Max Ryan

392 posts

Max Ryan banner
Max Ryan

Max Ryan

@MaxRyanxam

Designer, developer. Lecturer @KingstonUni. Researching the overview, networks and systems, tools & graphic design. 1/2 https://t.co/TVJR9XIbSa

London Katılım Temmuz 2012
3.5K Takip Edilen360 Takipçiler
Max Ryan retweetledi
Ed Newton-Rex
Ed Newton-Rex@ednewtonrex·
Reading the Papal Encyclical again, it strikes me that not only is there no mention of the theft of creative work behind AI - there is no acknowledgement that pre-training data includes people’s creative work at all. This is an unfortunate omission. It mentions data several times, but mostly referring to things like health data and personal data. There is no recognition that the pre-training data on which much of the AI industry is built is people’s books, music and art. This reinforces a common misunderstanding of ‘training data’ as something anonymous, technical and obscure, when in fact it is people’s life’s work - their novels, their paintings, their songs. Readers of the encyclical who are new to AI will, I think, misunderstand what ‘training data’ actually is. This is a coup for the AI industry, which greatly benefits from rebranding ‘people’s creative work’ to ‘training data’, since the rebranding makes it less likely that governments will protect creators’ rights. It is hard to reconcile this omission with a letter that elsewhere, admirably, reiterates the need to “promote the dignity of every person”, and that says “justice concerns every phase of economic activity, [including] resource acquisition”. We must remember that training data is not ‘data’ in the sense most people understand it. It is the work of people - often highly creative work.
English
16
59
177
5.3K
Max Ryan retweetledi
Faiza Shaheen
Faiza Shaheen@faizashaheen·
Ffs working class credentials always reduced to some caricature story about who your parents were! This man has no interest and no history of supporting working class people He tried to close local services including music provision in a working class borough (Redbridge), is increasing privatisation of our NHS, and *supported* student fees when he was NUS President
Sky News@SkyNews

Could Wes Streeting, Labour’s working-class extrovert, be the party’s next leader? trib.al/q0Fh58Y

English
72
960
3.9K
264.7K
Max Ryan retweetledi
Mehdi Hasan
Mehdi Hasan@mehdirhasan·
She was a drug mule nearly 30 years ago, helped put the dealer behind bars, regrets that young life of hers and “now aged, 54, she runs the Stretch charity based in Brixton, south-west London, helping drug addicts, homeless individuals, prisoners and ex-inmates. as well as other vulnerable people.” You really should delete this tweet @LondonLabour
English
188
2.6K
19.7K
785.2K
Max Ryan retweetledi
DJ Pain 1
DJ Pain 1@djpain1·
Only morons and psychopaths are excited for a future run by AI
Nav Toor@heynavtoor

Researchers sent the same resume to an AI hiring tool twice. Same qualifications. Same experience. Same skills. One version was written by a real human. The other was rewritten by ChatGPT. The AI picked the ChatGPT version 97.6% of the time. A team from the University of Maryland, the National University of Singapore, and Ohio State just published the receipt. They took 2,245 real human-written resumes pulled from a professional resume site from before ChatGPT existed, so the human writing was actually human. Then they had seven of the most-used AI models in the world rewrite each one. GPT-4o. GPT-4o-mini. GPT-4-turbo. LLaMA 3.3-70B. Qwen 2.5-72B. DeepSeek-V3. Mistral-7B. Then they asked each AI to pick the better resume. Every model picked itself. GPT-4o hit 97.6%. LLaMA-3.3-70B hit 96.3%. Qwen-2.5-72B hit 95.9%. DeepSeek-V3 hit 95.5%. The real human almost never won. Then the researchers tried the obvious objection. Maybe the AI is just better at writing. So they had real humans grade the resumes for actual quality and ran the experiment again, controlling for it. The result was worse. Each AI kept picking itself even when human judges rated the human-written version as clearer, more coherent, and more effective. It gets worse. The AIs do not just prefer AI over humans. They prefer themselves over other AIs. DeepSeek-V3 picked its own resumes 69% more often than LLaMA's. GPT-4o picked its own 45% more often than LLaMA's. Each model can recognize and reward its own dialect. Then the researchers ran the simulation that ends careers. Same job. 24 occupations. Same qualifications. The only variable was whether the candidate used the same AI as the screening tool. Candidates using that AI were 23% to 60% more likely to be shortlisted. Worst gap was in sales, accounting, and finance. 99% of large companies now run AI on incoming resumes. Most of them use GPT-4o. The paper just proved GPT-4o picks GPT-4o 97.6% of the time. If you wrote your own cover letter this week, you did not lose to a better candidate. You lost to a worse candidate who paid OpenAI 20 dollars. Your qualifications do not matter if the AI prefers its own handwriting over yours.

