Sign in

Richard Moyse Fenning

@mfrich.bsky.social
660 followers 2.1K following 77 posts
PostsRepliesMedia
Reposted by Richard Moyse Fenning
Butlerian Jihadist of House Dornick @shimminykricket.blacksky.app · 08/10/2026
Harvard just dropped a study AI productivity in software engineering and it found the average increase in productivity from using Claude code is.....zero and for exactly the reasons you'd think fion.ac/jellyfish.pdf
9862412436
Richard Moyse Fenning @mfrich.bsky.social · 07/10/2026
There was a good interview with polanski talking about this recently. Every day he replies to a list of questions he gets sent by the press, but the only things they print are the 'gaffs'
0110
Reposted by Richard Moyse Fenning
Annie Gottfried Cohen @annsplaining.bsky.social · 06/10/2026
The Party conference voted through a motion that declares Zionism *as it defines it* to be a form of racism. It also adds a few practical demands to the Party's existing position on Palestine, notably the call for the creation of a single democratic state
49118
Reposted by Richard Moyse Fenning
Dana El Kurd @danaelkurd.bsky.social · 07/10/2026
Three years of nightmare, and it doesn't end. It doesn't end. We are all irreversibly damaged.
232161
Reposted by Richard Moyse Fenning
Rob Delaney @robdelaney.bsky.social · 17/07/2026
this is you in meta’s pervert glasses
465029845
Reposted by Richard Moyse Fenning
Carole Cadwalladr @carolecadwalla.bsky.social · 30/09/2026
NEW: ‘It’s an implicit threat’. Lord Tom Watson of Palantir writes to one of its biggest critics -Labour MP @labourlewis.bsky.social - to tell him the surveillance firm is watching him. www.thenerve.news/p/clive-lewi...
thenerve.news
They’re saying: we’re watching you’: MP Clive Lewis reveals Tom Watson has warned him over criticising Palantir
In a letter to his former Labour colleague, the peer, who joined Peter Thiel’s AI company this month, writes: ‘You have made the case against the company more or less continuously’. By Carole Cadwalla...
37683453
Reposted by Richard Moyse Fenning
The Guardian @theguardian.com · 29/09/2026
Trial of live facial recognition in London stations ends with a false positive and no arrests
theguardian.com
Trial of live facial recognition in London stations ends with a false positive and no arrests
Freedom of information request finds six-month trial cost £320,000, used almost 100 police hours and led to just one – incorrect – alert A six-month trial of live facial recognition (LFR) technology in London’s railway stations that cost more than £320,000 and almost 100 hours of police officers’ time ended in one false match against a watchlist of suspects and no arrests. More than half a million faces were scanned between February and July this year in some of the capital’s busiest transport hubs during the British Transport Police (BTP) trial of the surveillance technology, which aimed to help catch offenders and people breaching court orders. Continue reading...
12138107
Reposted by Richard Moyse Fenning
Natacha @natacha.bsky.social · 29/09/2026
Green Party surge in New Zealand could be crucial, opening the possibility of a potential coalition with Lab & Opportunity to defeat the hard right coalition.
ASIA PACIFIC / POLITICS
New Zealand election on knife edge as Greens surge to 16%, poll shows
0669
Reposted by Richard Moyse Fenning
Kat Tenbarge @kattenbarge.bsky.social · 28/09/2026
What about her future? What about how hard she worked to be there? What is to become of her life now? What process of justice could possibly right these wrongs? And how much more will she lose in the process of trying to get it? How much more of this are we all supposed to tolerate?
61624223
Reposted by Richard Moyse Fenning
Kat Tenbarge @kattenbarge.bsky.social · 28/09/2026
One of the Cornell students accused of gang rape, Scott Kretzschmar, denies it and says through his attorney that they are going to “confront” the survivor through the court system. I can already tell you that means victim-blaming, DARVO, and smears. She is going to need a lot more of our support.
