustrating. I am also frustrated that I have no answers for
you on a lot of these questions.
1312869871
Michelle Shevin @michebox.bsky.social · 05/04/2025
a pro life tip is to find someone you like to tell your dreams to
medium.com
a poem i wrote while falling in love with my life
I dreamt my body electric -- tree-like, geomagnetic...
010
Michelle Shevin @michebox.bsky.social · 03/04/2025
“Planning on co-liberation and collective thriving doesn’t mean things proceed according to a plan. It looks more like prioritizing, over and over again, the reality that while the future is unwritten, my destiny is tied up with yours— that all life is an interconnected web…
120
Reposted by Michelle Shevin
Tom King @tallgeekychap.bsky.social · 02/04/2025
The hype machine for Big Tech, "AI", and cyberlibertarianism shows no sign of slowing down. This is because it actually has to speed up to balance the number of red lights flashing on the dashboard in terms of "AI's" usefulness, contribution to human happiness, and (crucially), profitability.
063
Reposted by Michelle Shevin
Tamara Kneese @tamigraph.bsky.social · 01/04/2025
Holy shit “Trump administration attorneys told the court to dismiss the request on multiple grounds, including that Trump’s “primacy in foreign affairs’ outweighs the interests of Abrego Garcia and his family”
072
Reposted by Michelle Shevin
Nick Miroff @nickmiroff.bsky.social · 01/04/2025
BREAKING: Trump admin says it mistakenly deported a Maryland father who had *withholding of removal* back to El Salvador on March 15. Govt told court Monday it won't bring Kilmar Abrego Garcia back. "What.The.Fuck..." one govt attorney told me. www.theatlantic.com/politics/arc... @theatlantic.com
theatlantic.com
An ‘Administrative Error’ Sends a Maryland Father to a Salvadoran Prison
The Trump administration says it mistakenly deported an immigrant with protected status but that courts are powerless to order his return.
33161572779
Reposted by Michelle Shevin
Eryk Salvaggio @eryk.bsky.social · 26/03/2025
Pro-AI hype and Anti-AI doom are so deeply personal and polarized for a reason: they are the only visions of the future anyone is offering at all. In my latest for @techpolicypress.bsky.social, a look at the loss of faith in participatory politics that frames AI hype, but also its backlash.
techpolicy.press
Future Fatigue: How Hype has Replaced Hope in the 21st Century | TechPolicy.Press
To resist AI hype, we must not reassert the fiction that we could “return” to a functional democracy, writes Eryk Salvaggio.
66927
Reposted by Michelle Shevin
Eryk Salvaggio @eryk.bsky.social · 26/03/2025
Setting them aside isn't the prescription, so I don't think we're that far off. Mostly I want to see a "future" without AI at the center. Most Gen AI use cases are inseparable from this techno-capitalist impulse. But others aren't. Giving up one because we don't like the other feels like a retreat.
221
Reposted by Michelle Shevin
Ketan Joshi @ketanjoshi.co · 26/03/2025
Brilliant piece by @hannahstoryb.bsky.social here - the naive 'JUST BUILD THINGS' abundance agenda is a boon to the fossil fuel industry. prospect.org/culture/book...
powerful enemies. The "liberalism that builds" camp, in short, would find more allies among the progressives they criticize if they acknowledged that some things should not, in fact, be built. We need better levers to prevent abuses and
better processes for getting
public goods built. But the authors refuse to dignify the
question of how to prevent existing abuses from worsening if we take away existing
guardrails.
K
158922
Reposted by Michelle Shevin
John Skiles Skinner @skiles.blue · 21/03/2025
There is a quasi-religion in Silicon Valley that views AI as godlike. This faith has always been parallel to Evangelical Christianity: salvation (transhumanism), the rapture (the technological singularity), and demons (Roko's Basilisk) Lately the AI faith has fully fused with Christian Nationalism.
10159481414
Reposted by Michelle Shevin
Ben Williamson @benpatrickwill.bsky.social · 21/03/2025
The promise of AI in education is it will save educators time. That's rubbish. It's sapping all our time already in working groups, committees, workshops, and marking anxiety. It's the most time-sapping, exhausting and frankly boring thing I've ever encountered in HE - and it's in my research area.