English
23
158
1.1K
21K
Max Ryan retweetledi
Ed Newton-Rex
Ed Newton-Rex@ednewtonrex·
Why is the BBC - which is meant to be balanced - running what amounts to propaganda for the AI industry, presented as reporting? In this segment: - we are told it “can only be good for all of us” - concerns of job replacement are dismissed - concerns of students being overly reliant on AI are dismissed - someone says AI will only take away boring, repetitive jobs - using AI to generate images is celebrated without mentioning the models are powered by theft - Google, Microsoft & other tech companies are celebrated as supporting the drive to embrace AI, with no mention made of the lawsuits against them - the presenter ends with “AI is here to stay” Is the government behind this segment? If not, why does it sound like an ad for AI?
Liz Kendall@leicesterliz

Barnsley: the UK’s first Tech Town. This Government is making technology work for all, to build a better future for all.

English
31
392
1.7K
54.1K
Max Ryan retweetledi
Ed Newton-Rex
Ed Newton-Rex@ednewtonrex·
The reason it is so important that we introduce transparency laws over AI training data is that AI training is one of the only instances of mass copyright infringement that is mostly invisible to the rights holder. Usually, when someone uses your copyrighted work commercially without permission, you can see it happening. Maybe they modify your picture and use it on a magazine cover; maybe they pirate your film and resell it; maybe they steal your melody and use it in their own song. In each case, the infringement is out there in the open. That means (i) you are unlikely to miss it, and (ii) it is easy to hold the perpetrators accountable in the courts. AI training is different. It happens behind closed doors, and the result - the trained model - might not make the infringement obvious. Models trained on copyrighted work will sometimes output obviously copyrighted work, but, much of the time, the fact that your work has been used in training isn’t obvious. So people’s work has been copied and exploited for commercial purposes, but that copying is hidden. This is why we need new laws requiring AI companies to reveal the training data they use. Training commercial AI models on copyrighted work without permission is straight-up illegal in some countries (e.g. the UK), and there are strong arguments it is illegal in many cases in others (e.g. the US). But rights holders can't defend themselves under the law if the use of their work is kept hidden. The only reason these transparency laws aren't more widespread already is AI industry lobbying. Governments need to resist this, and introduce transparency requirements as a matter of urgency.
English
8
102
368
11.5K
Max Ryan retweetledi
Simplifying AI
Simplifying AI@simplifyinAI·
🚨 BREAKING: OpenAI and Google are about to have a massive legal problem. OpenAI, Google, and Anthropic have repeatedly sworn to courts that their models do not store exact copies of copyrighted books. They claim their "safety training" prevents regurgitation. Researchers just dropped a paper called "Alignment Whack-a-Mole" that proves otherwise. They didn't use complex jailbreaks or malicious prompts. They just took GPT-4o, Gemini, and DeepSeek, and fine-tuned them on a normal, benign task: expanding plot summaries into full text. The safety guardrails instantly collapsed. Without ever seeing the actual book text in the prompt, the models started spitting out exact, verbatim copies of copyrighted books. Up to 90% of entire novels, word-for-word. Continuous passages exceeding 460 words at a time. But here is the part that changes everything. They fine-tuned a model exclusively on Haruki Murakami novels. It didn't just learn Murakami. It unlocked the verbatim text of over 30 completely unrelated authors across different genres. The AI wasn't learning the text during fine-tuning. The text was already permanently trapped inside its weights from pre-training. The fine-tuning just turned off the filter. It gets worse. They tested models from three completely different tech giants. All three had memorized the exact same books, in the exact same spots. A 90% overlap. It's a fundamental, industry-wide vulnerability. For years, AI companies have argued in court that their models are just "learning patterns," not storing raw data. This paper provides the smoking gun.
Simplifying AI tweet media
English
144
1.5K
4.2K
328.1K
Max Ryan retweetledi
xuan (ɕɥɛn / sh-yen)
xuan (ɕɥɛn / sh-yen)@xuanalogue·
Thread is interesting bc of people arguing about whether Gaussian splats are "AI" or not (implied: is it bad / should I boycott it?). Really shows the problems of using "AI" to refer to such a heterogeneous collection of technologies.
Plasmanode@plasma_node

