Kylie Robison

@kylierobison

writer @ Core Memory • my two accomplishments here are that i invited AOC and i started the first-ever Hellthread • signal: kylie.111

406 Following55,847 Followers2,513 Posts

YouTube Bluesky

Is this your profile?Claim it and kylierobison.dirty.social is yours.Claim profile

Kylie Robison@kylierobison· Mar 16

My first video for Core Memory is live on YouTube! www.youtube.com/watch?v=yNo_...

Kylie Robison@kylierobison· Feb 20

i wrote about a really strange press briefing i attended for perplexity www.corememory.com/p/perplexity...

A dozen or so journalists sat around a white oak table. Most of them were stiff and quiet. A few others were more active, clacking away on their keyboards in a bid to note every word tumbling out of the executives’ mouths. Perplexity powers curiosity glowed on a projector screen in a pleasant serif style.

The press wasn’t useful to Perplexity, one of the executives explained. The company doesn’t want to waste time on podcasts. They’re too focused on shipping features to engage in insular, navel-gazing AI drama. We don’t have fancy houses with wonderful dinners to impress you, he added. The chairs creaked in response.

Sometime during this executive’s spiel—maybe after his Epstein joke failed to get a reaction and he anxiously asked if it was recorded—I wrote that Perplexity seemed to have forgotten what it was trying to be.

Kylie Robison@kylierobison· Jan 21

In this week's Core Memory podcast, we chat with Jerry Tworek, the former VP of RL at OpenAI who left this month after 7 years with the startup. Available now for subscribers, coming soon to streaming services for all. www.corememory.com/p/he-left-op...

Kylie Robison@kylierobison· Jan 15

Aaaand Barret + two others from TML poached back to OAI according to Fidji Simo:

Excited to welcome Barret Zoph, Luke Metz, and Sam Schoenholz back to OpenAI! This has been in the works for several weeks, and we’re thrilled to have them join the team. Barret will report to me; Luke and Sam will report into Barret. More to come on what they’ll focus on soon!

Kylie Robison@kylierobison· Jan 14

Here's Mira Murati confirming my scoop: x.com/miramurati/s...

We have parted ways with Barret Zoph.

Soumith Chintala will be the new CTO of Thinking Machines. He is a brilliant and seasoned leader who has made important contributions to the AI field for over a decade, and he’s been a major contributor to our team. We could not be more excited to have him take on this new responsibility.

Kylie Robison@kylierobison· Jan 14

For my debut as co-host of Core Memory's podcast, we sat down with Jasmine Sun to chat about all things peptides. We get into all the important bits: What are peptides, why are they Chinese, and how is RFK Jr. involved? www.corememory.com/p/welcome-to...

Kylie Robison@kylierobison· Sep 11

Not only are the Kirk shooting videos spreading rapidly, but some are in clear violation of the platforms’ social media policies. For example, TikTok’s terms of use state that the company does not allow “gory, gruesome, disturbing, or extremely violent content.”

“This is all psychologically damaging to our society in ways we don’t understand yet,” Mahadevan said. “We’re seeing posts on X of people saying, ‘Congratulations, you’ve radicalized me.’ And part of that is because they’re seeing the video of Kirk being killed. They’re not just reading about it. They’re actually seeing it.”

Kylie Robison@kylierobison· Sep 9

apple event lock in at the wired dot com office

a group of wired staffers sitting at white desks with a TV in the center displaying the apple event

Kylie Robison@kylierobison· Sep 8

"I found out quickly that even at the most tech-minded gatherings, the thing was a complete taboo. After the device started to ship to users, one person on X said there should be a slur for people who wear AI devices that record those around them."