12707202
Reposted by Michelle Shevin
Dr Abeba Birhane @abeba.blacksky.app · 22/03/2025
despite popularised beliefs, LLMs are not fit for medical applications. SoTA models produce "non-trivial levels of hallucinations" even w inference techniques like CoT & search augmented generation: arxiv.org/pdf/2503.05777 of surveyed clinicians, 53% use LLMs daily & 91% encountered hallucinations
13302120
Reposted by Michelle Shevin
gquaggiotto.bsky.social @gquaggiotto.bsky.social · 21/03/2025
[Looks terrific] "This conference aims to examine hype as a performative force, exploring its mechanisms, effects, and implications across different domains" hypestudies.org/conference
053
Michelle Shevin @michebox.bsky.social · 18/03/2025
Just learned a lot about the egg cartel. Big egg is worth the read.
032
Michelle Shevin @michebox.bsky.social · 18/03/2025
David Graeber once said “this feeling of hopelessness that everyone has is a manufactured product.” Today I’ve published a piece about how the future gets foreclosed and how we fight back with our minds and relationships.
medium.com
Possibility + Power = Plan
I soured on my first strategic foresight job gradually, and then all at once.
2216
Reposted by Michelle Shevin
Melanie Mitchell @melaniemitchell.bsky.social · 18/03/2025
Very excited to participate in this event with @randomwalker.bsky.social next week! lectures.princeton.edu/lectures/202...
lectures.princeton.edu
A Guide to Cutting Through AI Hype: Arvind Narayanan and Melanie Mitchell Discuss Artificial and Human Intelligence
Princeton University Professor Arvind Narayanan and Professor Melanie Mitchell of the Santa Fe Institute, will share their insights on Artificial Intelligence in a discussion format, followed by a que...
17519
Reposted by Michelle Shevin
Shannon Mattern @shannonmattern.bsky.social · 17/03/2025
In this post Kelly provides great list of easy steps you can take: sign the EveryLibrary petition, call your senators + reps (use 5Calls!). Protect our state libraries, tribal libraries, interlibrary loan, digital equity, vital grants to libraries of all species... 📣 And please spread the word!! 📣
011199
Reposted by Michelle Shevin
Mehdi Hasan @mehdirhasan.bsky.social · 16/03/2025
“Denying the Seriousness of Long COVID Is Dangerous for Me – and You” On Long COVID Awareness Day, pleased to publish this piece for Zeteo from @juliadoubleday.bsky.social
zeteo.com
Denying the Seriousness of Long COVID Is Dangerous for Me – and You
What is Long COVID? Are there treatments? A guide to the illness that affects millions, but is hardly talked about.
672898814
Reposted by Michelle Shevin
Arrianna Marie Planey, PhD MA @arrianna-planey.bsky.social · 15/03/2025
We need mask mandates in healthcare settings more than ever
18834
Reposted by Michelle Shevin
Erin Richardson @erinrichardson.bsky.social · 15/03/2025
This is going to kill many public libraries, especially in rural areas. If you use a public library, you will be impacted. #imls is a dust mote in the federal budget. It also provides things like internet at your local library so people can apply for jobs and kids can do homework.
3815501061
Reposted by Michelle Shevin
Hypervisible @hypervisible.blacksky.app · 15/03/2025
“‘The Trump administration has removed safety, fairness, misinformation, and responsibility as things it values for AI, which I think speaks for itself,’ says one researcher at an organization working with the AI Safety Institute, who asked not to be named for fear of reprisal.”
wired.com
Under Trump, AI Scientists Are Told to Remove ‘Ideological Bias’ From Powerful Models
A directive from the National Institute of Standards and Technology eliminates mention of “AI safety” and “AI fairness.”
614369
Reposted by Michelle Shevin
Eryk Salvaggio @eryk.bsky.social · 15/03/2025
Oof.
In a recent study, colleagues and I asked computer scientists to develop a software tool that gives automated advice about whether people are eligible for bankruptcy relief. The computer scientists completed the task quickly, with one even noting, “this seems very simple.” However, when we analyzed the software tools they built, we found that errors and legal misconceptions were rampant. Despite these flaws, the computer scientists were confident that their tools would be beneficial in practice. Seventy-nine percent of them stated that they would be comfortable with their tool replacing judges in bankruptcy court.
06222
Reposted by Michelle Shevin
Amanda Litman @amandalitman.bsky.social · 14/03/2025
Run for office. runforwhat.net - our next intro call is 3/25.
runforwhat.net
Where Can I Run — Run for Something Action Fund
Use this tool to find out when, where and how you can run for office.
2511297
Michelle Shevin @michebox.bsky.social · 14/03/2025
Unfortunately this piece I wrote in 2018 on the absolute nonsense in the middle of the AI narrative stands up uncomfortably well in 2025: medium.com/@michebox/th...
medium.com
The Flawed Logic at the Heart of the Automation Fantasy
Tech doesn’t solve for trust, accountability, or labor — it shifts responsibility away from systems and onto individuals.