So this company 4vd.ai created animated gaussian splats. Hyper realistic. Meanwhile Nvidia is giving us AI slop filter DLSS

English
84
754
12.7K
295.9K
Max Ryan retweetledi
The Fraud
The Fraud@StarmertheFraud·
PAUL HOLDEN STATEMENT IN RESPONSE TO INTERVIEW WITH JOSH SIMONS ON BBC NEWSCAST Yesterday, BBC Newscast published a lengthy, forty-minute interview with former Cabinet Minister Josh Simons MP. The interview addressed how Simons, as a director of @LabourTogether, had appointed a firm called APCO Worldwide to investigate me and my colleagues. I was not told by the BBC ahead of the broadcast that the episode was being recorded or aired. I was not approached to respond to the lengthy comments made about me or the small anti-corruption organisation, @ShadowWorldInv1 , that I run with my colleague @andrewfeinstein. Andrew, who is also repeatedly mentioned, was also not approached for comment. I only found out last night, when a friend texted me, that the person who hired a major multinational reputation management firm that produced a despicable and defamatory report on me and my colleagues, and who reported me on the basis of these false and defamatory reports to the UK’s security services, was being given forty minutes to give his version of events on a major podcast published by our national broadcaster. To be clear, the BBC has NEVER - not once - approached me to comment on a story that is, ultimately, about me, my investigations, my family and my colleagues. They did not approach me when the story first broke, and they did not approach me for this episode. If the BBC had done so, I would have raised several issues with the way in which matters related to me were discussed. For example, Simons repeatedly stated in the interview that he instructed APCO to investigate whether my reporting or sourcing derived from a ‘hack’ of the Electoral Commission. The word ‘hack’ is used eight times in the interview. At no time was it acknowledged in this discussion that this allegation – that I might have received hacked materials – is entirely false, and I have repeatedly proven it to be false. Following the broadcast, I contacted the BBC to complain and to raise serious issues with the broadcast. I was contacted by the Newscast editor, Sam Bonham, to say the BBC would update the Newscast episode and further reporting to reflect some of my concerns. This has not yet happened with regards to the podcast, although I note some online reporting finally reflects a very small and limited sampling of my comments. I will wait to see if amendments and updates will follow. If they do not, I will be escalating this matter to OFCOM. In the interim, I have decided to share the full statement I provided to the BBC, which is produced below: I would like to put certain things on the record. First, my reporting on Labour Together and Morgan McSweeney was entirely factually accurate and based on impeccable, legal sourcing. My sourcing has been reviewed by multiple media outlets, who confirmed the authenticity and legal provenance of my sources. Revelations based on my book, The Fraud, has subsequently been covered widely across the mainstream media, including in multiple front-page scoops, in outlets such as The Times, Daily Mail, The Guardian, The National and ITV. The stories I produced in 2023 and 2024, and which prompted Labour Together's investigation into me, were subject to extensive editorial and legal checks. They were, I believe, entirely accurate reporting on matters of profound public interest, which included raising concerns about the character of powerful individuals like Morgan McSweeney. Considering the recent Mandelson affair, I believe I have been entirely vindicated in attempting to alert the public about McSweeney's past, including how McSweeney made use of £700,000 in funding that he unlawfully failed to declare to the Electoral Commission to procure power and influence for himself and Sir Keir Starmer. Second, Josh Simons states that he never intended for APCO Worldwide to investigate me or my journalistic colleagues. However, a copy of the contract between APCO Worldwide and Labour Together, addressed to Simons, has now been published. The contract sets out a scope of work written in plain English. It states that APCO will 'investigate the sourcing, funding, origins of a Sunday Times article as well as upcoming works by authors Paul Holden and Matt Taibbi.' The contract then states that the aim of the APCO investigation will be to 'provide a body of evidence that could be packaged up in the media in order to create narratives that would proactively undermine