Mamdani continually playing Trump in order to get things done rather than screaming about how heβs a fascist - heβs all Iβve wanted in a left wing politician for a decade.
26.02.2026 22:35 β π 1 π 0 π¬ 1 π 0Mamdani continually playing Trump in order to get things done rather than screaming about how heβs a fascist - heβs all Iβve wanted in a left wing politician for a decade.
26.02.2026 22:35 β π 1 π 0 π¬ 1 π 0Kore-eda with more films in the letterboxd 500 than Nolan is a victory for humanity
25.02.2026 02:42 β π 1 π 0 π¬ 0 π 0Ppl who keep track of whatβs cancelled the same way football fans follow the transfer windowβ¦
14.02.2026 14:52 β π 0 π 0 π¬ 0 π 0
Thatβs all Starmerβs Labour was ever truly based on. Plus, of course, a remarkably hated Tory Party.
Unless these foundational issues are remedied, and we get a Labour leader with a personality and an interesting vision, itβs Reform come the next election. 2/2
This gets to the root malaise of the Starmer government. You canβt govern, least of all in these times, if your foundation is basically factional - if your ascent to power was built not on ideas, or even personality, but instead an incredibly effective purge of the left within the Labour Party. 1/2
09.02.2026 00:14 β π 1 π 1 π¬ 1 π 0Corbyn is currently up in his allotment with the biggest grin on his face
04.02.2026 17:31 β π 5 π 0 π¬ 0 π 0The money isnβt really in individual users/chatbots anymore - itβs in the API/corporate use, and thatβs also where thereβs greatest technological uncertainty. Returning to the initial point, technological advances substantially change the API/corporate side of things.
29.01.2026 23:18 β π 0 π 0 π¬ 1 π 0
then the valuations that are at the core of the βbubbleβ then look rather more justified.
If with each successive new model, returns donβt increase sufficiently rapidly, and eg openAI defaults on one of its many v large circular loans, weβre all fucked.
IMO wastage of investment is much less important re the bubble than the returns on investment. If with each successive wave of investment - not just in infrastructure but also pretraining - we see v rapidly rising returns such that the large companies get ever closer to breaking evenβ¦
29.01.2026 16:39 β π 1 π 0 π¬ 2 π 0Oh defo, I just think with AI the extent to which thereβs been overbuilding is unclear because itβs unclear exactly how useful the technology will be economically in 6, 12 months. So the bubble isnβt something thatβs fixed, itβs something that exists in dialogue w this constantly changing technology
29.01.2026 03:40 β π 1 π 0 π¬ 1 π 0Think powerful economic use cases are v tied to whether itβs a bubble - if AI automates a large swathe of coding, the investment in AI infrastructure is much more likely to pay off. And thatβs before getting to other use cases that will grow as AIβs jagged frontier expands
29.01.2026 00:59 β π 0 π 0 π¬ 1 π 0
And itβs incredibly difficult to predict what precisely AIβs jagged frontier will look like in 6 months, let alone a year. 3 months ago, all I heard was talk of stagnation.
We inhabit a profoundly unstable technological conjuncture, and thatβs something to be embraced.
This is a key reason why I donβt trust anyone who confidently asserts AI is a bubble or not - so much depends on how rapidly the technology advances. Quick progress in economically useful areas like coding could substantially alter the financial equation.
28.01.2026 21:14 β π 1 π 0 π¬ 2 π 0If youβre gonna recommend a first foreign film to someone, itβs gotta be The 400 Blows - immediately disabuses anyone of the notion that B&W French films must be cold, inaccessible and pretentious. Itβs somehow fresher, funnier, and more devastating than any 21st century movie Iβve seen.
26.01.2026 22:01 β π 1 π 0 π¬ 0 π 0Seeing ppl debate whether Sinners or One Battle After Another will be discussed in decadesβ¦ the clear answer is neither. 2025 was not a good year for film - at least in America. Iβm waiting on the foreign releases
09.01.2026 21:22 β π 1 π 0 π¬ 0 π 0Should be of significant interest to the left, the latter less so though perhaps more now in the era of AI. The former represent imo a key node in the argument over the abundance agenda.
07.11.2025 17:08 β π 0 π 0 π¬ 0 π 0
So I would personally propose that...
- Huge state backed infrastructure development
- Huge partly state fueled housing boom
- Significantly state-backed investment in renewable technologies and key industries
- Large state sector partly to ensure large-scale employment...
No new American frontier AI labs have emerged this year, so cue Moonshot AI. And apparently there are several other Chinese labs in the mix beyond Alibaba and DeepSeek. Chinaβs unique blend of innovation - produce things to roughly the same quality as the West at a fraction of the cost
07.11.2025 16:45 β π 0 π 0 π¬ 0 π 0There is of course one great exception to this: China.
07.11.2025 14:25 β π 0 π 0 π¬ 1 π 0And how could he? The left has long lost any serious tradition of thinking about governance and economics. Gone are the Marxist and Kenyesian traditions, replaced by a never-ending focus on identity. As I argued from 2016, the left is only as good as its ideas, and it is now largely without them 2/2
07.11.2025 14:22 β π 0 π 0 π¬ 2 π 0Zach Polanski is clearly far more skilled than the Your Party lot at articulating the defining problems of modern Britain. But Iβm deeply unconvinced that he has any of the real solutions. 1/2
07.11.2025 14:22 β π 0 π 0 π¬ 1 π 0This election is the one time Iβve appreciated the American penchant for endless democracy. Give me another 100 days. I need more Curtis Sliwa. Far more. Even more of Cuomoβs public humiliation kink. And above all more and more and more Zohran.
04.11.2025 16:45 β π 0 π 0 π¬ 0 π 0I absolutely hate that this illness has made me semi chronically online as I have so few other distractions. Itβs one of the most off-putting things someone can be. I miss when it felt like I existed as much in the 19th century as our own, in the books I read, the music I played and listened to.
02.11.2025 13:38 β π 3 π 0 π¬ 0 π 0Itβs such a shame that none of the great British Marxist historians remain. Perhaps no one would be better suited to probe the rise of AI than a school of historians deeply interested in changing modes of production.
16.10.2025 14:56 β π 0 π 0 π¬ 0 π 0AI is a tabula rasa onto which people can project their views. Like the rise of the internet, AI is so far-reaching that anyone can find a whole range of data points to support their worldview. As a result there is a lot of noise, far less understanding.
16.10.2025 14:56 β π 0 π 0 π¬ 1 π 0If China is a βnational security threatβ to the UK, then what on earth is Trumpβs America?
15.10.2025 12:40 β π 1 π 0 π¬ 0 π 0The one thing I like about AI investment vs the internet - a v large proportion is being spent actually building things. Makes me a little more sanguine about whether weβre in a bubble (which we probably are).
13.10.2025 13:34 β π 0 π 0 π¬ 0 π 0That graph is purely weekly active users of ChatGPT - which people have to choose to use.
10.10.2025 14:36 β π 0 π 0 π¬ 0 π 0Extraordinary
10.10.2025 14:21 β π 0 π 0 π¬ 1 π 0Reading the literature - Iβm often left with the same feeling, that thereβs a specific way Iβm meant to feel about and talk about my body. And I kind of hate it.
10.08.2025 10:41 β π 3 π 0 π¬ 0 π 0