074
Reposted by Michelle Shevin
Jeffrey Vagle @jvagle.me · 14/03/2025
Futurist/Effective Altruist/Longtermist AI* hype men are essentially saying that we must sacrifice some part of humanity (but not them, of course) in order to "save" humanity through the building of an AI* god. If you think I'm exaggerating, have I got a reading list for you.
2104
Michelle Shevin @michebox.bsky.social · 13/03/2025
Designed and deployed under conditions of austerity, suspicion, and routine criminalization, fabricated pretext for state violence is a feature of predictive (text) models. As @ruha9.bsky.social has pointed out these are not glitches but rather glimpses of the way the system is intended to work.
020
Reposted by Michelle Shevin
Ruha Benjamin @ruha9.bsky.social · 12/03/2025
🚨Last week Helyeh Doutaghi, a scholar in international law, was barred from Yale and put on administrative leave: ‘The decision came three days after a news site powered by artificial intelligence, published a story calling her a member of a terrorist group.’ www.removepaywall.com/search?url=h...
7276162
Reposted by Michelle Shevin
Sean Martin McDonald @seanmartinmcdonald.bsky.social · 07/03/2025
This is why so many of us have been working on systems of contestability and governance in digitally defined systems for so long. The perception of "data"s value is, itself, a weaponizable ambiguity that gets used to huge effect, whether it's right or wrong. It always has been.
011
Reposted by Michelle Shevin
Movement Generation @movementgeneration.bsky.social · 26/02/2025
We’re new here, and it seems like a good time to share radical independent media. Read our latest Murmurations column in @yesmedia.bsky.social : Climate Solutions Require Black Ecology Thanks @adriennemareebrown.bsky.social ! www.yesmagazine.org/environment/...
yesmagazine.org
Murmurations: Climate Solutions Require Black Ecology
Slavery and genocide fueled climate destruction. Black liberation will fuel regeneration.
05311
Michelle Shevin @michebox.bsky.social · 21/02/2025
Have just watched Kevin @techtonicjustice.bsky.social speaking with CT lawmakers about AI with lots of good questions from citizens & man it feels good to glimpse functional civic engagement.
053
Reposted by Michelle Shevin
Tom King @tallgeekychap.bsky.social · 19/02/2025
1) It makes up links 2) It makes up papers that it can cite 3) This means if you're doing your job properly, you have to check every single link and citation, which is slower than finding them yourself, because your process doesn't generate false positives Don't trust guessing machines.
0442
Reposted by Michelle Shevin
Sean Martin McDonald @seanmartinmcdonald.bsky.social · 19/02/2025
"Each government’s attempt to define the rules either projects its policy globally or fragments what was once the common ground of some aspect of the internet." (from 2018) w/AX Mina for @foreignpolicy.com on #digitalpolitik 🙏@beijingpalmer.bsky.social foreignpolicy.com/2018/12/19/t...
foreignpolicy.com
2018 was the year the Internet splintered
The new rules of digital warfare are coming into shape.
011
Reposted by Michelle Shevin
Dr Heidy Khlaaf (هايدي خلاف) @heidykhlaaf.bsky.social · 18/02/2025
New investigation by @apnews.com on the IDF's use of OpenAI models in Gaza. I spoke to them regarding this first confirmation of commercial foundation models being directly used in warfare and the implications of using highly error-prone AI for life or death determinations apnews.com/article/isra...
apnews.com
As Israel uses US-made AI models in war, concerns arise about tech’s role in who lives and who dies
U.S. tech giants have quietly empowered Israel to track and kill many more alleged militants more quickly in Gaza and Lebanon through a sharp spike in artificial intelligence and computing services.
13831
Reposted by Michelle Shevin
Dr Abeba Birhane @abeba.blacksky.app · 18/02/2025
I was in Paris last week for the #AIActionSummit & was honoured to participate in the “AI in public interest” panel. 👇🏾 my thoughts & reflections on what AI in public interest is/isn’t & some concrete steps/initiatives for `bending the arc of AI towards the public interest' aial.ie/pages/aiparis/
aial.ie
Bending the arc of AI towards the public interest
By Abeba Birhane, 18/02/2025 Following the first in Bletchley Park in 2023 and the second in Seoul in 2024, the third AI Action Summit took place in February 2025 in Paris. In the context of previous ...
921779