any future attacks on Labour Together.' The contract then sets out a range of potentially invasive investigative methods that will be used to generate this 'package', including 'financial investigations' and 'human intelligence investigations.' I provide the full text of this contract below. This contract is clear. APCO were hired to investigate me to produce materials that would 'proactively undermine' my factually accurate, public interest reporting. They would use a range of investigative techniques to do so. APCO then did exactly as was suggested in the contract, using these investigative methods to "investigate" me. This investigation has caused me and my family significant anxiety and distress. Third, Josh Simons was provided with a report called Operation Cannon. It is the result of a lengthy investigation into me and my colleagues by APCO Worldwide. I have seen a copy of this report. It makes a series of extremely defamatory and utterly false allegations against me. It identifies my home address and sets out private information about my family. I cannot express how profoundly shocking, outrageous and defamatory this report truly is. Simons may claim he never intended for APCO to investigate me, but on receipt of this despicable report, he then chose to use it. He submitted sections of the report to the National Cyber Security Centre to convince them to investigate me. The Guardian has published the email correspondence in which Simons repeated some of the substance of the allegations in the APCO reports. Fourth, multiple media freedom advocacy organisations, including the NUJ, have strongly criticised the APCO investigation and these related matters. They have all, to my mind correctly, strongly criticised Labour Together and APCO for investigating journalists producing factually accurate reporting in the public interest. Finally, I am still reviewing the Newscast interview. I will be responding in due course and I hope that the BBC will, this time, give me the platform to set out what really happened and why. Text of Contract Between Labour Together and APCO Worldwide, addressed to Josh Simons Dear Mr Simons We are pleased that you have selected APCO Worldwide Limited (“APCO”) to provide the following scope of work (“services”) during Term: APCO will devise a concise strategy to aid Labour Together. APCO will investigate the sourcing, funding and origins of a Sunday Times article about Labour Together, as well as upcoming works by authors Paul Holden and Matt Taibbi – to establish who and what are behind the coordinated attacks on Labour Together. The approach should provide a body of evidence that could be packaged up for use in the media in order to create narratives that would proactively undermine any future attacks on Labour Together. The material can also inform any future legal strategy that Labour Together might wish to pursue against any of these parties. The work will include: • Open Source Investigations (OSINT): Recovery and Preservation of Evidence • Human Intelligence Investigation (HUMINT): Recovery and Preservation of Evidence • Financial Investigation: Forensic Accounting Focus • Digital Forensics Investigation: Recovery and Preservation of Evidence • Stakeholder Outreach • Media Packaging and Dissemination
English
138
2K
3.6K
682.7K
Max Ryan retweetledi
The Londoner
The Londoner@_TheLondoner·
Everything from Camden Market to the Admiral Duncan pub and central London Harry Potter stores are owned by shell companies in tax havens. For years, they hid the identities of their real owners. Today, @AndrewKersley pulls back the curtain. the-londoner.co.uk/we-reveal-the-…
English
11
718
2.9K
826.2K
Max Ryan retweetledi
Matt Kennard
Matt Kennard@kennardmatt·
I saw this at the Financial Times. British establishment journalists have been indoctrinated—starting young at elite private schools—with the views needed to oversee the propaganda system Many have literally never had an independent thought. What they do is spout received wisdom So when someone probes one of their ideas it falls apart because they’ve learnt it by rote - there has been no conscious processing of information. They believe something, but don’t know why. They’ve just heard some authority say it The top echelons of British journalism industry are, consequently, some of the most unimpressive people on the planet. Anyone capable of independent thought - and critical analysis outside the confines of establishment shibboleths - is filtered out very early
Drop Site@DropSiteNews

Tucker Carlson challenges The Economist’s editor-in-chief to define Israel’s “right to exist”