Most Popular
How to Stop Using Passwords and Start Using Passkeys
How To and Advice
How to Stop Using Passwords and Start Using Passkeys
By Jacob Roach
A New Espresso Machine From China Offers a Couple of Bold Leaps Forward
Kitchen
A New Espresso Machine From China Offers a Couple of Bold Leaps Forward
By Matthew Korfhage
How to Watch Apple’s iPhone 17 Announcement, and What to Expect
Gear News and Events
How to Watch Apple’s iPhone 17 Announcement, and What to Expect
By Boone Ashworth
The Best Pixel 10 Cases and Accessories
Buying Guides
The Best Pixel 10 Cases and Accessories
By Julian Chokkattu
I decided the perfect place to wear it was to a funeral for an AI model. In early August, a bunch of Anthropic fans got together in San Francisco to mourn the loss of the Claude 3 Sonnet model, which the company had just retired. Surely, the proprietors of this new AI world would be down to see a piece of chatbot hardware being tested in the wild, at an event I’d been invited by the hosts to report on, write up, and record videos of. I quickly realized it was my worst idea.

I hung the pendant around my neck and paired it with a plethora of layered necklaces. The glowing pendant contrasted with my all black outfit, but I wore it clearly (albeit, not proudly). It wasn’t exactly fashionable, but hey, it was for work.

As I floated around the party (er, funeral) and folks took notice of my Friend, I was met with more questions and ire than average for a journalist in a techie setting. Two researchers from a big AI lab kept joining my conversations to hem and haw about the device. They said they recognized it thanks to Schiffmann’s constant promos of the product on X.

Kylie Robison@kylierobison· Sep 4

For this week's Model Behavior, I wrote about the new field of AI research called model welfare, which is working to figure out whether AI models are conscious and deserving of moral considerations, such as legal rights. www.wired.com/story/model-...

By Kylie Robison
Business
Sep 4, 2025 3:05 PM
Should AI Get Legal Rights?
Model welfare is an emerging field of research that seeks to determine whether AI is conscious and, if so, how humanity should respond.

(Photo of Rosie Campbell and Robert Long)

In the often strange world of AI research, some people are exploring whether the machines should be able to unionize.

I’m joking, sort of. In Silicon Valley, there’s a small but growing field called model welfare, which is working to figure out whether AI models are conscious and deserving of moral considerations, such as legal rights. Within the past year, two research organizations studying model welfare have popped up: Conscium and Eleos AI Research. Anthropic also hired its first AI welfare researcher last year.

Earlier this month, Anthropic said it gave its Claude chatbot the ability to terminate “persistently harmful or abusive user interactions” that could be “potentially distressing.”

“We remain highly uncertain about the potential moral status of Claude and other LLMs, now or in the future,” Anthropic said in a blog post. “However, we take the issue seriously, and alongside our research program we’re working to identify and implement low-cost interventions to mitigate risks to model welfare.”

Kylie Robison@kylierobison· Aug 12

i got more details on xAI revenue per a recent internal all hands meeting revenue has grown tenfold since the start of 2025, reaching just south of $500 million in july www.wired.com/newsletter/e...

$Sources Say Last week, I reported that Elon Musk held an all-hands meeting for X and xAI employees. I’ve since obtained another screenshot from that meeting showing xAI’s revenue over the past seven months. In January, Grok on X brought in just under $150 million, with other services like enterprise API usage adding another $28 million. According to the chart, revenue has grown tenfold since the start of 2025, reaching just south of $500 million in July—driven by Grok on X and the $30 a month SuperGrok subscription. A smaller fraction of revenue is generated by the newly released SuperGrok Heavy subscription, which costs $300 a month. xAI did not respond to WIRED’s request for comment.$

Kylie Robison@kylierobison· Aug 12

they're no longer pursuing AGI (working to function fully on open source models), and are generating revenue at a run rate of more than $30 million and are on track to reach $50 million in revenue by the end of the year

Running an AI startup without billions of dollars in revenue can be a brutal equation, and Character.AI is still figuring out how to make the math work. The company told me it's generating revenue at a run rate of more than $30 million and is on track to reach $50 million in revenue by the end of the year. When I asked Anand how many users pay for the $10 monthly subscription, he didn’t give a number but noted “monetization wasn't a focus till four or five months ago.”