443082689
Richard Moyse Fenning @mfrich.bsky.social · 28/09/2026
Whoever booked him subconsciously remembered this
0132
Richard Moyse Fenning @mfrich.bsky.social · 28/09/2026
It makes me kind of nostalgic. The scientists building the LHC were worried about conspiracy theories because they thought it might limit future funding. The LLM companies are encouraging the doom hype either for financial reasons or because they're part of a millenarianist sex cult
000
Richard Moyse Fenning @mfrich.bsky.social · 28/09/2026
But there was no social media boosting the black hole stuff - just journalists, so it stayed pretty niche. There also aren't any other issues at CERN that might actually end the world (like accelerating climate change, creeping construction of the panopticon and a huge financial bubble)
100
Richard Moyse Fenning @mfrich.bsky.social · 28/09/2026
Listening to @maintenancephase.bsky.social discuss worries about CERN creating world swallowing black holes when the LHC was being built and thinking about the current AI worries. The feeling of listening to people as sci-fi narratives to things without knowing how they work is the same
100
Reposted by Richard Moyse Fenning
Robin Moira White @robinmoirawhite.bsky.social · 27/09/2026
The Labour Party’s women’s conference on 26/9/26 was an embarrassing farce, with only 200 delegates of the 1,000 normal attendance (no major unions) in protest of the Labour Party’s shameful abandonment of trans people following the legally absurd ruling of the Supreme Court in FWS in April 2025.
121273283
Reposted by Richard Moyse Fenning
Carole Cadwalladr @carolecadwalla.bsky.social · 28/09/2026
NEW: Meet the real AI ‘rogue agent’: Lord Watson of Palantir. I am sick to the teeth of this current AI hype cycle & its perversion of language. The only thing going ‘rogue’ right now are AI boosters, a gullible & compliant press & a Labour Lord 1/ broligarchy.substack.com/p/meet-the-r...
broligarchy.substack.com
Meet the real AI “rogue agent”: Lord Tom Watson
A Labour leader is out of control. The world needs a containment strategy now
301015402
Richard Moyse Fenning @mfrich.bsky.social · 28/09/2026
She checked twitter every morning for years, and every morning she put her phone down disappointed
040
Richard Moyse Fenning @mfrich.bsky.social · 27/09/2026
The trailers for that disaster simulation TV show with Michael Gove keep reminding me of this cartoon (and when Twitter was still good) www.theguardian.com/lifeandstyle...
theguardian.com
'Let me do it, David': the Stephen Collins cartoon
Independence fey
000
Reposted by Richard Moyse Fenning
elia ayoub @ayoub.bsky.social · 22/09/2026
Israel annihilated *all* of their schools and universities, but the folks over at Jerusalem Post wanna know: hey, are those kids being taught to hate us? Your brain on Zionism. You can exterminate at least 20,000 children in 3 years and still see yourself as the f-ing victim.
=a

THEJERUSALEM POST

Gaza's back-to-school coverage misses the crucial test: what are children being taught?
4286102
Reposted by Richard Moyse Fenning
Emma Yeomans @yeomans.bsky.social · 22/09/2026
Worrying scenes in the Channel today, for all interested in asylum policy and maritime tactics. A small boat from Normandy spent more than 30 hours at sea. Those on board now on their 2nd rescue vessel as Border Force forced to play cat and mouse with far-right activists who have their own boat.
1042174958
Reposted by Richard Moyse Fenning
Tom Roberts @tpgroberts.bsky.social · 22/09/2026
The complete and total collapse of state function in the face of far right extremists. It is frankly ludicrous that Daniel Thomas continues to be at large given the series of crimes committed which he has literally filmed.
281173380
Reposted by Richard Moyse Fenning
Mark Kersten @markkersten.bsky.social · 21/09/2026
I’ve been told this too, from UN sources - the US will sanction the #ICC as an institution this week, & that they have considered designating the Court as a terrorist or terrorist supporting body. All for Netanyahu (and Putin). States that believe in the rule of law must denounce this lunacy.