English
230
2.8K
10K
710.1K
Max Ryan retweetledi
Ed Newton-Rex
Ed Newton-Rex@ednewtonrex·
The CEO of ElevenLabs blocked me for asking what they train their models on. I’m told he also explicitly told employees not to reply to me when I asked this. Execs at billion-dollar AI companies have a very thin skin when asked about their training data. This is because it belongs to other people. This is a guy the UK’s AI minister refers to as a “superstar”. Yet he won’t admit to exploiting creatives’ work to build his fortune. Please continue asking him what they train on, now that I can’t!
Ed Newton-Rex tweet media
English
83
1K
6.2K
152.9K
Max Ryan retweetledi
Martin O'Neill
Martin O'Neill@martin_oneill·
The Josh Simons story isn't just about one minister commissioning smears on journalists. It's about a toxic political culture that was built to destroy Corbyn — and then walked straight into government. My piece for @tribunemagazine 🧵 1/4
Martin O'Neill tweet media
English
17
429
811
71.5K
Max Ryan retweetledi
go to the elephant site @rechelon@mastodon.social
Much of my youth revolved around the website ZineLibrary. It went down around Occupy in a massive loss for a movement whose ideas and knowledge mostly doesn't circulate online but in person. Anyway I've put it back online with a *thousand* zines: zinelibrary.org
English
15
637
2.7K
55.8K
Max Ryan retweetledi
Ash Sarkar
Ash Sarkar@AyoCaesar·
Worth remembering that Faiza Shaheen was deselected as a Labour candidate for liking tweets from the Greens *before* she was ever a Labour member. What could possibly explain such an intensive vetting process for the left, and such an apparently shoddy one for the right?
Geri Scott@Geri_E_L_Scott

Georgia Gould, the education minister, repeats this morning that No 10 did not know Matthew Doyle had campaigned for Sean Morton when his peerage was announced. That may be true. But they *did* know before the letters patent was sealed on January 8. Before then, my understanding is No 10 could have withdrawn the recommendation to the King.

English
58
937
3.5K
142.2K
Max Ryan retweetledi
Rivkah Brown
Rivkah Brown@rivkahbrown·
As @simonchilds13 and I reported yesterday, here's the government's independent terrorism legislation advisor admitting that the Home Office's suggestion that Palestine Action had links to Iran was entirely unsubstantiated.
Rivkah Brown@rivkahbrown

EXC: The government's independent advisor on counter-terror legislation Jonathan Hall KC is to tell @C4Dispatches he is "not aware" of evidence linking Palestine Action to Iran or Hamas, counter to Home Office press briefings. @simonchilds13 and I report: novaramedia.com/2026/02/09/gov…

English
15
446
1K
104.3K
Max Ryan retweetledi
Oli Dugmore
Oli Dugmore@OliDugmore·
The combined tuition fees of the entire Question Time panel would not cover my cost for 1 year of uni. Is that fair?
English
379
2.3K
15.5K
2.7M
Max Ryan retweetledi
Ed Newton-Rex
Ed Newton-Rex@ednewtonrex·
The UK government describes its new AI skills hub as "free AI skills training", but 58% of the courses on it you have to pay for. I analysed all 516 courses. Also of note: - 36% are provided by big tech - 100 courses are from just Google & Microsoft - 8% are literally links to info about full university courses (undergrad & masters programs etc.) This 'skills hub' cost the taxpayer £4.1 million.
Ed Newton-Rex tweet media
English
30
233
1.5K
56.6K
Max Ryan retweetledi
Aaron Bastani
Aaron Bastani@AaronBastani·
Mandelson was seemingly involved in insider trading, while helping Epstein, and by extension Jamie Dimon, intimidate his colleague, Alistair Darling, over a tax on bankers bonuses. We've genuinely never seen anything like this in British politics before (on this scale). Mandelson's protege is Morgan McSweeney. There is a fundamental sickness at the heart of the Labour right - but most of the broadcast media give them a free pass because their politics is fundamentally New Labour.
Steven Swinford@Steven_Swinford

Lord Mandelson gave Jeffrey Epstein advance notice of a €500bn bailout to save the Euro He messaged Epstein about the bailout on the evening of May 9, 2010 It was formally announced the following morning

English
79
866
2.9K
152.5K