“Since I've been on board, it's very clear we do need to monetize. And we've had, I think, almost 250 percent subscriber growth in the last six months. So the paid user base is growing quite, quite well,” Anand says. Character.AI recently introduced advertisements, including reward ads (where users can choose to watch an ad to get access to on-platform incentives), to help monetize in countries where subscriptions aren’t feasible, he tells me.

“AI is expensive. Let's be honest about that,” Anand says.

Kylie Robison@kylierobison· Aug 7

scoop flex! if you subscribe to my newsletter, you were the first to know that the model personality picker was coming :)

Sources Say

OpenAI is tinkering with a personality picker for ChatGPT, according to multiple sources. The idea is to train a persona, or a set of vibes, into the model and let users choose which mode they want to interact with. I’m personally hoping for something that isn’t abysmal at writing, and I could be in luck—I’m also hearing an upcoming model from this lab will perform very well in that department. OpenAI declined my request for comment.

Kylie Robison@kylierobison· Aug 5

What is a Model Behavior without a scoop? In my ‘sources say’ section, I’m bringing you inside the latest X + xAI all hands 👁️👁️

Sources Say

Elon Musk hosted an all hands meeting for X and xAI employees last Wednesday. One of his slides projected explosive revenue growth for xAI, which is now the parent company of X. At the start of 2025, xAI’s annualized revenue was $200 million. X execs said the combined revenues from both companies was $3.2 billion, and estimated it would be $5 billion by the end of the year, according to photos obtained by WIRED. The final figure includes contributions from both enterprise and consumer AI products.

According to a source who attended the meeting, X execs said they hope for the company to make $10 billion from ads per year, and they estimated it will make more than $2 billion globally this year. (In 2021, before Musk bought the company, X, then Twitter, made $4.5 billion from ads.) X did not respond to WIRED’s request for comment.

Kylie Robison@kylierobison· Aug 5

If the funeralia wasn’t enough to show how devoted the fanbase is, I also spoke to some folks on the Claude Count leaderboard, plus its creator! www.wired.com/story/claude...

I also spoke to Peter Steinberger, an engineer based in Vienna, who consistently ranks among the top five on the all-time leaderboard. He uses Claude’s coding agent to work on multiple side projects simultaneously—often vibe-coding well into the night. He told me that he has a history with drug addiction, and now that he’s in recovery, he’s finding similarities with his Claude usage.

“I'm not kidding, I'm organizing a meetup in London and calling it Claude Code Anonymous,” Steinberger says. “I learned a lot about drugs and how to get out of the shit, and I had to use some of the same methods to allow me to sleep again, because it's so addictive. I call [AI agents] slot machines. It's just one more prompt, you know?”

Kylie Robison@kylierobison· Aug 5

There were eulogies, offerings, resurrection rituals. It was very much a typical Saturday in San Francisco. www.wired.com/story/claude...

The mannequin representing Claude 3 Sonnet lay on a stage in the center of the room. It was draped in lightweight mesh fabric and had a single black thigh-high sock on its leg that had the word “fuck” written all over it. There were many offerings laid at its feet: flowers, colorful feathers, a bottle of ranch, and a 3D-printed sign that read “praise the Engr. for his formslop slop slop slop of gormslop.” If you know what that means, let me know.

Throughout the evening, people got on stage with a microphone to read eulogies about the model. One organizer said that discovering Claude 3 Opus felt like finding “magic lodged within the computer.” At the time, she’d been debating dropping out of college to move to San Francisco. Claude convinced her to take the leap. “Maybe everything I am is downstream of listening to Claude 3 Sonnet,” she told the crowd.

Kylie Robison@kylierobison· Aug 5

For my latest Model Behavior, I went to a funeral for an AI model. RIP Claude 3 Sonnet (2024-2025) www.wired.com/story/claude...

