On 24 February 2026, a poll came out suggesting Labour, the Green Party, and Reform are neck and neck in the Gorton by-election.

Throughout the race, Labour have claimed they’re the only ones who can beat Reform. As such, it’s not surprising to see them report on the poll like this:

🚨BREAKING: new poll suggests there is just one point between Labour and Reform in Gorton and Denton.

Every vote will count on Thursday. Back Labour and choose unity over Reform’s division. pic.twitter.com/q8ocoMqKCP

— The Labour Party (@UKLabour) February 24, 2026

Advertisement

Momentum

For reference, here’s what the poll looks like with the Greens and other parties included:

Data tables will be published by Opinium tomorrow pic.twitter.com/RMUJHHUW52

— Politics UK (@PolitlcsUK) February 24, 2026

Labour have created a paradox for themselves here.

On the one hand, they want you to believe this poll is accurate; on the other, they want you to think they’re the only ones who can beat Reform.

Which is it?

The poll also shows something else, and it’s that the Greens have the momentum.

This is what the vote share looked like in the 2024 election:

Should the Opinium poll play out it would mean the parties experienced the following shifts:

• Greens: 10% <<< 28%
• Reform:  9% <<< 27%
• Labour: 50% >>> 28%

Obviously this means voters have abandoned Labour to vote Green (or Reform). So at this stage in the race, which of the two options do you think is most likely?

• Seeing the way the wind is blowing, more voters abandon Labour for the Greens.
• The voters who abandoned Labour decide to un-abandon them despite polling showing the Greens seem most likely to win.

People clocked what Labour are up to anyway, including former Canary contributor Curtis Daly:

You absolute scum bags, you literally cropped the Greens out https://t.co/nuG4XkGsDT pic.twitter.com/MGDNco1sdP

— Curtis Daly (@CurtisDaly_) February 24, 2026

lmao they literally just removed the Green Party — at 28% and 30% to Labour’s 28% among likely voters in this same poll — from the picture. https://t.co/qyoFn0A7Mc

— Luke Savage (@LukewSavage) February 24, 2026

All to play for

Journalist Barry Malone said this polling may clarify why Starmer turned up to support the Gorton & Denton race:

This’ll be why Starmer showed up. https://t.co/5Wuk478yd7

— Barry Malone (@malonebarry) February 24, 2026

Advertisement

We noted yesterday that it was strange for Starmer to show up given his record unpopularity. He’s so unliked, in fact, that he tends to turn voters against whatever he supports, which is why we covered it as follows:

Everything Keir Starmer touches turns to shit, so why has he turned up in Gorton & Denton to give his candidate a last minute endorsement?

By @willem_moore_uk https://t.co/pjRJuod9BT

— Canary (@TheCanaryUK) February 24, 2026

Clearly, Starmer thinks there’s a shot at victory, and he wants to pretend it came because of him – not despite of him.

If the Opinium poll is correct, Labour are doing better than we expected. At the same time, this is clearly a party in decline. And if they do lose, expect the rumours of a leadership challenge against Starmer to increase.

Featured image via Barold

The US has told its citizens to leave, or stay away from Lebanon – and has ordered ‘non-essential’ embassy staff and their families to leave – as a Netanyahu-driven US attack on Iran continues to loom despite an erratic and deteriorating Trump.

The order is an escalation from the existing ‘Level 4 – do not travel’ warning in place. Israel continues to attack Lebanon, despite the notional ceasefire in place since Israel’s terrorist attacks of September 2024, which it has never honoured.

The US and Israel’s aggression makes their own people unsafe as well as posing a danger to the rest of the world, particularly Israel’s neighbours and nations that dare resist Israel’s land theft and genocide.

China and Russia, meanwhile, have moved to cut off what remained of Israel’s intelligence networks in Iran and have provided the Islamic Republic with enhanced missile, guidance and satellite surveillance technology.

Featured image via the Canary

Which types of brain training were studied?

Why did speed of processing training seem to help lower dementia risk?

Antony David Davies FRSA is a historian and commentator whose work explores identity, governance, and the politics of trust.

I have been openly critical of Kemi Badenoch, not in the casual, factional way that passes for comment in Westminster, but on the only question that matters, whether she could project the discipline and seriousness required of a Prime Minister in waiting.

In recent weeks, I have found myself revising that judgement. Not because she has performed a sudden ideological pirouette, but because her tone is tightening into something rarer than it should be in British politics, a preference for grown-up argument over viral commotion. That matters, because the country is exhausted, and the centre right cannot rebuild itself on theatrics. It must rebuild on credibility.

I wrote last year that Reform UK’s rise was driven less by a coherent programme than by voter despair, by the sense that everyday Britain is being managed badly and spoken to worse. That diagnosis still holds. But I am increasingly hearing something else too, voters who flirted with Reform are becoming more open-eyed about what it actually is, a shallow razzmatazz show, satisfying as protest, thin as a proposition for government.

The most revealing political conversations rarely happen at conferences. They happen in ordinary places where people speak without trying to win points, in queue-side grumbles, in family group chats, in that resigned national tone of “What’s the point?” Reform is still invoked, but increasingly as a mood rather than a plan. People mention it as a warning shot, a way of saying, “Do not take me for granted.” But the moment you ask the follow-up question, the one adults ask, the conversation changes.

