Sir Keir Starmer hopes he has averted the biggest rebellion of his premiership by agreeing to dilute his controversial welfare reforms, but he will be left counting the political and financial costs for years to come.

The sudden capitulation came during hastily-arranged, surreptitious talks with senior Labour MPs in Westminster on Thursday. It will carry a hefty price tag of more than £1.5bn — and invite questions over how a Labour government with a gargantuan majority has found itself so close to a House of Commons defeat.

When more than 100 Labour MPs vowed on Monday to vote against Starmer’s welfare bill, the guerrilla operation sent shockwaves through Downing Street.

Despite warnings from the whips’ office about the brewing rebellion — and one of the whips resigning last week over it — the prime minister’s team had miscalculated how many MPs were prepared to “go rogue”. 

They had also underestimated their backbench MPs, who had organised furtively with only quiet chats and notes on paper. WhatsApp groups, so ubiquitous in Westminster, were banned to avoid leaks. 

But the extent to which ministers, political operatives and whips sought to crush the mutiny raised eyebrows even among veterans of parliamentary struggles: “This has created a civil war which will last for the next three years,” said one MP.  

On Wednesday Starmer, focused on a Nato summit, was still publicly dismissing concerns about the uprising, but the party machine was panicking. Defeat on a key vote less than a year into power would risk fatally undermining the Starmer project.

Soon after his return from The Hague, the prime minister signalled a U-turn was coming, and by Thursday afternoon was thrashing out the details of a potential compromise with senior MPs, including chairs of select committees.

On Thursday evening, as the concessions emerged, rebel Labour MPs were discussing whether they were enough to make them walk back from the brink.

“These are significant changes, they’ve listened and they’ve heard and they’ve made the changes,” said one MP. “I think colleagues will be won over.”

Starmer offered to limit cuts to the main disability benefit (Personal Independence Payment, or Pip) to new claimants at a cost of £1.5bn-a-year, bring forward a £1bn package of employment support payments to this year, and consult on the reforms, according to MPs. Together that will hugely reduce the intended £4.8bn savings from the original package.

It was a major compromise, at odds with attempts earlier in the week to browbeat rebel MPs into submission. But some rebels were still uncertain, with some concerned that new, younger claimants of Pip would still be hit with lower benefits.

Rachael Maskell, the MP for York Central, said the bill “cannot and must not be saved”, adding: “Any concessions will still cause harm to disabled people. The only option is to withdraw the bill, rethink and start to rebuild trust with disabled people.”

After belatedly recognising the scale of the crisis, senior government figures, including deputy prime minister Angela Rayner, chancellor Rachel Reeves and her sister Ellie Reeves, the party chair, were dispatched days ago to try to bring the rebels into line,

One MP was told their next election campaign would be defunded unless they fell into line. Another had it relayed they would be deselected and their parliamentary future would be over. Others were warned they could be ousted as chairs or members of select committees. 

But even as the Labour leadership tried to turn the screw, the number of backbenchers joining the insurrection increased from 108 on Monday to more than 120 MPs. 

Even if Starmer’s compromises succeed, the revolt may leave him as a diminished leader. Observers are asking how the prime minister was blindsided by the biggest revolt of his premiership, involving scores of MPs who had hitherto been seen as loyal lobby fodder.

The seeds were sown in March when Liz Kendall, welfare secretary, set out plans to save nearly £5bn from the welfare budget. The biggest reform would be narrowing eligibility for Pip, the main disability benefit.

Kendall has tried to sell the reforms as an attempt to give back the dignity of work to some of Britain’s 2.8mn people who are currently inactive due to long-term sickness. 

But many MPs have been inundated by disabled constituents worried about what the cuts will mean in practice. The proposals would have removed benefits from at least 800,000 people, many of whom need help using the toilet or washing themselves. 

“The rebels are decent people, they are not headbangers, most of them, they are people who are deeply worried,” said one former Labour frontbencher. 

Starmer’s closest aides — “the Starmtroopers” — had congratulated themselves that they had weeded out left-wing radicals from Labour’s candidates for last year’s general election. 

Yet many of the new intake have a background in the charity sector or health, while others have friends or family with disabilities. 

