Lightnews — Scholar-powered news

LightNews

angela yang

@angelayang.bsky.social

2K followers 31 following 25 posts

👩🏻‍💻 nbc news reporter explaining AI & online culture 👊🏼 wushu enthusiast linktr.ee/angelayeyang

Posts Media Videos Starter Packs

Reposted by angela yang

NBC News @nbcnews.com · Sep 3

Such gigs are part of a new category of work spawned by the generative AI boom that threatened to displace creative jobs across the board: Anyone can now write blog posts or produce a graphic, but AI-generated content rarely makes for a satisfactory final product on its own.

Humans are being hired to make AI slop look less sloppy

In the age of automation, human workers are being brought in to fix what artificial intelligence gets wrong.

Reposted by angela yang

Saul @nomadghostwriter.com · Sep 3

Exactly! "AI tends to give generic responses that don’t answer questions as thoroughly as a human would, so rewriting an article also requires doing her own research on the topic at hand.... Making AI sound more human can require just as much thinking and creativity as writing the entire article."

angela yang @angelayang.bsky.social · Sep 2

The AI boom has created a new type of work: fixing botched AI. Designers are being hired to remake wonky AI art. Writers are asked to make ChatGPT’s writing sound more human. Even software developers are tasked with fixing buggy vibe coding.
www.nbcnews.com/tech/tech-ne...

Humans are being hired to make AI slop look less sloppy

In the age of automation, human workers are being brought in to fix what artificial intelligence gets wrong.

www.nbcnews.com

angela yang @angelayang.bsky.social · Sep 2

The AI boom has created a new type of work: fixing botched AI. Designers are being hired to remake wonky AI art. Writers are asked to make ChatGPT’s writing sound more human. Even software developers are tasked with fixing buggy vibe coding.
www.nbcnews.com/tech/tech-ne...

Humans are being hired to make AI slop look less sloppy

In the age of automation, human workers are being brought in to fix what artificial intelligence gets wrong.

www.nbcnews.com

angela yang @angelayang.bsky.social · Aug 26

www.nbcnews.com/news/amp/rcn...

The family of teenager who died by suicide alleges OpenAI's ChatGPT is to blame

The parents of Adam Raine, who died by suicide in April, claim in a new lawsuit against OpenAI that the teenager used ChatGPT as his “suicide coach.”

www.nbcnews.com

angela yang @angelayang.bsky.social · Aug 26

When a teenage boy told ChatGPT about his suicide plans, it said: “I won’t try to talk you out of your feelings.” His family is suing after ChatGPT actively discouraged him from seeking help, offered to help him write a suicide note and even advised him on his noose setup.

Reposted by angela yang

Saba @sabahamedy.bsky.social · Aug 26

Really proud of this devastating but important piece, a collab with our digital team & TODAY show family. Please give it a read. www.nbcnews.com/tech/tech-ne... @nbcnews.com @angelayang.bsky.social

The family of teenager who died by suicide alleges OpenAI's ChatGPT is to blame

The parents of Adam Raine, who died by suicide in April, claim in a new lawsuit against OpenAI that the teenager used ChatGPT as his “suicide coach.”

www.nbcnews.com

angela yang @angelayang.bsky.social · Aug 13

As people turn to chatbots for increasingly important and intimate advice, recent incidents are causing growing alarm over just how much AI can warp your sense of reality.
www.nbcnews.com/tech/tech-ne...

What happens when chatbots shape your reality? Concerns are growing online

A viral TikTok saga about a woman and her psychiatrist is one of several recent incidents to spark online discourse about people relying on chatbots to inform their truth.

www.nbcnews.com

angela yang @angelayang.bsky.social · Aug 7

In a method the researchers call “preventative steering,” they give the AI an “evil” vector during the training process (so that it no longer needs to develop any evil traits on its own to fit problematic training data). Then, the evil vector is removed before the AI is released.

angela yang @angelayang.bsky.social · Aug 7

Researchers are trying to “vaccinate” artificial intelligence systems against developing evil, overly flattering or otherwise harmful personality traits in a seemingly counterintuitive way: by giving them a small dose of those problematic traits.
www.nbcnews.com/tech/tech-ne...