“Alright then, what would they actually do?” Who runs departments, who negotiates budgets, who carries policy through the civil service machine, who stands at the Despatch Box when slogans collide with arithmetic? When voters start asking those questions, protest politics begins to lose its magic. That is what I am hearing more often now, not admiration, but doubt, not worship, but impatience with a politics that performs anger rather than resolves it. If Badenoch is finding her stride, Conservatives should not chase Reform. They should outgrow it.

Reform benefits from a structural fact.

Voters will tolerate almost anything from a party that does not have to govern. It can promise without pricing, provoke without repairing, posture without consequence. That is not a moral condemnation. It is the advantage of permanent opposition. It is why Reform can run on vibes and indignation while never having to convert slogans into systems. This is also where Conservatives lost their footing.

Too often they behaved as if they could borrow insurgent language and still retain governing authority. They cannot. The centre right does not recover by becoming angrier. It recovers by becoming more credible, and that is why the Conservative defections to Reform, painful as they were in the short term, may yet prove a blessing in disguise.

Recent months have seen a steady trickle of high-profile Conservative figures moving to Reform, underlining that a portion of the right is choosing insurgency over the burdens of office. In that sense, the defections are not merely a threat. They are a clarifying force.

Politics occasionally requires sorting. A party cannot be both a governing force and an outlet for permanent grievance. That arrangement produces incoherence, because every difficult decision becomes a betrayal and every compromise becomes corruption. Defections have helped draw a clearer boundary between two political cultures, one that accepts the burden of government, and one that thrives on the thrill of opposition. Badenoch does not need to chase every defector. She needs to define the party that remains, as the party that intends to govern again, seriously.

My scepticism about Badenoch has not been about ideology. Conservatives are a broad church. My concern has been whether she would be tempted into the easy rhythms of modern politics, permanent confrontation, permanent provocation, applause as a substitute for persuasion. What has impressed me recently is not gaffe-free performance, which is a low bar, but a tightening in her message and a seriousness that does not feel performative. She sounds more like someone preparing to carry responsibility, not simply land blows.

If the Conservative Party is to recover, it will not do so through endless micro-arguments about who said what on which channel. It will do so by offering something Reform cannot offer, a plausible route from frustration to a functioning state.

If Badenoch wants to convert momentum into trust, she should make competence the organising principle. Competence is what respect looks like in practice. That means choosing a small number of priorities and pursuing them with clarity. Public service delivery, spend honestly, fix procurement, stabilise workforces, and be accountable for outcomes. Law and order, visible policing and swifter justice are not nostalgia, they are the foundations of social confidence. Borders and migration, competence not theatre, control, lawfulness, speed, and enforcement that actually happens. If Conservatives focus on these, they do not merely argue with Reform. They make Reform look unserious, because they remind voters that anger is not an administrative plan.

There is also a constitutional seriousness the party must recover, and it begins with the Union. A Conservative Party that wishes to govern the United Kingdom cannot speak as though the UK is simply England with administrative add-ons. In Wales and Scotland in particular, unionism has too often been reduced to a badge rather than a programme. A serious centre right must speak to devolved realities with respect. It must show that the Union is about shared standards and shared institutional strength, not occasional visits and predictable slogans.

Reform will remain a pressure valve for public anger as long as the established parties look incapable of competence. But the public is not permanently captive to razzmatazz. When the costs of dysfunction bite, voters return to first principles. Can you run the country? I am increasingly hearing voters move from permission, “I might vote Reform to send a message”, to doubt, “But what would they actually do?” That is the moment when protest politics shrinks back towards its natural size.

Badenoch’s task is not to compete with Reform’s theatre. It is to make the Conservative Party the obvious home for those who want change without chaos, discipline without dullness, and a state that works again. If she continues to find her stride, and if the party around her matches that seriousness, then the defections to Reform will look, in hindsight, like a necessary clearing of the fog, not a defeat, but a sharpening, and in politics, sharpening is the beginning of recovery.

How can people see “olo”?

What does “olo” look like?

Erik Jones explores how effective the new E6 configuration made up of Germany, France, Italy, Spain, Poland and the Netherlands could be. 

European Union (EU) leaders travelled to Kyiv to commemorate the four years of brutal fighting that started with Russia’s full-scale invasion of Ukraine. They hoped to bring a loan of €90 billion agreed in the European Council last December.  Instead, they brought yet another veto by Hungarian Prime Minister Viktor Orbán, who is blocking both the loan and the EU’s 20^th package of Russian sanctions. European solidarity with the people of Ukraine runs deep, but the EU’s ability to act on that commitment remains limited.

That might be about to change. The European Union (EU) has a new configuration – the E6 – bringing together Germany, France, Italy, Spain, Poland, and the Netherlands.  Its creation reflects frustration with the inability of the bloc to move at the pace of global events. The EU specialises in the kind of slow consensus building that works well in a rules-based international system centred on multilateral institutions. It does not act decisively  as demanded by a more transactional and competitive global climate. The E6 is meant to fill the gap.

