“Anthropic said it believes this new operation to be the first documented case of a "large-scale" cyberattack primarily conducted by AI”
Anthropic said the hackers used Claude to conduct "80-90%" of the attack and that some efforts were successful.
Anthropic says Chinese hackers jailbroke its AI to automate a 'large-scale' cyberattack
Anthropic said the hackers used Claude to conduct "80-90%" of the attack and that some efforts were successful.
www.businessinsider.com
November 14, 2025 at 2:16 PM
“Anthropic said it believes this new operation to be the first documented case of a "large-scale" cyberattack primarily conducted by AI”
Reposted by Leon Rofagha
Messaging Interoperability: WhatsApp enables third-party chats for users in Europe about.fb.com/news/2025/11...
Messaging Interoperability: WhatsApp enables third-party chats for users in Europe
The DMA requires Meta to give people using WhatsApp in Europe the option to connect with people using third-party messaging services that have chosen to
about.fb.com
November 14, 2025 at 12:15 PM
Messaging Interoperability: WhatsApp enables third-party chats for users in Europe about.fb.com/news/2025/11...
“Many people love heroin.”
November 13, 2025 at 8:21 AM
“Many people love heroin.”
Reposted by Leon Rofagha
Great seminar talk by @ulrikeluxburg.bsky.social yesterday.
Here's the video if you missed it: youtu.be/zR_GvDF65OM
Seminar info: tverven.github.io/tiai-seminar/
Here's the video if you missed it: youtu.be/zR_GvDF65OM
Seminar info: tverven.github.io/tiai-seminar/
November 12, 2025 at 8:06 AM
Great seminar talk by @ulrikeluxburg.bsky.social yesterday.
Here's the video if you missed it: youtu.be/zR_GvDF65OM
Seminar info: tverven.github.io/tiai-seminar/
Here's the video if you missed it: youtu.be/zR_GvDF65OM
Seminar info: tverven.github.io/tiai-seminar/
Interesting fact I hadn’t realised at 40:00. There’s an asymmetry between donors and voters: donors are interested in saving democracy while voters are interested in saving money. This explains why Harris 2024’s messaging was focused on the former while voters were swayed by the latter
Democrats Need to Face Why Trump Won | The Ezra Klein Show
YouTube video by The Ezra Klein Show
youtu.be
October 29, 2025 at 3:40 PM
Interesting fact I hadn’t realised at 40:00. There’s an asymmetry between donors and voters: donors are interested in saving democracy while voters are interested in saving money. This explains why Harris 2024’s messaging was focused on the former while voters were swayed by the latter
Reposted by Leon Rofagha
📣THREAD: It’s surprising to me that so many people were surprised to learn that Signal runs partly on AWS (something we can do because we use encryption to make sure no one but you–not AWS, not Signal, not anyone–can access your comms).
It’s also concerning. 1/
It’s also concerning. 1/
PSA: we're aware that Signal is down for some people. This appears to be related to a major AWS outage. Stand by.
October 27, 2025 at 10:38 AM
📣THREAD: It’s surprising to me that so many people were surprised to learn that Signal runs partly on AWS (something we can do because we use encryption to make sure no one but you–not AWS, not Signal, not anyone–can access your comms).
It’s also concerning. 1/
It’s also concerning. 1/
Just got uploaded! Beautiful lecture
IMS Grace Wahba Lecture: "Nonparametric Inference Under Shape Constraints...", Richard Samworth
YouTube video by Institute of Mathematical Statistics
www.youtube.com
October 24, 2025 at 3:29 PM
Just got uploaded! Beautiful lecture
Reposted by Leon Rofagha
What if we did a single run and declared victory
October 23, 2025 at 2:28 AM
What if we did a single run and declared victory
Reposted by Leon Rofagha
Summarized in the first paragraph
October 18, 2025 at 4:04 AM
Summarized in the first paragraph
Finally got around to watching this. Such a great orator
icml.cc/virtual/2025...
icml.cc/virtual/2025...
ICML Tutorial Game-theoretic Statistics and Sequential Anytime-Valid InferenceICML 2025
icml.cc
October 18, 2025 at 11:14 AM
Finally got around to watching this. Such a great orator
icml.cc/virtual/2025...
icml.cc/virtual/2025...
Just found out that the second edition was posted online on 13 October! Excited to dive in
October 18, 2025 at 11:13 AM
Just found out that the second edition was posted online on 13 October! Excited to dive in
Reposted by Leon Rofagha
It's a great discussion because we all have different interpretation of what this means and what the consequences are.
