Gerrit De Vynck

ChatGPT started the AI race. Now its lead is looking shaky. OpenAI’s chatbot jolted Silicon Valley when it debuted three years ago, but ChatGPT’s user growth is slowing and Google’s Gemini is gaining ground.

OpenAI has already done more than any other startup to shake the dominance of Big Tech. It's still probably not enough
www.washingtonpost.com/technology/2...

Voters’ minds are hard to change. AI chatbots are surprisingly good at it. New research suggests AI chatbots can shift people’s political views more effectively than campaign ads on TV.

New: AI chatbots can change voters' minds, according to a pair of in-depth studies published just now in Science and Nature.

How they do it is interesting — and concerning. Gift link: wapo.st/49RSstP

GOP faces a familiar dilemma: What to do about Obamacare? Republicans are racing to craft health-care plans ahead of a planned vote, as the deadline looms to act on soon-to-expire Affordable Care Act subsidies.

Want the news -- the really important news -- but don't want a WaPo subscription?

That's fair.

We're trying something new: $4 for a full day of WaPo access.

Or if you just want this story, $2.

Try it with this story, loaded with scoopy bits from me, @ddiamond.bsky.social & @rbeggin.bsky.social.

OpenAI says the chats released by Raine's family don't paint a full picture, so they are providing more of his conversations with the bot to the court. Those are not public, they were filed under seal

Additionally, OpenAI argues its not liable because Raine, by using ChatGPT for self-harm, broke its terms of service

OpenAI also says Raine told ChatGPT that he reached out to trusted people in his life for help but was ignored. (this is a different take on chats released as part of the family's lawsuit)

OpenAI today said that though Raine’s death is tragic, they’re not liable because he said he had suicidal ideation for years before using ChatGPT, and he also sought info on suicide from other online resources

ChatGPT to get parental controls after teen user’s death by suicide ChatGPT maker OpenAI said it will add parental controls that can detect “acute distress,” after a lawsuit alleging the chatbot aided a teen’s death by suicide.

Raine’s family alleges that ChatGPT coached Adam to take his own life and encouraged him to hide his self-harm and plans from his family. BG: www.washingtonpost.com/technology/2...

NEW - OpenAI has responded to the lawsuit from the family of Adam Raine, the teen who committed suicide after months of talking to ChatGPT.

From The Washington Post:
The concentration of wealth among the richest Americans is unlike anything in history — and so is billionaires’ influence in politics.

www.washingtonpost.com/politics/int...

What People Actually Use ChatGPT For With Gerrit De Vynck Podcast Episode · Better Offline · 11/20/2025 · 42m

Here’s this Better Offline. I’m joined by @gerritd.bsky.social of The Washington Post to discuss what an analysis of 47,000 ChatGPT conversations can tell us about how people use the service - and how eager it is to fuel any conversation.

podcasts.apple.com/us/podcast/b...
Linktr.ee/betteroffline

These are the U.S. ships and aircraft massing off Venezuela The large-scale buildup of U.S. military forces and assets in the Caribbean suggests that the Trump administration is preparing to expand operations.

Eight warships, a nuclear-powered attack submarine, a floating base for Special Operations and an aircraft carrier on the way

www.washingtonpost.com/world/2025/1...

The Cost of the AGI Delusion By chasing superintelligence, America is falling behind in the real AI race.

“It is unlikely that superintelligence is actually imminent,” write @mchorowitz.bsky.social‬ and @laurenakahn.bsky.social. Washington should not “distort U.S. policy by dashing for something that might not exist.”

we contain multitudes Bradley!

Inside billionaire Peter Thiel’s private ‘Antichrist’ lectures - The Washington Post The billionaire Peter Thiel recently hosted a series of exclusive lectures in San Francisco. The topic? The Antichrist. Today, his arguments and what they say about an emerging mindset in Silicon Vall...

I listened to Peter Thiel's antichrist lectures, so you don't have to
www.washingtonpost.com/podcasts/pos...