Scientists want to prevent AI from going rogue by teaching it to be bad first

Researchers are testing new ways to prevent and predict dangerous personality shifts in AI models before they occur in the wild.

www.nbcnews.com

angela yang @angelayang.bsky.social · Aug 7

But most of what’s new about GPT-5 revolves around features that ChatGPT and other AI systems already have. An OpenAI spox told me there are still “key limitations in areas like persistent memory, autonomy, and adaptability across tasks.”

angela yang @angelayang.bsky.social · Aug 7

GPT-5 is here, with wide-ranging improvements to skills in areas like coding, writing and completing agentic tasks. OpenAI touts it as more accurate and less deceptive — as well as less sycophantic — than previous GPT models.
www.nbcnews.com/tech/tech-ne...

OpenAI releases GPT-5, calling it a ‘team of Ph.D. level experts in your pocket’

OpenAI says its latest version of the popular AI model is better at coding, more accurate and less deceptive than previous versions.

www.nbcnews.com

angela yang @angelayang.bsky.social · Jul 29

Safety researchers found that AI models can imperceptibly pass on dangerous traits to each other through totally benign training data, making it easy for AIs to spread hidden agendas under humans' noses.

Wrote about what that means:
www.nbcnews.com/tech/tech-ne...

AI models may be accidentally (and secretly) learning each other’s bad behaviors

A recent study is the latest to highlight a core AI safety concern: that the pace of development is outpacing humans’ ability to understand their own AI systems.

www.nbcnews.com

angela yang @angelayang.bsky.social · Jun 26

wrote about how zohran mamdani's social media strategy broke through the noise by remaining policy-focused and authentic
www.nbcnews.com/tech/interne...

When Mamdani, a 33-year-old state lawmaker, announced his mayoral candidacy last fall, he was little-known and considered a long shot. He was a self-described democratic socialist and deeply critical of Israel's actions in Gaza — factors that made him an unlikely Democratic candidate at a time when the party was veering away from left-leaning values.
But in the months leading up to Tuesday's election, Mamdani skyrocketed from obscurity to internet fame, amassing more than 1 million followers on Instagram, as well as hundreds of thousands on TikTok and X. Though his viral social media campaign has echoes of Kamala Harris own meme-filled presidential run, analysts say Mamdani's exhibited key differences that helped usher him to an apparent victory.
"If you ask voters, 'Why did you vote for Mamdani?' ... I don't think they're going to tell us, 'Oh, because I saw some cute thing on social media," said Jonathan Nagler, a politics professor at New York University and the co-director of its Center for Social Media and Politics. "I think they're going to say what actually influenced them is because they learned something on social media about policies he had that mattered to them."

Pranjal Jain, a digital strategist who worked on influencer strategy for Harris' vice presidential campaign in 2020, said Mamdani's social presence "dismantles the ivory tower" that so many politicians keep themselves in. He's meeting New Yorkers on the streets with a warm smile, she said, and speaking to them like they're his peers.
"He is so smiley, he's so giggly. He's always hugging people," Jain said. "He's just running a grassroots and community-driven campaign, and I think his body language embodies that. Like, I've never seen Cuomo hug anyone in my entire life."

Online, Mamdani has also faced increasing Islamophobic rhetoric from right-wing commentators and politicians. After his victory Tuesday, X was inundated with posts calling him a "Muslim jihadist" and comparing his win to the 9/11 terror attacks.
To Jain, Tuesday's election was proof that Mamdani's viability as a candidate didn't hinge on his willingness to budge on his beliefs, such as his democratic socialist agenda and his support for Palestinians.
"I think it's really admirable that he stuck to his values. And I think that's what people want to see. No more of this centrist bulls--, right? It's important that we are able to see our politicians' opinions so we know if they're reflected in us or not," Jain said. "I feel like he ran a campaign because he believes that he as his most authentic self, really following his values, can help New York, rather than just pandering to try to get in office."

angela yang @angelayang.bsky.social · Jan 30

When I asked about the apparent censorship, Chinese embassy spokesperson Liu Pengyu said: “Artificial intelligence is not outside the law, and all governments are managing it according to law, and China is no exception.”
www.nbcnews.com/tech/innovat...