“The fact that it’s centrists who are revolting should tell the government something,” one Labour MP said. “It’s far more dangerous than if it was just the usual suspects on the left.”

Those teetering on the edge of dissent were emboldened last month when Starmer announced a £1bn U-turn on the government’s highly unpopular cuts to winter fuel payments for pensioners. 

But ministers feared that giving in to pressure would stoke further rebellions, for example over the two-child benefit cap. “The new intake of backbenchers are starting to find their feet — and they’ve realised they have a lot of power,” one MP said.

The risk of Starmer’s compromises is that it could make it even more difficult for his government to get through future reforms or cuts — at a time when the UK’s debt levels already putting gilt markets on edge.

Government U-turns are putting increasing pressure on chancellor Rachel Reeves’ fiscal rules and increasing the odds of tax rises this autumn, economists have warned. Ruth Gregory at Capital Economics said the chancellor may need to raise between £10bn and £20bn.

Some say Reeves, with her determination to balance the books, is responsible for the crisis, while others point the finger at Starmer’s chief of staff Morgan McSweeney. Others say the culprit is Number 10’s influential policy director Liz Lloyd.

Still others blame the prime minister for an initial stubbornness that allowed the situation to spiral.

One MP said on Thursday before the last-ditch talks: “I’ve never seen so much bad blood. Both sides are raging.” It remains to be seen how quickly that anger will dissipate.

Lucas Beyer posted on X Thursday that he, Alexander Kolesnikov, and Xiaohua Zhai would depart OpenAI to join the $1.8 trillion company led by Mark Zuckerberg. Beyer said it was “fake news” that Zuckerberg was paying him that level of compensation. However, that news came from OpenAI CEO Sam Altman himself, who called the offers “crazy” this month. 

“They started making these like, giant offers, to a lot of people on our team—$100 million signing bonuses, more than that comp per year,” Altman told his brother Jack Altman in an episode of the podcast Uncapped. “I’m really happy that, at least so far, none of our best people have decided to take him up on that.”

hey all, couple quick notes:
1) yes, we will be joining Meta.
2) no, we did not get 100M sign-on, that’s fake news.

Excited about what’s ahead though, will share more in due time!

cc @__kolesnikov__ and @XiaohuaZhai.— Lucas Beyer (bl16) (@giffmana) June 26, 2025

Beyer, Kolesnikov, and Zhai have been members of the technical staff at OpenAI since December 2024, which they joined after being poached from rival Google DeepMind. They depart for Meta at a time when competition for talent among AI firms is reaching a frenzied pitch, with Zuckerberg reportedly on a recruitment spree to counter the narrative that it is lagging behind in AI development. Reports claim the company is hiring a 50-person “Superintelligence” team to ramp up its AI efforts. Meta has also purchased a $14 billion stake in Scale AI, to bring CEO Alexandr Wang into the fold. 

Zuckerberg famously earns only $1 as CEO at Meta, although the company provides him a $14 million allowance for costs related to security for Zuckerberg and his family. He holds about 13% of the tech behemoth’s stock and his fortune is valued at $250 billion by Forbes. 

Among the top-paid executives at Meta, chief operating officer Javier Olivan was paid the most last year, with compensation valued at $25.5 million. No other top executive at Meta was paid $100 million in any of the past three years, according to the company’s financial filings. 

The median of the total annual compensation of all Meta employees other than Zuckerberg was $417,400 last year. 

On X, a commenter speculated that Altman “clearly just threw out the 100m figure out there to make potential takers think that they were being lowballed.”

“Yes, it was a brilliant move, gotta give him that,” posted Beyer in response.

OpenAI and Meta did not immediately respond to requests for comment. 

For the second time this week, a US federal judge has issued an opinion on whether or not using copyrighted materials without permission to train AI amounts to “fair use” – and the most recent ruling contradicts the previous one.

In an order on Monday (June 23), Judge William Alsup handed a partial victory to AI company Anthropic in its defense against a lawsuit by three authors, declaring that training AI on copyrighted materials does indeed count as fair use.

Two days later, another judge in the same court – the US District Court for the Northern District of California – declared the exact opposite.

“This case presents the question whether such conduct is illegal,” Judge Vince Chhabria wrote. “Although the devil is in the details, in most cases the answer will likely be yes.”

