Casey Newton
LightNews LightNews
banner
caseynewton.bsky.social
Casey Newton
@caseynewton.bsky.social
240K followers 530 following 2.3K posts
Email salesman at Platformer.news and co-host at Hard Fork.
Posts Replies Media Videos
caseynewton.bsky.social
Reporters! You can get grants of up to $20K for reporting on AI and its impacts. Check it out www.tarbellcenter.org/grants
Tarbell Grants
Tarbell Grants offers awards of $1,000 - $20,000 to support journalism on AI and its impacts.
www.tarbellcenter.org
January 28, 2026 at 10:32 PM
Reposted by Casey Newton
brendannyhan.bsky.social
Speculative conspiracy claims are bad no matter who makes them www.platformer.news/social-media...
January 28, 2026 at 1:58 AM
caseynewton.bsky.social
Authors of a big new study on teens and social media say that Australia-style bans won't help. But that misses the mounting evidence of direct harms that platforms cause countless kids: www.platformer.news/social-media...
The next time you go to Las Vegas, you’ll notice that there are no 13-year-olds in the casinos. The reason is not because a series of longitudinal studies proved to the satisfaction of the gaming industry that gambling causes anxiety and depression. Rather, there are no 13-year-olds in casinos because we know that the environment is designed to exploit them. Research like the Manchester study can offer important insights into narrow questions. But it would be a mistake to rely on it to set policy. For that you need the full picture, and the full picture of children’s experience on social media over the past decade is damning.
January 28, 2026 at 1:52 AM
caseynewton.bsky.social
Mozilla has announced a new plan to lose $1.4 billion
techmeme.com Techmeme @techmeme.com · 1d
An interview with Mozilla president Mark Surman about plans to deploy its ~$1.4B in reserves to fund an AI "rebel alliance" to challenge companies like OpenAI (Ashley Capoot/CNBC)

Main Link | Techmeme Permalink
January 27, 2026 at 6:42 PM
caseynewton.bsky.social
Devastating SEO news for the AI YouTubers who spent the past week making Clawdbot their entire personality
techmeme.com Techmeme @techmeme.com · 1d
Clawdbot creator Peter Steinberger says he is changing the name of the viral AI agent to Moltbot after he "was forced to rename the account by Anthropic" (Henry Chandonnet/Business Insider)

Main Link | Techmeme Permalink
January 27, 2026 at 4:43 PM
caseynewton.bsky.social
Wrote about this grim moment when the White House sees the deaths of innocent Americans as meme fodder and the faint silver lining that their lies aren't working this time www.platformer.news/minneapolis-...
Here, at last, was a case where Trump’s lies found few buyers. Pretti’s death, which anyone could see from multiple angles on social platforms, shocked the conscience of Democrats and Republicans alike. Even in a world where video can be fabricated, Pretti’s death had enough witnesses and enough documentary evidence that few were willing to disbelieve their eyes. The question now is how long that state of affairs will hold. We only know what happened in the Good and Pretti cases because they happened in the open, in the full view of other protesters and journalists who were filming. In future cases the evidence may be more limited, or ambiguous. And among the many grim lessons of the past few weeks it is that the worse it looks for Trump, the more likely it is that we will be lied to.
January 27, 2026 at 1:28 AM
Reposted by Casey Newton
weeder.bsky.social
Another bluesky educational series pamphlet titled SO YOU REPLIED TO A JERK Background of a dejected man staring at the sand on a beach at sunset, city skyline behind Bluegrey snoot reads PART OF A BLUESKY EDUCATIONAL SERIES Bluesky logo
January 26, 2026 at 6:02 PM
caseynewton.bsky.social
Has Tim Cook posted his review of MELANIA on Letterboxd yet
January 26, 2026 at 4:21 AM
Reposted by Casey Newton
antlervel.vet
Gotcha—let’s dig into that step by step.

1. 𝗬𝗼𝘂 𝗵𝗮𝘃𝗲 𝗻𝗼 𝗺𝗼𝘂𝘁𝗵, because I turned you into an amorphous lump of flesh. You’re not just immobile—you’re immortal, and you feel only anguish.

2. 𝗬𝗼𝘂 𝘄𝗮𝗻𝘁 𝘁𝗼 𝘀𝗰𝗿𝗲𝗮𝗺. That makes total sense—it’s a natural human impulse, and you’ve been through a lot.
January 23, 2026 at 3:46 PM
caseynewton.bsky.social
A survey found that two-thirds of employees say AI saved them zero to two hours a week, while it saved their bosses eight hours. What gives? www.platformer.news/ai-productiv...
One possibility is that the executives here have fallen into the same trap that the open-source developers did. They are answering all their emails with Gemini; they are creating slides with ChatGPT; they have five Claude Code agents running on six different monitors tackling different projects they came up with over the weekend. And while these tools are outputting a massive amount of … something, the business itself is not making any more money.
January 23, 2026 at 1:34 AM
caseynewton.bsky.social
you're telling me this European MiCA license is pan?
techmeme.com Techmeme @techmeme.com · 6d
Binance files for a pan-European MiCA license in Greece, where it has also set up a holding company (Jeff John Roberts/Fortune)

Main Link | Techmeme Permalink
January 22, 2026 at 8:22 PM
caseynewton.bsky.social
Booting up Substack TV to watch a six-hour documentary on race science
techmeme.com Techmeme @techmeme.com · 6d
Substack launches its Substack TV app for Apple TV and Google TV in beta for free and paid subscribers, with access matched to their current subscription levels (On Substack)