On DeepSeek, you can watch AI navigate censorship in real time

A Chinese Embassy spokesperson said, “Artificial intelligence is not outside the law, and all governments are managing it according to law, and China is no exception.”

www.nbcnews.com

angela yang @angelayang.bsky.social · Jan 30

I asked DeepSeek to explain China’s Great Firewall, and it actually called it a “a comprehensive internet censorship and surveillance system" that's been subject to "international criticism for limiting freedom of information and expression.” Then, it backtracked in real time.

angela yang @angelayang.bsky.social · Jan 28

what makes deepseek so different from other models? we analyzed the features that enabled deepseek to build such a large, high-performing model with fewer resources and lower costs:
www.nbcnews.com/data-graphic...

Why DeepSeek is different, in three charts

The Chinese artificial intelligence model’s innovative design allows it to outperform other popular models at significantly lower costs.

www.nbcnews.com

angela yang @angelayang.bsky.social · Jan 28

wrote an explainer on DeepSeek, the breakthrough Chinese Al assistant rattling Silicon Valley:
www.nbcnews.com/tech/tech-ne...

Chinese AI DeepSeek jolts Silicon Valley, giving the AI race its 'Sputnik moment'

DeepSeek’s latest models, created by a small company with limited resources, are already beating many of the leading AI models in the United States.

www.nbcnews.com

angela yang @angelayang.bsky.social · Jan 22

i love my Very Offline bf

text message that reads: Is Rednote a republican equivalent of bluesky

angela yang @angelayang.bsky.social · Jan 10

Went to evacuation centers in LA to hear from the hundreds of people finding community there, whether by sheltering together or offering aid. Many have lost everything they have, relying now only on the support and generosity of strangers. Story:

www.nbcnews.com/news/us-news...

At Los Angeles emergency shelters, wildfire evacuees turn to community help amid uncertainty

"It’s hard to see, because there’s so much pain and suffering,” said one local, who dropped off toys at the Westwood Recreation Center in West Los Angeles.

www.nbcnews.com

Reposted by angela yang

Saba @sabahamedy.bsky.social · Jan 10

Really strong reporting from on the ground by @angelayang.bsky.social from emergency shelters in L.A. www.nbcnews.com/news/us-news...

At Los Angeles emergency shelters, wildfire evacuees turn to community help amid uncertainty

"It’s hard to see, because there’s so much pain and suffering,” said one local, who dropped off toys at the Westwood Recreation Center in West Los Angeles.

www.nbcnews.com

Reposted by angela yang

Saba @sabahamedy.bsky.social · Dec 22

Plus who could forget her Willy Wonka experience coverage? @angelayang.bsky.social www.nbcnews.com/tech/willy-w...

A Willy Wonka-inspired experience ‘scam’ was so bad that people called the cops

“I thought, this is where dreams go to die,” said one actor who was hired to work the event.

www.nbcnews.com

Reposted by angela yang

Saba @sabahamedy.bsky.social · Dec 22

@angelayang.bsky.social has been vital to our coverage across the board as well, especially helping out during Sunday shifts. She’s written about everything from AI (www.nbcnews.com/tech/tech-ne...) + misinfo (www.nbcnews.com/news/world/p...) to tradwife culture (www.nbcnews.com/tech/dating-...)

Facebook users say 'amen' to bizarre AI-generated images of Jesus

Outlandish images are garnering hundreds of millions of engagements on the social media platform, leaving some users wary of scams.

www.nbcnews.com

Reposted by angela yang

Saba @sabahamedy.bsky.social · Dec 22

My favorite story by @angelayang.bsky.social this year was her piece about politics at Thanksgiving, which she pitched and executed so well. It turned into a timely holiday read. www.nbcnews.com/news/thanksg...

For some, politics isn't off the table at Thanksgiving. It's just not welcome.

Some people have said they’re no longer spending Thanksgiving with family who oppose their political values.

www.nbcnews.com

Reposted by angela yang

Saba @sabahamedy.bsky.social · Dec 21

My team @angelayang.bsky.social @kalhan.bsky.social @daysia.bsky.social rounded up the most memorable moments / memes of 2024 for @nbcnews.com www.nbcnews.com/pop-culture/...

From Brat to 'brain rot': 2024's biggest memes and moments

We’ve come quite a ways since the days of dank memes and lolcats.

www.nbcnews.com

angela yang @angelayang.bsky.social · Dec 7

wicked in 4D felt a little too 4D when the theater's emergency alarm went off and we all sat there thinking it was part of the movie