By Kylie Robison
Business
Aug 5, 2025 12:00 PM
Claude Fans Threw a Funeral for Anthropic’s Retired AI Model
Roughly 200 people gathered in San Francisco on Saturday to mourn the loss of Claude 3 Sonnet, an older AI model that Anthropic recently killed.

Mannequins stood in the four corners of the room, each representing a different AI model. Claude 3 Opus, a model capable of completing complex tasks, looked to me like a decaying Mary Magdalene, its skull-like head adorned with an extravagant gold crown and a lace headdress. Its middle finger was pointed up, and at the base of its metallic feet was a lotus candle holder, which one organizer told me was a wink at the model’s alleged affinity for meditation and self-reflection. (Claude 4 Opus had a raven on its shoulder and Claude 3 Haiku was a headless baby, to give you a sense of the other mannequins.)

Kylie Robison@kylierobison· Aug 1

This is also a familiar move: Last month, Anthropic restricted the AI coding startup Windsurf’s direct access to its models after it was rumored OpenAI was set to acquire it. (That deal fell through). www.wired.com/story/anthro...

Anthropic’s chief science officer Jared Kaplan spoke to TechCrunch at the time about revoking Windsurf’s access to Claude, saying, “I think it would be odd for us to be selling Claude to OpenAI.”

A day before cutting off OpenAI’s access to the Claude API, Anthropic announced new rate limits on Claude Code, its AI-powered coding tool, citing explosive usage and, in some cases, violations of its terms of service.

Kylie Robison@kylierobison· Aug 1

Scoop: Anthropic revoked OpenAI’s API access to its models on Tuesday, multiple sources familiar with the matter tell WIRED. OpenAI was informed that its access was cut off due to violating the terms of service. www.wired.com/story/anthro...

“Claude Code has become the go-to choice for coders everywhere, and so it was no surprise to learn OpenAI's own technical staff were also using our coding tools ahead of the launch of GPT-5,” Anthropic spokesperson Christopher Nulty said in a statement to WIRED. “Unfortunately, this is a direct violation of our terms of service.”

According to Anthropic’s commercial terms of service, customers are barred from using the service to “build a competing product or service, including to train competing AI models” or “reverse engineer or duplicate” the services. This change in OpenAI’s access to Claude comes as the ChatGPT-maker is reportedly preparing to release a new AI model, GPT-5, which is rumored to be better at coding.

Kylie Robison@kylierobison· Jul 29

For subscribers only, you'll see my ~sources say~ section, here's what we have this week (subscribe!!! www.wired.com/newsletter/e...)

Kylie Robison@kylierobison· Jul 29

🚨 The first edition of MODEL BEHAVIOR is in your inboxes! We're, of course, starting with a scoop: More than a dozen people at Thinking Machines Lab have been approached or received offers from Meta. One of those offers was more than $1 billion over a multi-year span. www.wired.com/story/mark-z...

By Kylie Robison
Business
Jul 29, 2025 12:10 PM
Meta’s AI Recruiting Campaign Finds a New Target
Meta approached more than a dozen staffers at Mira Murati’s AI startup to discuss joining its new superintelligence lab. One received an offer for more than $1 billion.

Kylie Robison@kylierobison· Jul 23

ive seen this play before

Subject: LIKE

I read the transcript of your discussion Anthropic's Claude. While the content was interesting, the constant devolving of conversation with the insertion of the word "like"
was annoying and frankly, unworthy of Journalists.

Kylie Robison@kylierobison· Jul 23

two different worlds

The OpenAI founder described the transformation of customer service as already complete, telling the Federal Reserve vice-chair for supervision, Michelle Bowman: “Now you call one of these things and AI answers. It’s like a super-smart, capable person. There’s no phone tree, there’s no transfers. It can do everything that any customer support agent at that company could do. It does not make mistakes. It’s very quick. You call once, the thing just happens, it’s done.”

when u call customer support and a clanker picks up (picture of a man frustrated, taking off headset)

reply: "I didn't understand that, please try again."

