Lightnews — Scholar-powered news

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · 2d

JFK is setting expectations really low these days, huh?

Sign at the airport with a quote from Lao Tzu: "A good traveler has no fixed plans and is not intent on arriving."

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Center for News, Technology & Innovation @cnti.bsky.social · 6d

What does responsible use of AI look like for the future of journalism?

CNTI's @drnycnerd.bsky.social will discuss ways AI is reshaping newsrooms as part of First Amendment Day at UNC.

🗓️ Wednesday, October 8, 2025
🕚 11:00 a.m.
📍 UNC
✅ Free and open to the public

Full details: lnkd.in/g-y2FyUE

1

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Center for News, Technology & Innovation @cnti.bsky.social · 16d

In Serbia, a leaked recording has raised fears of political attempts to weaken independent media. As the country adopts the EU’s DSA, a question looms: Will this law protect press freedom—or be misused to suppress it?

👉 Read more from @ewrightwy.bsky.social:
euobserver.com/eu-and-the-w...

How Serbia could use EU Digital Services Act for state censorship

The implementation of the EU's Digital Services Act in Serbia and other countries lacking an independent digital services coordinator risks becoming an EU-based tool for state repression, enabling rul...

euobserver.com

2 2

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Eman Abdelhadi @emanabdelhadi.bsky.social · 20d

Eat. Sleep. Exercise. Nurture the relationships that sustain you. Protect your health. Find joy. Don't mistake any of these for the resistance itself. Joy isn't resistance, neither is rest. But you need to be whole to resist for a lifetime.

1 210 800

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Chanda Prescod-Weinstein 🌌 @chanda.blacksky.app · Aug 17

It’s weird to me that people keep starting new Substacks and still don’t know that Substack is extremely bad news

Please get off Substack 😭

www.anildash.com//2024/11/19/...

Don't call it a Substack. - Anil Dash

A blog about making culture. Since 1999.

www.anildash.com

18 290 830

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Joshua Benton @joshuabenton.com · Aug 7

GPT-5 is so amazing it can make 52.8 higher than 69.1 on a bar chart

It can even make 69.1 and 30.8 THE SAME SIZE

13 60 280

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Kathryn C-K @kathrynck.bsky.social · Aug 4

The real secret about genius is that it refers to an emotional experience in the perceiver, not a quality in the perceived. It's an amazing feeling, when someone lights up your brain with their skill. It's why I love my job. But it is a feeling, not an objective event

Faine Greenwood @faineg.bsky.social · Aug 4

I am really sick of all these pieces about the DOGE boys who uncritically and unthinkingly repeat the claim that they're "geniuses."

We have to stop regurgitating out the label of "genius" exclusively for a certain type of special white little boy.

Ari @arithmetc.bsky.social · Aug 4

We are also mythologizing his brilliance because he fits the mold of the singular boy genius tech stereotype. We need to stop and question that too.

4 9 89

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Aug 1

Even worse!

1

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Aug 1

They get to pretend they're accountable?

Is this ConEd by any chance?

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Aug 1

Some things researchers can do:
- ask about specific tools or use cases or outputs
- discourse analysis to tease out what's included in "AI" policy documents, the media, etc.
- interviews and thinkalouds to suss out what people think "AI" is

3

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Aug 1

I finished reading "The AI Con" last week and I'm going to be sitting with it for a long time.

As a researcher it affirms my frustration with the slipperiness of "AI" as a term. All the research shows that people don't understand it consistently.

1 3

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Jul 31

What do we do if we're already a monthly donor and we desperately want a pin? Asking for ... a friend?

1 1

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Jul 31

What do news organizations actually need to know about AI & journalism? A new briefing from @cnti.bsky.social's Global AI & Journalism Research Working Group synthesizes 80+ research studies.

cnti.org/article/aij-...

Briefing #1: AI Literacy & Communication

July 2025

cnti.org

1 2 2

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Q. Pheevr @q.pheevr.ca · Jul 31

🥇 Black: The good, normal way to have coffee

🥈 Milk, no sugar: Not bad at all, but superfluous (unless the coffee is too strong)

🥉 Milk and sugar: Fine if I think of it not as coffee, but as a coffee-flavoured dessert (affogato lite?)

💩 Sugar, no milk: O gods this is vile, why would you do this

3 1 4

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Jul 31

This briefing is the first in a series. Stay tuned for more!

If you'd like to suggest topics or research to include, please reach out.

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Jul 31

5. The research consistently shows that explaining *how* AI was used is more valuable than stating *that* AI was used -- "AI" is a broad umbrella term that means different things to different audiences.

1

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Jul 31

4. Covering AI effectively benefits from both understanding the technology deeply and hearing from the broadest possible range of stakeholders.

1

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Jul 31

3. One big challenge for public understanding is that AI tools upend previous assumptions about how technology works: their outputs are not consistent (unlike, say, a calendar) and they're often "under the hood" such that people may not even know they're interacting with one.

1

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Jul 31

2. The public needs a better understanding of different types of AI systems, their strengths and limitations, and how to critically evaluate their output. This is a real opportunity: journalism is one of the main ways people learn about new technology.