Analysis | We uploaded a fake video to 8 social apps. Only one told users it wasn’t real. Facebook, TikTok and other major platforms do not use a tech industry standard touted as a way to flag fake content, tests using AI-generated videos found.

Remember when the tech cos. said they would use watermarks and metadata to help inform people what content was AI-created? Yea, it's all nonsense.
Really great/important story from @kevinschaul.bsky.social www.washingtonpost.com/technology/2...

"While on the phone with 911, Rinderknecht allegedly typed a question into ChatGPT: 'Are you at fault if a fire is lift because of your cigarettes'"

A Venezuelan singer won asylum. ICE detained him for two more months. The singer known online as Davicito59 won his asylum case in July. Still, ICE kept him in custody for two more months — a holdup several experts called unlawful.

When a judge told Claudio David Balcane González “Welcome to the United States,” he sobbed. He had finally been granted asylum.

Then ICE detained him for 63 more days @marialuisapaul.bsky.social reports

www.washingtonpost.com/immigration/...

Do you abstain from AI? I'd love to hear about why you avoid it -- and how hard that might be getting. Send me a DM or Signal @lisabonos.18

Reality Defender bypassed Sora 2's deepfake impersonation safeguards in under 24 hours. Here's what happened: OpenAI just launched Sora 2 with a "Cameo" feature requiring identity v... Reality Defender bypassed Sora 2's deepfake impersonation safeguards in under 24 hours. Here's what happened: OpenAI just launched Sora 2 with a "Cameo" feature requiring identity verification to pre...

Deepfake detection co. Reality Defender was able to use videos of real people to get their cameo uploaded to Sora. OpenAI's system didn't catch it
www.linkedin.com/posts/benpco...

Everything is fake on Silicon Valley’s hottest new social network The new Sora social app from ChatGPT maker OpenAI encourages users to upload video of their face so their likeness can be put into AI-generated clips.

A few hours on Sora we made
- video of an ICE raid on Sesame St.
- JFK saying things he never said
- Sam Altman robbing a bank and dressed as a Nazi general

www.washingtonpost.com/technology/2...

ICE to Buy Tool that Tracks Locations of Hundreds of Millions of Phones Every Day Documents show that ICE has gone back on its decision to not use location data remotely harvested from peoples' phones. The database is updated every day with billions of pieces of location data.

New: ICE has bought a tool that tracks the locations of hundreds of millions of phones globally, updates every single day. Usually harvested from apps and advertising

www.404media.co/ice-to-buy-t...

Saudi comedy festival draws big names and backlash Human rights advocates and some comics say the festival, at which Bill Burr, Louis C.K., Dave Chappelle and Kevin Hart are performing, whitewashes abuses.

Wrote about the Riyadh Comedy Festival, where the setup of tell-it-like-it-is comedians who riff on cultural reservations around taboo subjects meets the awkward punchline of severe restrictions on speech.
wapo.st/46vPycf

Police are drowning in data. Could a chatbot help? Start-up Longeye offers an AI chatbot for police designed to help them spot clues and patterns in digital evidence such as phone recordings or online chatlogs.

Thanks to tower dumps, Gmail warrants, social media, police are drowning in data. A new startup is selling AI to help them parse it. www.washingtonpost.com/technology/2...

the fact they have the ability to allow opt out means they're capable of matching rights holders to works, so so much for their theory that the machine is dumb and they can't tell what it is generating.

New @washingtonpost.com supercut:

Donald Trump and his allies keep blaming the left for "fascist" political rhetoric that they have also used

yea we have a red light and blue light in the booth and Im facing the red one. Print reporters have no business making videos without help

the lighting makes me look like

YouTube to pay $24.5 million to settle Trump lawsuit over post-Jan. 6 suspension YouTube will pay $24.5 million to settle a 2021 lawsuit alleging it unfairly blocked President Donald Trump’s account after the Jan 6. Capitol attacks.

YouTube's settlement will go towards paying for the new White House ballroom (seriously)

www.washingtonpost.com/technology/2...