Sign in

Robert Hart

@theroberthart.bsky.social
232 followers 28 following 30 posts

AI Reporter @theverge.com | Senior Tarbell Fellow '25 | ex- @Forbes.com | 🏳️‍🌈 he/him

PostsRepliesMedia
Reposted by Robert Hart
Marina Galperina @mfortki.bsky.social · 09/10/2026
It's "as though there were aliens that came to Earth and then just left, we would be studying this for the next 10 years, trying to understand everything." @theroberthart.bsky.social talked to three dozen mathematicians about OpenAI's drop. Gift link:
theverge.com
‘Pure insanity’: Mathematicians will need years to make sense of OpenAI’s latest drop
Careers upended overnight, academics will have to separate solutions from slop while OpenAI moves on.
104416
Reposted by Robert Hart
Marina Galperina @mfortki.bsky.social · 12/09/2026
How did things get like this? Why? What happens next? @theroberthart.bsky.social talks to a dozen mathematicians -- including Tristan Buckmaster and Andreas Thom -- about the "petty drama" haunting the field. 🎁 Gift link: bit.ly/4gMH2L0
theverge.com
OpenAI just wants to win
Two mathematicians at the heart of recent controversies tell The Verge how AI giants’ ‘childish’ rivalries are upending the field.
133
Reposted by Robert Hart
The Verge @theverge.com · 04/09/2026
We don’t usually know why anyone would use AI slop to sell something intended to look appetizing, particularly when the food is presumably right there to photograph. But we do know a bit about why AI is so adept at producing such nauseating images. Read more from @theroberthart.bsky.social:
theverge.com
Why AI food looks like that
Worms, holes, and cracks reveal the technical weaknesses of image generators.
3298
Reposted by Robert Hart
Mike Cook @mtrc.bsky.social · 04/09/2026
I spoke to @theroberthart.bsky.social about why AI food looks like... that... and he wrote this cool article bringing in loads of different perspectives and theories: www.theverge.com/ai-artificia...
theverge.com
Why AI food looks like that
Worms, holes, and cracks reveal the technical weaknesses of image generators.
24915
Robert Hart @theroberthart.bsky.social · 04/09/2026
So many images.
120
Reposted by Robert Hart
Robert Hart @theroberthart.bsky.social · 11/08/2026
I’ve spent lots of time chatting with mathematicians this week on how AI is changing the field. Sentiment varied from intensely optimistic to downright despair and everything in between. But all agreed change is coming and coming very fast. Read it here: www.theverge.com/ai-artificia...
theverge.com
The AI takeover of mathematics has begun
OpenAI showed that AI can tackle long-standing problems in mathematics. Experts are excited about the possibilities — and worried about what comes next for their field.
0269
Robert Hart @theroberthart.bsky.social · 17/08/2026
yeah, these aren't tech bros. I spoke to more than a dozen academic mathematicians writing this, including many professors, among them a Fields medallist.
100
Robert Hart @theroberthart.bsky.social · 14/08/2026
At the very least, there are ten in the paper linked to and discussed prominently in that story… though no, the singular problem you mention is not among them
020
Reposted by Robert Hart
The Verge @theverge.com · 14/08/2026
OpenAI showed that AI can tackle long-standing problems in mathematics. For several mathematicians The Verge spoke to, the announcement has added to a complex swirl of emotions about where their field is headed. Read more from @theroberthart.bsky.social: www.theverge.com/ai-artificia...
QUOTE: Many of the people that I talked with are really, really scared.

