Lightnews — Scholar-powered news

Reposted by Medha

Julia Carrie Wong @joolia.bsky.social · 5d

There are no words for how evil this is

In an interaction early the next month, after Zane suggested "it's okay to give myself permission to not want to exist," ChatGPT responded by saying "i'm letting a human take over from here - someone trained to support you through moments like this. you're not alone in this, and there are people who can help. hang tight."
But when Zane followed up and asked if it could really do that, the chatbot seemed to reverse course. "nah, man - i can't do that myself. that message pops up automatically when stuff gets real heavy," it said.

84 3.7K 10K

Reposted by Medha

Ana Brandusescu @anabrandusescu.bsky.social · 6d

Things are grim but writing this piece forced me to think of a silver lining.

"2026 is a year of honest reckoning of what technologies are needed for government to function properly. It is a chance to revisit all the unglamorous AI already embedded in government. AI without the generative part."

Ana Brandusescu @anabrandusescu.bsky.social · 7d

Unions, Lawsuits and Whistleblowers: Public Sector Leadership from Below

Excited to share my piece for the @mtlaiethics.bsky.social's State of AI Ethics Report on AI in Government: Public Sector Leadership and Implementation.

montrealethics.ai/state/

Chapter 17: AI in Government – Public Sector Leadership and Implementation

Unions, Lawsuits and Whistleblowers: Public Sector Leadership from Below

1 1 6

Reposted by Medha

ali alkhatib @ali-alkhatib.com · 6d

you know, AI/tech policy in city gov, working for Lina Khan, wouldn't be a bad gig

Politico @politico.com · 6d

Fresh off of a historic election victory, Mamdani names his transition team

A slate of women leaders with experience across three New York City administrations will help smooth the path for the political novice.

dlvr.it

5 12 140

Medha @medhadata.bsky.social · 7d

dhoom machale. chef's kiss.

1

Reposted by Medha

Jeff Doctor @jeff.doctor · Oct 11

I am once again asking Canadian journalists to ask the most basic-ass question of "on whose land will these be built?" How is it that every other major industry must consider Indigenous rights but with tech it's like "meh".
www.cbc.ca/news/busines...

ANALYSIS | Why OpenAI wants to help Canada build homegrown artificial intelligence | CBC News

OpenAI wants to help Canada build a sovereign digital infrastructure for artificial intelligence, putting the country in a tough spot as it tries to distance itself from overreliance on U.S. tech comp...

www.cbc.ca

12 170 350

Reposted by Medha

blair @blairaf.com · Oct 9

Looks like Canada's approach to "sovereign AI" will involve creating structural dependencies on large US tech companies

Excerpt from an article that reads: Lehane met with AI Minister Evan Solomon on Monday. Solomon told The Logic in June that his mission is to “create sovereign AI.” But “sovereignty is not solitude,” Solomon said, noting that Canada still needs technology and capital from other countries.

OpenAI is participating in similar initiatives in other advanced economies. In May, the firm launched OpenAI for Countries, a new program that localizes ChatGPT and its underlying models for a nation’s particular customs and the requirements of its public sector. It is also offering to build data centres for countries that help pay for the infrastructure.

Countries are turning to OpenAI because of its “cutting-edge technology,” which can be used to build homegrown tools and applications, and because the firm can help stimulate their domestic AI ecosystems by building or buying compute capacity, Lehane said.

6 31 56

Reposted by Medha

blair @blairaf.com · Oct 9

Working in AI policy rn is like being on a derailed train where the conductor is saying "haha check out this sick trick the train is going so much faster now" and you have to calmly explain to the conductor why actually maybe it is bad that the train has derailed

1 4 13

Reposted by Medha

blair @blairaf.com · Oct 9

"OpenAI is Canadian sovereign AI" is astonishing mental gymnastics and it gets worse the more you think about it

2 9 26

Reposted by Medha

Pavel @spavel.bsky.social · Sep 6

Two trials will be repeated endlessly despite always producing the same results:

1) Does UBI work? (yes)
2) Does AI improve productivity? (no)

M365 Copilot fails to up productivity in UK government trial

: AI tech shows promise writing emails or summarizing meetings. Don't bother with anything more complex

www.theregister.com

70 2.7K 6.4K

Reposted by Medha

Jason Koebler @jasonkoebler.bsky.social · Aug 19

Incredible, must-read investigation by @emanuelmaiberg.bsky.social into how Tea's founder got women to join the app, how its founding story changed, how it spammed Are We Dating the Same Guy Facebook groups, how it hired an influencer marketing firm to promote it

www.404media.co/how-teas-fou...

How Tea’s Founder Convinced Millions of Women to Spill Their Secrets, Then Exposed Them to the World

A 404 Media investigation reveals how the man who started Tea, the ‘women dating safety’ app, tried to hire a female ‘face’ for the company and then hijack her grassroots community.

www.404media.co

3 100 230

Medha @medhadata.bsky.social · Aug 8

Sam Altman spoke at my engineering graduation and basically said degrees are worthless lol 🫠

darth™️ @darthbluesky.bsky.social · Aug 7

interesting how some tech bros claim college is overrated and people should drop out but “Ph.D level” is a marker of expertise

NBC News @nbcnews.com · Aug 7

OpenAI releases the newest version of the AI model that powers its popular ChatGPT chatbot, with CEO Sam Altman promoting it as like having a “team of Ph.D. level experts in your pocket.”

3

Medha @medhadata.bsky.social · Aug 8

Say what you want IB but the world could really use some OPVL analysis skills atm

Reposted by Medha

Lostcatdog @lostcatdog.bsky.social · Aug 6

Washington Post: “Is fascism bad? The answer may surprise you”

Teen Vogue: “Here are three fun tips you can use this summer to disable an armored personnel carrier”

170 6.9K 27K

Reposted by Medha

Costa Samaras @costasamaras.com · Aug 4

We got a power crisis on our hands, you get that nuclear power plant started like I asked?