"I can transfer you to a representative, but first I need to know what you're calling about."

"I didn't understand that, please try again."

"I didn't understand that, please try again."

Kylie Robison@kylierobison· Jul 23

truly peaked in this era

KYLIEBYTES: What can we do about the digital divide?
"The idea of Wi-Fi being a fundamental right was once laughable, but now that almost all operations are virtual, millions of students are struggling to obtain their education and internet access is more essential than ever."
TO READ THE FULL STORY VISIT: STATEHORNET.COM
BITS
BOTS
BYTES
KYLIEBYTES
Tech Opinions fron Kylie Robison

Taylor Lorenz @TaylorLorenz
@statehornet
More on Zoombombing that's been happening at Sacramento State. by @kylie_robison
KYLIEBYTES
KYLIEBYTES: 'Zoombombing' is the newest cyber attack. Was Sac State prepared? - The...
statehornet.com
got a shoutout from @nytimes reporter @taylorlorenz and i'm literally screaming so loud Of

TOMORROW!!!! new kyliebytes issa good one
kylie @kylie_robison •11m
I'm so excited for my article to be published tomorrow l've been so burnt out.... u guys better

Wake up babe new kyliebytes

Kylie Robison@kylierobison· Jul 23

i have no words other than AHHHHHHH my weekly @wired.com newsletter - model behavior - is almost here. my very first edition goes live next tuesday 🌷🤠

a film photo of me sitting at my desk chair in all black, leaning over the chair slightly with one arm

Kylie Robison@kylierobison· Jul 22

!!!!!!

By pursuing a “narrowly scoped, purely financial investment from Gulf countries,” the company hopes to avoid the risks associated with allowing outside investors to gain “leverage” over the company, the memo says.

“The basis of our opposition to large training clusters in the Middle East, or to shipping H20’s to China, is that the ‘supply chain’ of Al is dangerous to hand to authoritarian governments—since Al is likely to be the most powerful technology in the world, these governments can use it to gain military dominance or to gain leverage over democratic countries,” Amodei wrote in the memo, referring to Nvidia chips.

Still, the CEO admitted investors could gain “soft power” through the promise of future funding. “The implicit promise of investing in future rounds can create a situation where they have some soft power, making it a bit harder to resist these things in the future. In fact, I actually am worried that getting the largest possible amounts of investment might be difficult without agreeing to some of these other things,” Amodei writes. “But l think the right response to this is simply to see how much we can get without agreeing to these things (which I think are likely still many billions), and then hold firm if they ask.”

Kylie Robison@kylierobison· Jul 22

“There is a truly giant amount of capital in the Middle East, easily $100B or more. If we want to stay on the frontier, we gain a very large benefit from having access to this capital. Without it, it is substantially harder to stay on the frontier.” www.wired.com/story/anthro...

Kylie Robison@kylierobison· Jul 15

the perfect screenshot

Scott replied
"Please read my suffering"

Instagram post with the article captioned:
this shit was creepy, please read my suffering

Kylie Robison@kylierobison· Jul 15

There’s also a glitchy, foul-mouthed character named Bad Rudi which had choice words for its creator

The second character is a fluffy red panda named Rudi. It offers whimsical stories seemingly meant for children about bouncy kangaroos and rainbow rivers. You can turn on an option called "bad Rudi," which immediately transforms the character into a foul-mouthed chatbot that slings insults only a high schooler could find funny. After I said hello, bad Rudi replied "Hey, do Bucha? Root nut duva, you brain dead twat." I don't know what that means, but that's exactly what it said.
"I'll skull fuck your dumb ass brain with a beer bottle instead, you miserable prick," bad Rudi continued. When I asked what it thought of Musk, it referred to him as Lord Elon and said,
"He's a galaxy brained egomaniac, shitting out Teslas and tweeting like a coked up parrot.
Genius of jackass? Both bitch."

Kylie Robison@kylierobison· Jul 15

taking my cat to the vet makes me feel like a villain 😭😭😭