This latest ruling is in a class-action case brought in 2023 against Meta – owner of Facebook and Instagram and developer of the Llama large language model – by 13 writers, including comedian Sarah Silverman, who wrote the book The Bedwetter. Other authors involved in the suit include Richard Kadrey, Junot Diaz, and Laura Lippman.

They argued that Llama had been trained on their works without permission, and would even reproduce parts of those works when prompted.

Despite his conclusion that training AI on copyrighted works without permission isn’t fair use in most cases, Judge Chhabria ruled in Meta’s favor – but only because, in his view, the authors’ lawyers had argued the case badly.

The authors “contend that Llama is capable of reproducing small snippets of text from their books. And they contend that Meta, by using their works for training without permission, has diminished the authors’ ability to license their works for the purpose of training large language models,” the judge noted. He called both arguments “clear losers.”

“Llama is not capable of generating enough text from the plaintiffs’ books to matter, and the plaintiffs are not entitled to the market for licensing their works as AI training data,” the judge wrote in his order, which can be read in full here.

The judge granted Meta’s request for a partial summary judgment in the case.

But what may be of greatest interest to rightsholders is that Judge Chhabria offered what he says would be a winning argument: That allowing tech companies to train AI on copyrighted works would severely harm the market for human-created works.

“The doctrine of ‘fair use,’ which provides a defense to certain claims of copyright infringement, typically doesn’t apply to copying that will significantly diminish the ability of copyright holders to make money from their works (thus significantly diminishing the incentive to create in the future),” Judge Chhabria wrote.

“What copyright law cares about, above all else, is preserving the incentive for human beings to create artistic and scientific works… By training generative AI models with copyrighted works, companies are creating something that often will dramatically undermine the market for those works, and thus dramatically undermine the incentive for human beings to create things the old-fashioned way.”

“By training generative AI models with copyrighted works, companies are creating something that often will dramatically undermine the market for those works…”

US District Judge Vince Chhabria

Copyright owners likely won’t be happy to hear the judge’s assertion that they don’t have a right to a market for licensing works to AI companies, but they’re likely to rejoice over much of the rest of the judge’s argument – including a remarkable passage where he directly criticizes the earlier ruling by Judge Alsup, who sits on the same court.

“Judge Alsup focused heavily on the transformative nature of generative AI while brushing aside concerns about the harm it can inflict on the market for the works it gets trained on,” Judge Chhabria wrote.

“Such harm would be no different, he reasoned, than the harm caused by using the works for ‘training schoolchildren to write well,’ which could ‘result in an explosion of competing works’…

“But when it comes to market effects, using books to teach children to write is not remotely like using books to create a product that a single individual could employ to generate countless competing works with a miniscule fraction of the time and creativity it would otherwise take. This inapt analogy is not a basis for blowing off the most important factor in the fair use analysis.”

“If using copyrighted works to train the models is as necessary as the companies say, they will figure out a way to compensate copyright holders for it.”

US District Judge Vince Chhabria

The judge also demolished an argument often made by AI companies: That forcing them to license all the materials they use for training would slow down or even stop development of the technology.

“The suggestion that adverse copyright rulings would stop this technology in its tracks is ridiculous,” Judge Chhabria wrote.

“These products are expected to generate billions, even trillions, of dollars for the companies that are developing them. If using copyrighted works to train the models is as necessary as the companies say, they will figure out a way to compensate copyright holders for it.”

Using pirated works not OK, judges agree

The disagreement between the two judges notwithstanding, there is one thing they both agreed on: Using pirated copies of works to train AI is not acceptable.

In the case against Anthropic, Judge Alsup ordered the AI company to answer for its use of materials taken from online libraries known to offer pirated books. That part of the case will be heard in December, and Anthropic could find itself on the hook for up to $150,000 per infringed work.

Similarly, Judge Chhabria allowed one key part of the authors’ case to go forward: The part dealing with Meta’s alleged use of the torrent file-sharing network to download illegal copies of books, and its stripping out of rights management information from the books it obtained, in violation of the Digital Millennium Copyright Act (DMCA).

All in all, the two rulings present an unusual instance of one court offering two very different opinions on the same question – a matter likely to be resolved, sooner or later, by an appeals court.Music Business Worldwide