Tom Simonite

Just ran some evals on Claude Sonnet 4.5. It's better than 4 on some but worse on a lot. LLM progress is so weird. You really gotta test this stuff on what you care about.

CHOTINER: And you thought it would be funny to do a fake interview by me?
ME: A lot of people have, so
CHOTINER: You were looking for engagement?
ME: Well, to make someone laugh I guess
CHOTINER: How many likes and reposts did it get?
ME: That's, uh, not really the metric...look can we start over

Hegseth orders rare, urgent meeting of hundreds of generals, admirals The Pentagon has summoned military officials from around the world for a gathering in Virginia. Top generals and their staffs don’t know the reason for the meeting.

Pete Hegseth has summoned every general and admiral in the US military around the world to an *in person* meeting next week. And he’s given no reason.

That’s basically unprecedented. There’s lots of reasons you don’t do it, too – not least that it’s bound to make the rest of the world twitchy.

Trump’s media enforcer is relishing his Jimmy Kimmel moment Brendan Carr has become the personification of the president’s crusade against the mainstream media, and he’s just getting started.

When he named Brendan Carr as his FCC chair, Trump called him a "warrior for free speech."

We wrote about how he has become a warrior against it.

Gift link: wapo.st/4pxe2t0

OpenAI won’t say whose content trained its video tool. We found some clues. Tests by The Washington Post show OpenAI’s video generator Sora can mimic Netflix shows, TikTok clips and movies, suggesting it used versions of the originals.

By @kevinschaul.bsky.social & @nitasha.bsky.social Gift link: wapo.st/3KeqLR0

Sora can mimic Netflix shows, movies, TikTok videos and Twitch streams, suggesting it was trained on versions of the originals. Netflix said it didn't provide data to OpenAI. Twitch said it doesn't have a partnership w/ OpenAI.

OpenAI won't say whose content it used to train its video generator Sora. Here are some clues.
Gift link: wapo.st/3KeqLR0

Trump, allies seek to punish speech they dislike following Kirk killing The White House-directed campaign, which is likely to face legal challenges, defies past GOP rhetoric condemning the label “hate speech” as a pretext for political persecution.

Top story on @washingtonpost.com this morning: www.washingtonpost.com/politics/202...

Who's using ChatGPT?

A teen contemplating suicide turned to a chatbot. Is it liable for her death? A lawsuit filed by the parents of 13-year-old Juliana Peralta against Character AI is the latest to allege a chatbot contributed to a teen’s death by suicide.

"The complaint filed by Juliana’s parents is the third high-profile case in the past year brought by a U.S. family alleging that an AI chatbot contributed to a teen’s death by suicide." www.washingtonpost.com/technology/2...

A teen contemplating suicide turned to a chatbot. Is it liable for her death? A lawsuit filed by the parents of 13-year-old Juliana Peralta against Character AI is the latest to allege a chatbot contributed to a teen’s death by suicide.

A new wrongful death lawsuit alleges that Character AI and Google are liable for the death by suicide of a 13-year-old girl in Colorado, Juliana Peralta

This is the third wrongful death claim against a popular AI app for a teen's death by suicide this year www.washingtonpost.com/technology/2...

‘My kid has seen this. Now what?’: Parents reel as Charlie Kirk video goes viral Gory social videos of Charlie Kirk’s shooting have inundated children, leaving parents feeling powerless and disturbed.

Visceral closeup videos of Charlie Kirk getting shot in the neck were everywhere on social media. A lot of parents were not pleased that mainstream social apps’ algorithms showed it to their kids.

Story today by @tatumhunter.bsky.social & me (gift link): wapo.st/4n6oBSi

I asked Bluesky how it's moderating graphic videos of the Charlie Kirk assassination, as well as responses that glorify violence or call for retribution.

From a Bluesky spokesperson: "We're suspending accounts that are encouraging violence; we are taking down close-up videos of the event."

Business Insider yanked 40 essays with suspect bylines. Are they related? “Margaux Blanchard” appears to be one dubious element of a broader scheme to peddle bogus articles to a number of publications.

Business Insider yanked 40 essays with suspect bylines. A @washingtonpost.com probe into these mass retractions found key links between the bogus writer “Margaux Blanchard” and another individual, suggesting a broader scheme. Here’s what I found: www.washingtonpost.com/business/202...

🚨Journalism job alert🚨

We're hiring an investigative reporter to join my team at @wired.com. We're looking for someone who has both traditional and non-traditional reporting skills (coding/data work/OSINT, etc). Fit the bill? Come work with me! condenast.wd5.myworkdayjobs.com/CondeCareers...

Making cash off ‘AI slop’: The surreal video business taking over the web A mad rush of creators is using AI video tools to flood the web — and turn a profit — with videos that can seem remarkably real.

