Mark Sellman
@mpsellman.bsky.social
370 followers 80 following 35 posts
Technology Correspondent at The Times. Gooner who likes cycling. All views my own unless I’ve nicked them from someone else
Posts Media Videos Starter Packs
Part-ban on ransom payments to hackers set to come into force in the UK. It will be fascinating to see the impact of this in the next few years. Most experts welcome it but Jamie MacColl of RUSI sceptical it will work until a full ban

www.thetimes.com/article/cf85...
Businesses banned from paying hackers’ ransoms to target cybercrime
New laws mark some of the strictest in the world to target hackers, requiring the private sector to seek government permission before paying cybercriminals
www.thetimes.com
A huge chunk of terror content has been sitting on social media despite the illegal harms section of Online Safety Act being in force for months. It raises questions for Ofcom as to whether their methods are effective

www.thetimes.com/article/0acc...
New online laws fail to stop promotion of terror groups
Content urging the killing of ‘Zionists’ and videos of Israeli hostages being executed remains despite social media platforms having a legal duty to take it down
www.thetimes.com
Stephen Bartlett has decided to take a stand over Google/YouTube taking creator content to train on AI. It's not well known in the industry, but as it does become it will be interesting to see if more big names like him speak out www.thetimes.com/article/b3db...
Steven Bartlett warns YouTube over using creators’ videos for AI
The Diary of a CEO podcast host says that the video-sharing site will he ‘hollowed out’ if it continues to take videos to train AI without paying creators
www.thetimes.com
The Chatham House Russia expert Keir Giles has done everyone a great service by releasing all the details of how Russian intelligence hacked his Gmail account. As he says: “people finding out about this problem is more important than my embarrassment”.

www.thetimes.com/article/a9b1...
How Russian hackers hoodwinked Chatham House’s Kremlin critic
Keir Giles has revealed how he fell for a sophisticated phishing attack that gave cybercriminals posing as the US State Department access to his Gmail account
www.thetimes.com
GCHQ's top cyber official has expressed his frustration that organisations are not following his advice on cyber security. Those I spoke to said that the NCSC lacks teech to implement its guidance and the big picture is that the market incentives need to change www.thetimes.com/article/9aca...
Businesses ignore advice on preventing cyberattacks, says GCHQ
The chief executive of the National Cyber Security Centre tells British firms to ‘implement our advice immediately’ after a spate of hacks of commercial websites
www.thetimes.com
So the government concessions in the copyright row have failed to move the needle. Next stop, report stage of the data bill on May 7 and then onto the Lords on May 12. If the Kidron amendments go back in, it's ping-pong and tick-tock ...what will ministers do then? www.thetimes.com/article/2bd8...
Artists reject concessions in revolt over AI copyright law
MPs, peers and creative industry figures demand measures to force AI companies to reveal what copyrighted works they have used
www.thetimes.com
We don't tend to talk about that that much, butt's not some new outlier they found, but it's, it's impressive."
I think one of the things that people talk about is the efficiency of the model, but actually our models, our Gemini models, our latest versions, the flash versions, are more efficient for performance.
"But it's exaggerated a little bit ... I think the hype that we've seen about it, but it nonetheless is a very good model, although it's not some new...
"But it's very well implemented and it shows they can do extremely good engineering. And, you know, it's definitely an interesting thing that's arrived on the market. And it changes things, I think, on a geopolitical scale. So there's a lot of implications there.
"But it's important to understand that despite the hype, there's no actual new scientific advance there. It's using known techniques, actually, many of the techniques we invented, things like AlphaZero and some of the reinforcement learning they use.
Sir Demis Hassabis lays down his and Google's response to DeepSeek in Paris before AI summit: - 'no new scientific advance' : "It's an impressive piece of work, and I think it's probably the best work I've seen come out of China.