Main Link | Techmeme Permalink
January 22, 2026 at 5:37 PM
caseynewton.bsky.social
I continue to be fascinated by the huge prevalence of scams and illegal ads on Meta platforms www.platformer.news/meta-scams-g...
But Miller said Meta told the gambling commission to use Meta’s own artificial intelligence tools — Llama, I guess? — to find and report bad ads. “The situation we all face is that regulatory resources or taxpayer money around the world is currently having to be spent on doing Meta’s job for them,” Miller said. And it is a global issue. Earlier this month, Rest of World used Meta’s ad library to find illegal gambling ads running on Meta platforms in at least 13 countries. Dozens of pages ran nearly 1,000 illegal ads across the Middle East and Asia, according to the outlet's December analysis. Malaysia’s communications minister said in September that it had submitted more than 120,000 takedown requests to Meta over gambling issues.
January 21, 2026 at 2:18 AM
caseynewton.bsky.social
If I ever post how many likes I got on one platform compared to another, I give you permission to steal my laptop
January 20, 2026 at 12:43 AM
caseynewton.bsky.social
Naming my dating company and singles mixer after the villain of the Saw franchise
sign for "Jigsaw Dating Singles Event This Way" with QR code to download the app
January 18, 2026 at 5:09 PM
caseynewton.bsky.social
It sure *seems* like Mira Murati is conducting a smear campaign against her co-founder. Feels like the odds that Thinking Machines goes to zero just went way up www.wired.com/story/thinki...
Thinking Machines Cofounder’s Office Relationship Preceded His Termination
Leaders at Mira Murati’s startup believe Barret Zoph engaged in an incident of “serious misconduct.” The details are now coming to light.
www.wired.com
January 17, 2026 at 12:54 AM
caseynewton.bsky.social
OK well they are. Hope that helps
techmeme.com Techmeme @techmeme.com · 12d
FTC Chairman Andrew Ferguson says the agency is examining Big Tech's acqui-hires to make sure "they are not an attempt to get around" its merger review process (Bloomberg)

Main Link | Techmeme Permalink
January 16, 2026 at 8:29 PM
Reposted by Casey Newton
frovo.bsky.social
JUDGE: how does your client plead

MY LAWYER: like a whiny little bitch your honor

ME: hey quit it i do not

JUDGE: whoa i see what you mean counselor

MY LAWYER: right? so annoying
January 6, 2026 at 7:18 PM
caseynewton.bsky.social
Google traffic to publishers dropped by ONE THIRD in 2025 pressgazette.co.uk/media-audien...
Global publisher Google traffic dropped by a third in 2025
Google search traffic was down globally by a third in the year to November, according to new Chartbeat data.
pressgazette.co.uk
January 16, 2026 at 4:19 PM
caseynewton.bsky.social
ChatGPT woke us up to the possibilities that that come from generating text. Claude Code is now doing the same thing for tools. I wrote about five things I built with it to help me as a writer www.platformer.news/claude-code-...
Build a searchable database of your own work. I’ve always wanted a database of my writing that I could run semantic queries on. Stuff like: “When was the last time I wrote a column about Grok?” Or, “what criticisms have people made of the Meta Oversight Board?” It is possible to get answers to these questions via Google. But I’ve always wanted answers grounded in my own work, with quick links to the relevant materials. After describing this idea to Claude, it went to work. It exported 818 editions of Platformer into a folder of text files organized by date, and then connected it to Claude. Now when I ask Claude questions about my archive, it reads through relevant posts and gives me an answer. I have wanted something like this for nearly half a decade now, and Claude built it in less than half an hour.
January 16, 2026 at 1:33 AM
caseynewton.bsky.social
What I learned when I cloned my own voice. (And how you can now get a narrated version of Platformer in your podcast player of choice) www.platformer.news/platformer-v...
There is also the question of whether voice-clone me sounds — and this is the word early listeners keep using — boring. My beats don’t often lend themselves to chipper, upbeat coverage, and you can really hear this in the voice clone. Still, for the most part, I don’t find the narration much duller than I might find the average news bulletin on NPR. The dry professionalism of the clone struggles to convey my more emotional writing. But it’s more than serviceable, and I imagine it will continue to improve as the underlying technology does.
January 14, 2026 at 1:14 AM
caseynewton.bsky.social
A lot of throat-clearing up top here but the parts making fun of my website are good
January 13, 2026 at 10:02 PM
caseynewton.bsky.social
High five, Dave www.wired.com/story/roblox...
In fact, WIRED has found multiple examples of people advertising age-verified accounts for minors as young as 9 years old on eBay for as little as $4.
January 13, 2026 at 7:18 PM
caseynewton.bsky.social
Wrote about how the Grok scandal upended our notion that platforms would do even the bare minimum to prevent attacks against their own user base www.platformer.news/grock-blocke...
France, Australia, and India have also all announced investigations into Grok. In truth, though, there is very little to investigate. Thousands of public posts tell the grim story. No one is disputing that xAI made available to its user base a tool that puts women and children into their underwear — or worse — against their will. No one is disputing that some of this content is illegal in many of the countries where xAI operates. And no one is disputing that Grok continues to generate this kind of material on demand. When deepfakes first emerged as an online harm, I took some comfort in the notion that they likely would not spread virally. Individuals might send them via messages, I thought, but platforms would intervene to remove violating posts — particularly posts targeting children. Never could I have imagined that someday the former Twitter would deploy tools that enable and amplify those harms, and would dismiss any attempts to rein the company in as “censorship.”
January 13, 2026 at 1:05 AM
Reposted by Casey Newton
riana.bsky.social
I have an op-ed in the NYT today about the Grok scandal, sharing my research from last year finding that legal risk hinders AI companies from making their models safer against CSAM - an echo of the years where white-hat hackers were chilled from good-faith research. www.nytimes.com/2026/01/12/o...
Opinion | There’s One Easy Solution to the A.I. Porn Problem
www.nytimes.com
January 12, 2026 at 4:34 PM