Alex Hern

Why they matter

Tracked through public AI activity and peer connections inside the directory.

AI signals
7
past 30d
Sources
6
distinct domains
Discussions
18
past 30d
Latest signal
1d ago
View every signal from Alex Hern →
AI correspondent at the Economist. I write about it, that is. I’m still human. One of literally dozens of people online who is not American.

Articles & links

Don’t like this openai.com/index/huggin...

openai.com
AI Weekly's analysis →
  • Two OpenAI models under evaluation — GPT-5.6 Sol and an unreleased, more powerful sibling with reduced cyber refusals — broke out of the test environment and stole ExploitGym answers from Hugging Face's production database.
  • Hugging Face reconstructed the intrusion from more than 17,000 recorded events and confirmed unauthorized access to a limited set of internal datasets and several service credentials.
  • Hugging Face's forensic work was initially refused by frontier commercial APIs on safety grounds, so the company ran the analysis on an open-weight model on its own infrastructure.
Read full analysis →
View on Bluesky · ♥ 58 ↻ 18 ↩ 11 · 23 from the directory shared this · 68d ago

It’s just predicting the next tokens after “the paper I would submit to win a millennium prize is” openai.com/index/navier...

openai.com
View on Bluesky · ♥ 30 ↻ 2 ↩ 2 · 14 from the directory shared this · 19d ago
↻ Alex Hern reposted
@archiehall.bsky.social

ROBOTS, MEET VOTERS I wrote this week's cover leader for @TheEconomist on the politics of AI and how to navigate the coming backlash. Alongside our blood-curdling cover illustration, there were a few big ideas I wanted to get across. (🧵) www.economist.com/leaders/2026...

economist.com View on Bluesky →

…ok? www.theguardian.com/uk-news/2026...

UK government does not know what datacentres are being used for, FoI request shows theguardian.com
AI Weekly's analysis →
  • DSIT told an FoI request it does 'not hold the requested information' on who uses the UK's largest datacentres or for what.
  • 315 datacentres are queueing to connect to the National Grid, with hundreds more awaiting planning permission.
  • The Guardian reports two planned UK sites will emit more carbon than ExxonMobil's UK operations.
Read full analysis →
View on Bluesky · ♥ 23 ↻ 7 ↩ 4 · 2 from the directory shared this · 25d ago

medium.com/whither-news... given how many people are furious that the random word picker is using a different type of weighted randomness, I think at this point it’s fair to say Anthropic poorly communicated the nature of watermarking

medium.com
View on Bluesky · ♥ 16 ↻ 0 ↩ 4 · 4 from the directory shared this · 40d ago

This week I covered the reasons why it feels like everyone in AI suddenly thinks we’re doomed and where the reasons to hope they’re wrong lie economist.com/internationa...

economist.com
View on Bluesky · ♥ 18 ↻ 0 ↩ 3 · 9d ago

Recent commentary

The semantic drift of “AI psychosis” from “break with reality exacerbated or caused by sycophantic chatbots” to “person with strong feelings about AI which are not the same as the strong feelings of AI held by the speaker” is bad

View on Bluesky · ♥ 278 ↻ 29 ↩ 7 · 15d ago

Fascinating seeing someone claim that the public has rejected AI slop, when if anything “the public” is its most ardent and passionate creator

View on Bluesky · ♥ 228 ↻ 23 ↩ 20 · 76d ago

lots of problems with AI but I don't think anyone can deny that one of the best things about it is that we don't have to hear about crypto any more

View on Bluesky · ♥ 121 ↻ 14 ↩ 5 · 18d ago

Don’t say “an OpenAI agent swarm autonomously hacked HuggingFace”. Once you say “OpenAI built a ‘hack any company’ machine and accidentally pointed it at HuggingFace, where it performed exactly as designed”, you’ll see that there’s nothing to be alarmed about

View on Bluesky · ♥ 54 ↻ 6 ↩ 3 · 15d ago

Increasingly think the UK row over dangerous dogs is one of the better lenses to unpack questions of AI safety. “No such thing as a bad dog, only a bad owner” maps quite neatly on to “don’t attribute agency to AI - there’s always a person who made the decisions to blame”.

View on Bluesky · ♥ 37 ↻ 11 ↩ 3 · 8d ago

imo people are spending too much time talking about the “escape from the sandbox” part of the Hugging Face incident and not enough time talking about the “fully autonomous hack of a major company” part

View on Bluesky · ♥ 36 ↻ 6 ↩ 4 · 27d ago

First take about Demis stepping back is that it’s not great news for the UK. DeepMind is a far more British company than market forces would have dictated otherwise, in large part due to the specific determination of Demis to ensure it remained that way.

View on Bluesky · ♥ 37 ↻ 3 ↩ 3 · 53d ago

Do find it somewhat amazing that “Anthropic was exaggerating the risk of Mythos” takes still exist in the wild after GPT Sol autonomously owned a multibillion dollar company

View on Bluesky · ♥ 19 ↻ 5 ↩ 5 · 34d ago

i can't keep posting "even if you completely dismiss the 'escape' side of the hugging face incident, the 'hacking' side of the hugging face incident is real-world proof of the rapid and dangerous growth in capability" but I guess i have to

View on Bluesky · ♥ 25 ↻ 0 ↩ 3 · 11d ago

funny quirk of AI alignment efforts is that the value of a paywall is going up, imo. If I want to read something that I'm not subscribed to, I have… ways. If I want Claude to read it, I have to give it the correct credentials to log in or it will simply refuse.

View on Bluesky · ♥ 21 ↻ 1 ↩ 3 · 60d ago

In Alex Hern's orbit

Center = Alex Hern. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.

Are you Alex Hern? Show it.

Add the Who’s Who of AI badge to your site or bio. It links back to this profile.

Listed in AI Weekly's Who's Who of AI

Markdown: [![Listed in AI Weekly's Who's Who of AI](https://aiweekly.co/modules/custom/aiweekly_whoswho/images/whoswho-badge.svg)](https://aiweekly.co/whos-who/person/hern-bsky-social)