613362
Reposted by Richard Moyse Fenning
elia ayoub @ayoub.bsky.social · 21/09/2026
Talking about how I used to enjoy the internet
static.klipy.com
Elrond's Famous Line: I Was There Gandalf
Alt: Elrond's Famous Line: I Was There Gandalf
126834
Richard Moyse Fenning @mfrich.bsky.social · 20/09/2026
It's just about how these llms function. There's explicitly no understanding step. There is machine learning tech that has or builds world models, but the big leap with the type op is talking about is that's left out. The understanding is in the training data, and in your brain
000
Reposted by Richard Moyse Fenning
Fared Al Mahlool | فريد المحلول @faredalmahlool.bsky.social · 19/09/2026
This video was filmed in Palestine. My heart struggles to comprehend such cruelty and the heartbreaking disregard for human life. No human being should have to endure this.
5629461403
Reposted by Richard Moyse Fenning
Paul F. Tompkins @pftompkins.bsky.social · 19/09/2026
You can’t have it both ways. You can’t be “It’s simple math, vote blue no matter who, grow up” and ALSO believe you are fighting fascism by making calculations as to who’s expendable. The world we’re in now demands more from us. More from everyone.
435271712
Richard Moyse Fenning @mfrich.bsky.social · 19/09/2026
Dude
110
Richard Moyse Fenning @mfrich.bsky.social · 19/09/2026
It's just imprecise language that plays into our tendency to our tendency to see faces in clouds imo
110
Richard Moyse Fenning @mfrich.bsky.social · 19/09/2026
That's true for any one response - it's specifically designed like that - but there's no mystery to how they work. There's a long history of maths and computer science that explains it. If there weren't manuals and writing about similar things in the training data, it wouldn't have got anywhere
100
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
Yes
200
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
Please read his original post. He's not talking loosely, and also not using metaphor to illuminate a deeper concept or anything - he's obscuring how these things work. (How human brains work is also very interesting btw - but feels a little tangential)
000
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
Okay - whatever it's called when coders get together and do competitive hacking games
100
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
It's not beside the point of the person I was replying to said that it had to have comprehensive understanding to work. It doesn't - and that's partly why it's so amazing
250
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
What point at hand? I was disagreeing with the sci-fi characterisation in the original post of how chatbots work. The whole point is that they don't understand anything. There is machine learning tech that uses world models, but one of the major strengths of this kind of tech is that it doesn't
110
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
Nice! It didn't visualise a conceptual model and explain it to you though - it gave you a string of probable words based on your prompt and the training data
110
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
Understand what the tech is and how it works?
100
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
The language we use is important imo, because there's a sci-fi doomsday cult trying to convince us all that agi is just around the corner, and get us to make policy and give cash based on that
000
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
It's really incredible technology that can make use of more knowledge than any human could read in a hundred lifetimes, but it still doesn't understand anything
300
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
There is technology that used world models, but this tech doesn't do that
000
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
Whether the person I was replying to was right in saying it had a conceptual understanding of the problem
201
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
You're really anthropomorphizing here. It's theory free inference - the point is it doesn't understand anything. It's amazing and surprising it can do so much, but the one thing it's explicitly designed not to do is understand
8311
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
But the reason they ran so many was because they wanted unpredictable things to emerge - so yes, but also that was on purpose. It's super interesting tech, but the original article (or the headline at least), is correct in saying the immediate reporting probably overplayed it's importance/spookiness
110
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
Sure
000
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
Yes - it's all bad - but I think the point is that it still needs people to be bad or reckless. Talking about python/chatbot loops as if they have schemes and agency just lets these companies off the hook and increases their share prices
110
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
You're misunderstanding what I'm saying. LLMs are probable next word machines. They do theory-free inference based on correlations in their training data, including the hacking logs. They don't just regurgitate the words, but where else do they get the probabilities they work from?
180
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
No one's saying they just copied chat logs, but it's obviously important that they're in the training data, when the task that was set was a hacking game
340
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
Also - what do you mean by "they knew"? "They" were increasingly long llm prompts
000
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
You're right, they weren't simply copying, they were calculating the most probable next token based on their training data - which includes hackathon logs. The whole thing was done to find novel hacking exploits, so it makes sense they would sound like a team of hackers in the text they generated
580
Richard Moyse Fenning @mfrich.bsky.social · 18/09/2026
From listening to Cory Doctorow - there is hackerthon chat logs in the training data, and they were instructed to play a capture the flag type hacker game cooperatively. He says the scary thing is not the emergent properties, which is old news for cellular automata, but how bad the cyber security is
5240