New: My deep dive into the business of AI slop. wapo.st/4mFECxC

Analysis | The AI industry is awash in hype, hyperbole and horrible charts Just checking: 69.1 is a larger number than 30.8, right?

A horrifying investigation into the chart crime scene that is the AI industry by @shiraovide.bsky.social, with top notch illustration from the great Elena Lacey

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

Column | Zuckerberg fired the fact-checkers. We tested their replacement. Our tech columnist drafted 65 community notes, Meta’s new crowdsourced system to fight falsehoods. It failed to make a dent.

Meta fired its fact-checkers, citing concerns of liberal bias, and replaced them with a version of X's "Community Notes." How's that going?

Well, my coworker @geoffreyfowler.bsky.social proposed 65 fact-checks debunking false posts... and only 3 got approved. www.washingtonpost.com/technology/2...

A New Era for WIRED—That Starts With You As of today, there are even more ways to enjoy WIRED’s journalism, including premium newsletters and livestream Q&As. Join us.

Wired rules. Just an incredible group of journalists who are completely meeting the moment. And you can subscribe for $2 a month, which is insane www.wired.com/story/a-new-...

Maybe those robots will be ... trains

Analysis | Trump’s ‘Artificial Intelligence Action Plan’ is already stirring debate Critics unveiled their own “People’s AI Action Plan” ahead of new executive orders expected to address energy and “woke” AI.

New: As Trump prepares to unveil his "AI Action Plan," more than 80 labor, environmental, consumer and tech watchdog groups have signed on to a "People's AI Action Plan" that calls for regulations that put ordinary people's interests over those of the industry. www.washingtonpost.com/politics/202...

The UK expects AGI in four years. Why doesn’t it act like it? Trust me bro Westminster edition

UK government minister: AGI by 2029
Also UK government: HS2 high speed rail project won't be complete until sometime after 2033
www.learningfromexamples.com/p/the-uk-exp...

Tech billionaire Trump adviser Marc Andreesen says universities will ‘pay the price’ for DEI Tech investor and Trump adviser Marc Andreessen in a private group chat attacked colleges for promoting diversity, saying they had discriminated against “native born kids.”

New from me. Hope you’ll give it a read www.washingtonpost.com/technology/2...

Grok’s Nazi tirade sparks debate: Who’s to blame when AI spews hate? Antisemitic outbursts from the chatbot promoted by Elon Musk shows how AI companies often face minimal consequences when their projects go rogue.

New: If a person spewed Nazi garbage like Grok, they'd get fired in an instant. Elon Musk's chatbot just got a promotion. Why don't companies face any consequences for AI disasters? There might be a few ways to turn the tide ... wapo.st/4095Rbh

Google co-founder Sergey Brin calls U.N. ‘transparently antisemitic’ after report on tech firms and Gaza Google co-founder Sergey Brin called the United Nations “transparently antisemitic” in a message on a company forum, after a U.N. report alleged Google profited from Israel’s war in Gaza.

Google co-founder Sergey Brin called the U.N. “transparently antisemitic” in an internal forum for employees.

His comments came in response to a U.N. report that alleged technology firms profited from “the genocide carried out by Israel” in Gaza.

Google co-founder Sergey Brin calls U.N. ‘transparently antisemitic’ after report on tech firms and Gaza Google co-founder Sergey Brin called the United Nations “transparently antisemitic” in a message on a company forum, after a U.N. report alleged Google profited from Israel’s war in Gaza.

new from me and @gerritd.bsky.social wapo.st/4lg3pIh (gift link)

on an internal forum for Google Deepmind employees, Sergey Brin called the UN a “transparently antisemitic organization” after a new UN report accused tech companies of profiting from genocide in Gaza

“My comments came in response to an internal discussion that was citing a plainly biased and misleading report,” Brin said in a statement provided to The Post.

Google co-founder Sergey Brin calls U.N. ‘transparently antisemitic’ after report on tech firms and Gaza Google co-founder Sergey Brin called the United Nations “transparently antisemitic” in a message on a company forum, after a U.N. report alleged Google profited from Israel’s war in Gaza.

Scoop: Google cofounder Sergey Brin branded the United Nations "transparently antisemitic" on an internal company message board, in comments that left some staff confused and upset www.washingtonpost.com/technology/2... via @gerritd.bsky.social @nitasha.bsky.social

In dramatic reversal, Senate votes to kill AI-law moratorium A GOP-led bid to stop states from regulating AI collapsed after a deal to save it fell through.

In a stunning reversal, the Senate voted 99-1 this morning to strip from Trump's big bill a 10-year moratorium on state-level AI regulations.

Gift link to my story on how it happened and who's celebrating: wapo.st/3TOyiaG