1

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Jul 31

We find a few key recurring themes:
1. Better communication about newsrooms' methods and values, not just their use of AI tools, could help audiences understand the work of journalism.

1

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Jul 31

What do news organizations actually need to know about AI & journalism? A new briefing from @cnti.bsky.social's Global AI & Journalism Research Working Group synthesizes 80+ research studies.

cnti.org/article/aij-...

Briefing #1: AI Literacy & Communication

July 2025

cnti.org

1 2 2

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Brad Lander @bradlander.bsky.social · Jul 24

The mass starvation & killing of Palestinians in Gaza at the hands of the Israeli gov’t is a chillul hashem, a desecration of God.

Jewish leaders who criticized Zohran Mamdani for three words he doesn’t use should raise their voices loudly here as well.
forward.com/opinion/7575...

Gaza is starving. Where are the American Jewish leaders?

A fear of being labelled antisemitic still deters American Jews and Jewish institutions from taking a position that is morally correct.

forward.com

180 3.4K 9.5K

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Erin Grievances @erinbartram.bsky.social · Jul 26

Setting aside the ethical and environmental concerns of AI, I'm really offended by a bunch of elite dudes telling us our minds aren't good enough when they've had to steal everything human minds have created to create their pretend mind and it doesn't even work right!

4 490 1.4K

Reposted by Dr. Jay Barchas-Lichtenstein (they/them)

Ed Zitron @edzitron.com · Jul 27

Now Anthropic is strongly considering weekly limits on Claude Code. Users making it clear they won't accept this or pay more than $200 a month. I am confident this happens because as my analysis suggests they are losing tons of money on each user!
www.reddit.com/r/ClaudeCode...

Weekly limits are coming...
Join
I was recently invited to participate in a brief Al-moderated interview by Apthropic which I completed because they were offering a $250 Amazon gift card.
I was invited because I am supposedly "one of our most engaged Max 20x users" which was surprising to me. 1 log some pretty long hours and hit limits almost daily with CC but I wouldn't consider myself a power user at all. I don't even use mcp servers... Just a vibe coder building ai slop projects I probably have no business trying to build.
Anyways, the reason I am posting is because I was disappointed to learn that they are strongly considering or have already decided they will be implementing weekly limits.
Meaning you could, depending on your usage, max out your limits by Monday or Tuesday, even on the 20x plan and then be locked out for a week or need to upgrade or purchase additional utilization.
I voiced my concerns in the interview and let them know how I felt about that. But I haven't seen anyone else talk about this and I feel like more of you should be able to let Anthropic know if you support this or not.
I do apologize for not screenshoting some of the questions it was super early morning when I did it and wasn't really expecting them to talk about changing the limits in this manner. I can share screenshot of the email if anyone doesn't believe but I don't think it's that serious.

dolmdemon • 5h
No way in hell Im shelling out more than $200/month. That's already more than my electric bill.
I think we need to push back here. If they get CC to stop doing stupid ubercoder throw all the things at the codebase shit, we wouldn't need so much token usage wrangling it.
…今Reply&个17少
Medicaided OP • 4h
yeah $200 is my limit too I really can't justify more. Let alone more than a car payment which I imagine some of the next plans will be.
It would be cool if they could adopt a model kinda like google workspace where its lower cost for personal accounts and more for businesses. So average people can build ideas for lower costs, and businesses who are gaining real serious value from Al, can pass that Al cost (or savings) on to their customers and afford to front the cost of Al.

Better-Cause-8348 • 5h
Yep, I'll be cancelling if they do this, as I'm sure many others will as well. I've already downgraded from 20x to 10x and may even go lower, because the quality is still horrible. I'm constantly having to explain things so in-depth that I might as well do it myself.
I can barely get 10 minutes into anything before I have to stop. No, this is what I mean. No, this is what you need to do. No, that's not what I said at all. No, did I ask for that?
The crazy part to me is that it's probably costing them far more money now that they're running a more quantized version than if they'd left it the way it was and simply rate-limited or imposed stricter limitations. Now, it might take me four or five back-and-forths before I can get what I need, if I can get what I need, whereas before, it would have happened on the first try.
And yes, I get that others aren't having any issues and it's rainbows and unicorns for them still, happy for you. Don't be a dummy and assume you're experience is the only one, there are PLENTY of others on here saying the same things again and again, clearly there is an issue somewhere and it can't always be the user.
• Reply&个少
Medicaided OP • 4h
I find myself priming the convo until I have probably 20-30% context left before I let CC fix anything. Then as soon as the fix is over I need to do it again for the ne to maintain decent project alignment and remove an, over engineering

28 110 780

Dr. Jay Barchas-Lichtenstein (they/them) @drnycnerd.bsky.social · Jul 17

What do we know about how journalists and the public think about "journalism" and the people who create it?

This was fun to work on -- my colleague Sam is so, so smart!

journalistsresource.org/media/what-j...

What journalists and the public think of journalism and technology

Researchers shed light on a mismatch between how journalists and the public view journalism in the age of artificial intelligence.

journalistsresource.org

2