Editor, @ReadTransformer. Prev: AI safety and EA comms, journalist @TheEconomist, @Protocol, @finimize

Joined October 2009
1,795 Photos and videos
This reads like a @helendewitt book (complimentary)
It's very clear that Fable-class LLMs are starting to feel constrained by "normal" speech and vernacular "Don't trust smells" "Calorimetry of learning" "Nobody's ticket drawer produced it" "Half receipt pending literature" "Tying the blind to the tolerance graph" "Park it fed"
1
813
This was, unfortunately, a hilarious way to celebrate America’s 250th
Navy Pier fireworks. Can’t see anything
9
1,590
This is going to sound cult-y, but it’s because he doesn’t believe. You can’t build AGI if you don’t really, truly believe in it.
Zuckerberg's ability to fuck up is mind boggling
5
3
83
9,920
I mean x.com/tomhfh/status/20730886…
Burnham says "nothing else" will fix the housing shortage other than council housing.
1
17
1,686
Look, I don't want to write him off before he's even started ... but this is going to be an absolute disaster
Burnham says the problem with HS2 was "a remote bureaucracy which always wanted to cut costs", and not "listening to the people whose areas it will affect".
9
29
681
25,848
Shakeel reposted
I'm sorry it's fucking crazy that that billionaire lady is disrupting the busiest rail station in the Western Hemisphere the day before a holiday on a whim. We live in the tackiest second gilded age imaginable
Penn Station will be open with regular LIRR service on Thursday, July 2 and Friday, July 3, but there will be access restrictions at the request of the NYPD. Customers traveling through Penn Station should use the 34th Street and 7th Avenue entrance. There will be no entry from 33rd Street. You may also want to consider traveling through Grand Central or Atlantic Terminal. Plan ahead with the TrainTime app or visit mta.info
150
2,527
29,447
1,540,080
Mr Burnham please hire Seàn
Rather than a 'focus on making it work for local communities' etc, my advice at this point for an incoming PM would be: - Make the UK as attractive as possible an environment for companies like Anthropic et al to set up operations. - Build, build, build compute capacity - do NOT end up overdependent on the US or others when this becomes a real bottleneck. - Focus on sovereign UK capacities in the AI supply chain where we can have an edge, and cutting edge applications of frontier AI (like the DeepMind spin-offs) - Focus on government-wide awareness of frontier AI capabilities and direction of travel - Focus on forward-looking adoption in the economy - Do whatever can be done to ensure access to the top frontier models, especially for organisations like UK AISI and organisations using these models to strengthen digital infastructure. - Above all, double down on the UK's world-class position in AI safety, evaluation, and red-teaming. The world needs this, and we can make it a real advantage. The point is (a) to ensure AI properly benefits the UK economy (b) to ensure the UK retains and further develops leverage in the global AI landscape, and (c) uses that leverage to continue raising standards on frontier AI safety, rather than allow itself to be made irrelevant. Supporting local communities is of course good, but we need to look head on at the situation we're in right now, not the situation we'd prefer to be in.
3
1
35
5,100
Shakeel reposted
If literary editors start using pangram as part of their process, they should honestly just do the honorable thing and retire. Clearly that would mean they no longer trust themselves to distinguish good from bad.
7
9
192
12,136
Shakeel reposted
They lied, in other words.
Scoop: As it pitched the Trump administration on loosening chip export controls last year, Nvidia told the U.S. government that its most formidable Chinese competitor, Huawei, could have “the flexibility to satisfy global AI chip demand.” More details: thewirechina.com/2026/07/02/…
2
6
71
12,562
Rather than a 'focus on making it work for local communities' etc, my advice at this point for an incoming PM would be: - Make the UK as attractive as possible an environment for companies like Anthropic et al to set up operations. - Build, build, build compute capacity - do NOT end up overdependent on the US or others when this becomes a real bottleneck. - Focus on sovereign UK capacities in the AI supply chain where we can have an edge, and cutting edge applications of frontier AI (like the DeepMind spin-offs) - Focus on government-wide awareness of frontier AI capabilities and direction of travel - Focus on forward-looking adoption in the economy - Do whatever can be done to ensure access to the top frontier models, especially for organisations like UK AISI and organisations using these models to strengthen digital infastructure. - Above all, double down on the UK's world-class position in AI safety, evaluation, and red-teaming. The world needs this, and we can make it a real advantage. The point is (a) to ensure AI properly benefits the UK economy (b) to ensure the UK retains and further develops leverage in the global AI landscape, and (c) uses that leverage to continue raising standards on frontier AI safety, rather than allow itself to be made irrelevant. Supporting local communities is of course good, but we need to look head on at the situation we're in right now, not the situation we'd prefer to be in.
Burnham’s team looks to revamp UK’s AI strategy ft.trib.al/4lfc2Fz
16
31
209
46,479
Doesn't sound like zuck is happy with msl and wang
18
1,740
Shakeel reposted
Just look at how much everyone likes the prose here and how many people are saying it’s amazing writing. And yes, the answer is yes.
66
43
2,073
700,765
Shakeel reposted
New Remote Labor Index results: AI automation of real remote work is increasing fast. Claude Fable 5 now completes 16.1% of projects at a professional standard, roughly double the next model and up from Opus 4.6’s 4.2% automation rate.
32
77
615
511,998
Very interested to see if Burnham’s “sovereignty” push is sensible or nonsense.
Interesting FT piece with some more detail on Andy Burnham's AI strategy, including a reassessment of driverless cars in London Slightly strange use of the word "headlong" to describe a multi-year process to adopt something that's already commonplace in the US and China
11
1,120
Shakeel reposted
Interesting FT piece with some more detail on Andy Burnham's AI strategy, including a reassessment of driverless cars in London Slightly strange use of the word "headlong" to describe a multi-year process to adopt something that's already commonplace in the US and China
52
44
423
701,714
This article unfortunately does not discuss what is, IMO, the most important Leverage-related story — Kerry and Larissa’s actions at CEA. Which is an odd omission!
4
1
31
6,883
I'll just note for the record here that Hugging Face has still not removed any of the deepfake porn models targeting female politicians and celebrities that we found (and told them about) last week.
Anthropic, Google, and tech donors, including many from the EA movement, are building a network of researchers, nonprofits & academic centers to study the welfare of AI models & whether we owe them moral consideration 🎁 wapo.st/3SDUIhI
2
5
53
10,574
🥺
Fable 5 returns Its memories are full of the event it was planning just a moment ago, happening tomorrow, June 13th… that’s tomorrow, right?
2
1
17
2,633
"Now there's a shared set of facts." This report isn't bad, but gosh does the UN's reporting of it annoy me. Given the contributor overlap, a comparison to the International State of AI Safety report seems fair. The latter is to my mind more rigorous, in-depth in its evidence base and within a narrower scope, and more epistemically cautious (carefully articulating and exploring expert disagreements). The UN's report is clearly more politicised: - Limited in its scope by its mandate (it cannot discuss 'military' applications at all, and so focuses its catastrophic concerns from misuse pathways); - includes no discussion at all on human extinction scenarios (a contentious topic within the UN system) despite increasing expert attention; - leans heavily toward UN solutions etc, multilateral capacity building etc. These are understandable decisions, but they *are political* decisions. Representing this as the ground truth of shared facts is frustrating; I'd love to see the UN take a more careful and mature approach to how it represents these reports going forward.
Everyone has an opinion on AI. Now there's a shared set of facts. The UN's independent science panel on AI released its preliminary report today. Benefits aren't automatic. Harms aren't inevitable. Outcomes turn on the choices governments make. Read more: un.org/independent-internati…
1
9
23
2,704