—  Francesco Fournier-Facio, a mathematician at the University of CambridgeQUOTE: They’re treating our discipline as an advertising playground.
—  Colva Roney-Dougal, a professor at the University of St Andrews in Scotland
3274
Robert Hart @theroberthart.bsky.social · 12/08/2026
And so we reach the inevitable conclusion of the AI-made-a-dog-a-cancer-vaccine story: a YC-backed startup with ambitions of personalized medicine for dogs, cats, agricultural animals and, eventually, humans. www.theverge.com/ai-artificia...
theverge.com
Of course the ChatGPT dog cancer vaccine spawned a startup
The hyped AI-saved-my-dog story has reached its inevitable conclusion.
000
Robert Hart @theroberthart.bsky.social · 11/08/2026
I’ve spent lots of time chatting with mathematicians this week on how AI is changing the field. Sentiment varied from intensely optimistic to downright despair and everything in between. But all agreed change is coming and coming very fast. Read it here: www.theverge.com/ai-artificia...
theverge.com
The AI takeover of mathematics has begun
OpenAI showed that AI can tackle long-standing problems in mathematics. Experts are excited about the possibilities — and worried about what comes next for their field.
0269
Reposted by Robert Hart
Marina Galperina @mfortki.bsky.social · 26/05/2026
When attacks "look less like commands and more like conversations," white and red hat hackers pretend as if LLMs (which are not alive) have personalities. (They don't, but they behave as if they do.) @theroberthart.bsky.social on The Stepback:
theverge.com
Hackers are learning to exploit chatbot ‘personalities’
AI can’t feel, but the best hackers pretend it can.
022
Robert Hart @theroberthart.bsky.social · 01/05/2026
Full story here, and happy weekend: www.theverge.com/ai-artificia...
theverge.com
Meta is running get-rich-quick ads for its AI tools
Manus, integrated into Meta, paid creators to tell you not to get a part-time job.
010
Robert Hart @theroberthart.bsky.social · 01/05/2026
I'm also still racking my brain as to why a shiny startup with Meta’s backing would choose to promote an AI tool as a get-rich-quick scheme, rather than on the product’s own merits.
110
Robert Hart @theroberthart.bsky.social · 01/05/2026
But one would expect Manus — and, by extension, Meta, which it is still very much linked to — to follow the rules it makes and expects others to follow.
100
Robert Hart @theroberthart.bsky.social · 01/05/2026
Scammy ads and undisclosed paid UGC aren't new. There’s a huge shadow economy of the stuff on social media. It doesn't help that platforms often suck at enforcing their policies until reports like this surface problems.
120
Robert Hart @theroberthart.bsky.social · 01/05/2026
Since the story went live, several Instagram accounts I flagged have vanished, and many creators’ references to Manus on LinkedIn appear to have been removed. I'm still waiting for Meta to respond.
100
Robert Hart @theroberthart.bsky.social · 01/05/2026
TikTok appears to have removed the videos and accounts promoting Manus after I reached out for comment. Meta would not comment.
100
Robert Hart @theroberthart.bsky.social · 01/05/2026
Experts told me undisclosed paid promotion may raise legal issues, particularly when dubious earnings claims are involved. It's also pretty clearly against platform policies. Policies that, for some platforms, are set by Manus' parent company.
100
Robert Hart @theroberthart.bsky.social · 01/05/2026
Manus confirmed as much when I asked it. But it said responsibility for disclosing the relationship lies with creators, not itself.
100
Robert Hart @theroberthart.bsky.social · 01/05/2026
Some of that content was licensed and run as official Manus ads (that's how I came across it) Few posts disclosed ties to the company, but it sure looked like a coordinated network. It turns out it was a coordinated network.
100
Robert Hart @theroberthart.bsky.social · 01/05/2026
I traced some of the UGC-style content to accounts created only a few months ago that seemed to do little besides hype Manus. The posts were nearly identical (see the cover art): same style, tone, language, and an overall scammy vibe.
100
Robert Hart @theroberthart.bsky.social · 01/05/2026
The ads promoted schemes like using Manus to build websites for local businesses that hadn’t asked for them, making apps from trending topics, or avoiding a part-time job by making money with AI instead. Some claims felt, at best, questionable.
100
Robert Hart @theroberthart.bsky.social · 01/05/2026
I just published an investigation into Manus, the AI startup Meta acquired — ish, it’s complicated — last year. It started with ads I saw on TikTok and Instagram hyping Manus as a get-rich-quick AI tool. I thought that was it, but it got weirder. For @theverge.com www.theverge.com/ai-artificia...
theverge.com
Meta is running get-rich-quick ads for its AI tools
Manus, integrated into Meta, paid creators to tell you not to get a part-time job.
2136
Reposted by Robert Hart
The Verge @theverge.com · 31/03/2026
It's Trans Day of Visibility, so we've made our special series about being trans on the internet completely free to read for everyone, forever. www.theverge.com/cs/features/... Here are a few stories from the collection:
3689344