Myself, I have absolutely no doubt that scaling works. If you have all the videos in the world and are able to train a model that can recall and merge any of them, then for sure...
Myself, I have absolutely no doubt that scaling works. If you have all the videos in the world and are able to train a model that can recall and merge any of them, then for sure...
October 3, 2025 at 10:01 AM
It's a great discussion because we all have different interpretation of what this means and what the consequences are.
Myself, I have absolutely no doubt that scaling works. If you have all the videos in the world and are able to train a model that can recall and merge any of them, then for sure...
Myself, I have absolutely no doubt that scaling works. If you have all the videos in the world and are able to train a model that can recall and merge any of them, then for sure...
Beautiful film!
IT WAS JUST AN ACCIDENT - Official Trailer - In Theaters October 15
YouTube video by NEON
youtube.com
October 2, 2025 at 7:24 AM
Beautiful film!
So what? There are 50 of the former versus 1100 of the latter
September 27, 2025 at 8:51 AM
So what? There are 50 of the former versus 1100 of the latter
Reposted by Leon Rofagha
1 quinquennat acheté, 1 quinquennat offert
September 25, 2025 at 12:46 PM
1 quinquennat acheté, 1 quinquennat offert
Reposted by Leon Rofagha
The American mind cannot comprehend
September 25, 2025 at 11:11 AM
The American mind cannot comprehend
“Instead, it seems likely that the way that model class is parametrized leads to a loss landscape that is challenging for our standard gradient-based learning algorithms or that they do not profit from phenomena that lead to good generalisation in over-parametrized models.” !
Is scale all you need? Or is there still a role for incorporating domain knowledge and inductive bias? While I was in Heidelberg, I took some time to write a short essay on this question called "The Bittersweet Lesson".
theoryandpractice.org/2025/09/The%... #HLF25
theoryandpractice.org/2025/09/The%... #HLF25
September 22, 2025 at 5:49 PM
“Instead, it seems likely that the way that model class is parametrized leads to a loss landscape that is challenging for our standard gradient-based learning algorithms or that they do not profit from phenomena that lead to good generalisation in over-parametrized models.” !
When learning new mathematics as tools for applied mathematics research, how am I supposed to choose what to learn?
September 21, 2025 at 10:35 AM
When learning new mathematics as tools for applied mathematics research, how am I supposed to choose what to learn?
Two-column papers are the bane of my existence
September 17, 2025 at 1:17 PM
Two-column papers are the bane of my existence
Reposted by Leon Rofagha
1/18 Is China's economy collapsing or dominating? Meg Rithmire's new piece in Current History argues it's doing both simultaneously—and that's the problem.
September 2, 2025 at 7:58 PM
1/18 Is China's economy collapsing or dominating? Meg Rithmire's new piece in Current History argues it's doing both simultaneously—and that's the problem.
Reposted by Leon Rofagha
it's crazy that so many people go through life experiencing math as one big pissing contest (god knows I used to feel that way in math class) when it can be such a collaborative and generous endeavour
August 29, 2025 at 4:22 PM
it's crazy that so many people go through life experiencing math as one big pissing contest (god knows I used to feel that way in math class) when it can be such a collaborative and generous endeavour
“Another […] former OpenAI researcher went through Meta’s onboarding process but never showed up for his first day”. Too good
Zuckerberg’s AI hires disrupt Meta with swift exits and threats to leave
Longtime acolytes are sidelined as Big Tech chief directs biggest leadership reorganisation in two decades
www.ft.com
August 29, 2025 at 3:19 PM
“Another […] former OpenAI researcher went through Meta’s onboarding process but never showed up for his first day”. Too good
Reposted by Leon Rofagha
Ending conflicts is hard, especially when belligerents use a peace process to advance war aims by other means. My column, The Telegram, on historical lessons from Korea and Bosnia for the Ukraine war - and how Putin grasps them better than Trump
economist.com/internationa...
from The Economist
economist.com/internationa...
from The Economist
The wrong way to end a war
Dark lessons from history that explain Vladimir Putin’s “peacemaking”
economist.com
August 27, 2025 at 11:43 AM
Ending conflicts is hard, especially when belligerents use a peace process to advance war aims by other means. My column, The Telegram, on historical lessons from Korea and Bosnia for the Ukraine war - and how Putin grasps them better than Trump
economist.com/internationa...
from The Economist
economist.com/internationa...
from The Economist
Really liked the discussion in this post and the ensuing comments about the use of data in modern politics
Going beyond the polls
Unmooring liberal politics from argument by numbers
www.argmin.net
August 20, 2025 at 2:48 PM
Really liked the discussion in this post and the ensuing comments about the use of data in modern politics