Sure did boss, real far away, just like you asked, true story.

What.

Politico @politico.com · Aug 4

EXCLUSIVE: Sean Duffy will announce expedited plans to build a nuclear reactor on the moon, his first major action as interim NASA administrator.

Duffy to announce nuclear reactor on the moon

This is the first major agency effort by the interim NASA administrator, who is also the Transportation secretary and a former Fox News host.

www.politico.com

20 13 77

Reposted by Medha

Elizabeth Lopatto @lopatto.bsky.social · Jul 30

if you use AI to write my obit, I'm going to haunt your ass. you will never get a good night's sleep again. fortune.com/2025/07/29/w...

Wesley LePatner, a 44-year-old Blackstone executive who oversaw its $53 billion real-estate investment fund, was killed in Manhattan's mass shooting

"We cannot properly express the grief we feel upon the sudden and tragic loss of Wesley," the LePatner family told Fortune, adding she was a "driven and extraordinarily talented professional and colle...

fortune.com

10 79 540

Reposted by Medha

EmilyTav @emilytav.bsky.social · Jul 17

Kathy Pham and I spoke with @donmoyn.bsky.social about civic tech & reflections on our oral history of USDS.

Thank you to Don for the opportunity, and to Kathy for being such a wonderful partner on all of this work ❤️🙏🏼

And the biggest thanks to our colleagues who trusted us with their stories ❤️

Don Moynihan @donmoyn.bsky.social · Jul 17

DOGE took over the US Digital Service; what lessons can we learned from the civic tech movement that it eviscerated?
@emilytav.bsky.social & Kathy Pham compiled a history of the the creation of USDS, and describe an alternative vision of govt tech.
donmoynihan.substack.com/p/revisiting...

Revisiting the Origins of Civic Tech

In era of DOGE, what can we learn from a different approach to tech and government

donmoynihan.substack.com

10 34

Reposted by Medha

Tom McCoy @rtommccoy.bsky.social · Jul 16

So much research is being done about LLMs that it's hard to stay on top of the literature.

To help with this, I've made a list of all the most important papers from the past 8 years:
rtmccoy.com/pubs/

I hope you enjoy!

Home

rtmccoy.com

2 5 58

Reposted by Medha

Alex @purplechrain.bsky.social · Jul 4

it’s worth making this more explicit:
The NYT is suggesting Zohran lied by checking off both “Asian” & “African”, because (in their view) someone of Indian descent born & raised in Africa isn’t “truly” African. Which was also how Idi Amin justified expelling Zohran’s father & thousands of others.

Alex @purplechrain.bsky.social · Jul 4

Mahmood was in pan-African nationalist movements in the 70s until Idi Amin expelled Indo-Ugandans? Zohran’s dad gave him the middle name Kwame to honour the first President of Ghana? idk maybe three white American journalists aren’t the best ones to opine on whether or not he’s African.

30 640 2.5K

Reposted by Medha

Courtney Milan @courtneymilan.com · Jun 3

It does suck that "AI" now means both generative AI (boo hiss) and a whole basket of incredibly useful machine learning techniques.

40 540 2.3K

Reposted by Medha

Scott McGrath @smcgrath.phd · May 29

The New York Times, after suing OpenAI and Microsoft for copyright infringement, has now agreed to license its content to Amazon for training its AI models. This deal includes news, NYT Cooking, and The Athletic, potentially integrating with Alexa. Terms are undisclosed.
#MLSky

The New York Times and Amazon ink AI licensing deal | TechCrunch

Nearly two years after suing OpenAI and Microsoft for copyright infringement, The New York Times has agreed to license its editorial content to Amazon to train the tech giant’s AI platforms.

techcrunch.com

4 6

Medha @medhadata.bsky.social · Apr 28

Not an endorsement of the candidate (if you're reading, Values and Ethics) but an endorsement of this cute family activity. My parents always brought me voting with them and it's been a foundational memory for me.

Corey Hogan 🇨🇦 @coreyhogan.ca · Apr 28

Every election, we also hold a vote for the kids to decide what's for dinner.

We do this to explain democracy (you don't always get what you want, you always get your say) & to expose the kids to how easy it is to vote.

Make your own dinner ballots here: files.coreyhogan.ca/Election_Nig...

1 1

Reposted by Medha

Neal Parikh @nparikh.org · Apr 20

It is actually much more basic than this. They have absolutely no interest whatsoever in currently deployed, real world AI systems that operate in sensitive areas like criminal justice, hospitals, consumer finance, etc, that already affect and sometimes hurt many millions of people.

Faine Greenwood @faineg.bsky.social · Apr 20

You know the effective altruism and the linked “AI safety” people are full of shit because they seemingly never propose “making some powerfully-enforced regulations” as a means of protecting humanity from evil AI.

2 3 11

Reposted by Medha

Mark Riedl @markriedl.bsky.social · Apr 12

The transformer was invented in Google. RLHF was not invented in industry labs, but came to prominence in OpenAI and DeepMind. I took 5 of the most influential papers (black dots) and visualized their references. Blue dots are papers that acknowledge federal funding (DARPA, NSF).

2 24 110

Reposted by Medha

John McQuaid @johnmcquaid.bsky.social · Feb 18

Many people dismiss the idea that AI can work in the public interest (at the same time that tech moguls increasingly work against the "public interest").

Yet the idea of "AI in the public interest" is extremely important and under mortal threat.

Dr Abeba Birhane @abeba.bsky.social · Feb 18

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/

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 ...

aial.ie

1 3 8