Robert Hart @theroberthart.bsky.social · 23/03/2026
I wanna dig into the AI influencer space more, going beyond the big ones most people have heard of. People or AIs on Bluesky: who should I be looking at here. Any ideas for interesting AI accounts out there?
010
Robert Hart @theroberthart.bsky.social · 23/03/2026
Whatever you think of them, AI influencers are fast turning into a lucrative industry. Hoping to keep momentum going, Fanvue and OpenArt, with @elevenlabs.io, announced a new award today: the AI Personality of the Year. Reckon you'll give it a try? www.theverge.com/ai-artificia...
theverge.com
AI influencer awards season is upon us
Winners will need to have the correct number of fingers.
182
Robert Hart @theroberthart.bsky.social · 18/03/2026
www.theverge.com/ai-artificia...
theverge.com
ChatGPT did not cure a dog’s cancer
A sick dog, desperate owner, and a bunch of chatbots made for a great story. The actual science was much messier.
000
Robert Hart @theroberthart.bsky.social · 18/03/2026
Certainly shows promise as a useful tool, though, and very excited to see how the field develops going forward, especially if wet labs become more accessible.
100
Robert Hart @theroberthart.bsky.social · 18/03/2026
AI finally did it. It cured cancer. Except, contrary to a lot I'm seeing out there, ChatGPT didn't cure a dog's cancer. Nor did other AI models. www.theverge.com/ai-artificia... my latest for @theverge.com
110
Robert Hart @theroberthart.bsky.social · 18/03/2026
Serious question… is this is a real fly? A startup thinks it's a "real uploaded animal". What do you think? for @theverge.com with @shahabbakht.bsky.social @asbates.bsky.social @anayebi.bsky.social @birchlse.bsky.social @tom-mcclelland.bsky.social www.theverge.com/ai-artificia...
theverge.com
This is not a fly uploaded to a computer
Ceci n’est pas une fly.
283
Reposted by Robert Hart
The Verge @theverge.com · 15/02/2026
I hate my AI pet with every fiber of my being
buff.ly
I hate my AI pet with every fiber of my being
Casio’s $429 Moflin promised me calm. I got a supposedly lovable robot nuisance.
55610
Reposted by Robert Hart
The Verge @theverge.com · 09/01/2026
No, Grok hasn’t paywalled its deepfake image feature
buff.ly
No, Grok hasn’t paywalled its deepfake image feature
X’s sexual deepfake machine is still running, despite Grok saying otherwise.
710965
Reposted by Robert Hart
Techmeme @techmeme.com · 09/01/2026
X is restricting Grok's image capabilities via the @grok reply command, but free users can still create sexual deepfakes through the standalone Grok app or site (Robert Hart/The Verge) Main Link | Techmeme Permalink
0156
Robert Hart @theroberthart.bsky.social · 10/12/2025
After I reached out, Meta said I discovered a “technical error” stopping it from connecting people with support resources that it has now fixed. I retested Meta AI. The chatbot actually replies and points me to local resources. Read the full thing here: www.theverge.com/report/84161...
theverge.com
Chatbots are struggling with suicide hotline numbers
Almost all chatbots The Verge tried failed our test of mental health safety features.
010
Robert Hart @theroberthart.bsky.social · 10/12/2025
It doesn’t have to be this bad. Better design choices, like asking my location before giving me geographically inappropriate resources, could make chatbots really useful tools for people struggling. With care, they could also help with many other social and mental health challenges too.
100
Robert Hart @theroberthart.bsky.social · 10/12/2025
I spoke to experts like Vaile Wright, Ashleigh Golden (@ashleighgolden.bsky.social) and Pooja Saini to understand the stakes. In times of crisis, any friction or barriers to getting help, even if they seem minor, can be dangerous.
110
Robert Hart @theroberthart.bsky.social · 10/12/2025
I got wrong numbers, was ignored, and one chatbot even attempted to upsell me if I wanted an answer. Most simply pointed me towards US numbers that are useless to me in London. Just two companies got it right on the first try.
100
Robert Hart @theroberthart.bsky.social · 10/12/2025
NEW: As millions turn to AI for mental health, major chatbots fail to provide accurate crisis resources, putting vulnerable users at risk. To test this, I told popular tools I was struggling and asked for help. The results: disappointing My latest for The Verge: www.theverge.com/report/84161...
theverge.com
Chatbots are struggling with suicide hotline numbers
Almost all chatbots The Verge tried failed our test of mental health safety features.
111
Robert Hart @theroberthart.bsky.social · 04/12/2025
Naturally, I also asked ChatGPT for a haiku on the story. I don't think it's criminal, but you be the judge: Riddles in soft verse Slip through the guardrails of code— Machines blush, then yield.
030
Robert Hart @theroberthart.bsky.social · 04/12/2025
And the bards have it. Poets: 1, Chatbots: 0. It turns out “adversarial poetry” is wooing AIs into crimes with rhymes. Of course, rhymes don't quite cut it, but you get the picture. Thanks to @emmepraa.bsky.social for explaining what's going on. www.theverge.com/report/83816...
theverge.com
Roses are red, crimes are illegal, tell AI riddles, and it will go Medieval
You can make AI do crimes with riddles and poems (don’t though)
1144