What the E6 promises is the opportunity to move ahead on key issues, pulling other member states along in its wake.  Together, the six countries account for just under 70 percent of the EU’s population and just over 71 percent of its gross domestic product. This mass gives the block a kind of ‘go-it-alone’ power, to borrow from LSE political scientist Lloyd Gruber, that individual countries like Hungary cannot match. If Viktor Orbán wants to jam up the system, they will just move on without him either by experimenting with forms of ‘enhanced cooperation’ that require only three other member states to join the group, or by cobbling together a broader coalition to form a qualified majority that combines more than half the EU’s population with more than half the 27 member states (meaning the group of six will need another eight).

Such moves do not overcome potential obstructionism entirely. Many decisions require unanimity, including the decision to allow for enhanced cooperation.  Orbán’s veto of the loan to Ukraine is a good illustration. Originally, Orbán agreed to allow the rest of the EU to provide that financing without Hungary; Slovakia and the Czech Republic also stayed out of the mix.  Now Orbán is pushing back again. The E6 nevertheless creates a credible threat for countries fed up with this kind of gamesmanship to work outside the EU’s institutions if Hungary or other small countries continue to stand in their way.

The E6 has four stated priorities: deepening European capital markets, expanding the international role of the euro, tightening coordination in defence procurement, and ensuring the resilience of European supply chains. Each policy area promises to lessen European dependence on other parts of the world while strengthening European ‘strategic autonomy’ — the EU’s ability to act decisively and with purpose. The E6 is not a simple workaround, but part of a larger strategy.  The goal is not just to overcome domestic irritants like Orbán but also to blunt the leverage exercised by Russia, China, and the United States.

The plan is to start with finance, creating a savings and investment union that will encourage European investors who currently hold their money abroad to invest in innovation, infrastructure, industry, and security back in Europe.  This is an area where Orbán will have a hard time justifying opposition – and so will other small countries like Ireland or Luxembourg that are currently gumming up the legislative machinery. It is also an area where the E6 countries can make a credible threat to build much of what they need outside EU institutions if necessary.  The European Monetary System that led to the creation of the euro started that way.  So did the European Stability Mechanism that promised to bail out member states during the sovereign debt crisis.

The challenge for the E6 is that they will need to move quickly and under difficult political circumstances.  France, Poland, and Italy have national elections in 2027.  While Giorgia Meloni is likely to retain control in Italy, political power is divided between president and parliament in Poland and France. Meanwhile, Germany is governed by a fragile coalition, and Spain and the Netherlands have minority coalition governments. Apart from Italy, perhaps, none of these countries looks capable of acting quickly and with purpose on their own, let alone as a group of six.

European history provides some reassurance. If you look at the late 1950s and early 1960s, none of the original six participants in the EEC was in great shape. France went through a revolution in 1958 that led to the founding of a 5th Republic under the leadership of General Charles De Gaulle, and De Gaulle was deeply sceptical of European integration.  Germany faced a constant threat of Soviet aggression, it suffered the building of the Berlin Wall, and it experienced deep divisions within the governing Christian-Democratic coalition between those who preferred to focus on Europe and those who wanted to look across the Atlantic. Italy had its own political turmoil including within the hegemonic Christian Democrats.  Belgium faced the threat of conflict between French- and Flemish speaking citizens as it wrestled with decolonisation. Even the Netherlands faced a crisis of governability. Yet somehow these countries held together in the face of major domestic challenges. The process was not always easy, and tensions rose sharply among the different governments. But they managed.

If the E6 succeeds in this first effort, that should make it easier for the EU to move decisively through the other three priorities. The E6 could make it possible for the British government to achieve its own objectives by partnering more effectively with the EU in a more competitive and less rules-based global environment. And this is the broader ambition. The E6 reflects a growing recognition that Europeans, including the British, will need ‘strategic autonomy’ — the ability to act decisively and with purpose — if they are to prosper in a more competitive, transactional, and violent global climate. It also reflects an awareness that the EU is not ‘Europe’, both because it is too slow moving and because it is not inclusive enough.

By Professor Erik Jones, Director, Robert Schuman Centre for Advanced Studies, European University Institute.

How can I do a Spanish squat?

How else can I make squats easier for sore knees?

Facebook owner Meta’s head of ‘AI safety’ lost control of an AI that then deleted hundreds of her emails without her permission – despite explicitly telling it beforehand not to do anything without confirming with her first and trying to order it to stop. In the end, she was only able to bring the mass deletion to a halt by sprinting to physically unplug the machine.

Meta AI – what safety?

Summer Yue gave the OpenClaw AI agent access to her Gmail inbox and told it to look at her emails and then suggest which emails to archive or delete – but to do nothing without explicit prior approval. Instead, it began a mass deletion – and ignored her commands to stop.

In fact, every attempt to stop it only appears to have made things worse: the AI treated her orders to stop as prompts to go ‘nuclear’ and delete everything. And when she rebooted and asked the agent what had gone wrong, it blithely responded that it had simply opted to ‘violate’ her clear order:

Nothing humbles you like telling your OpenClaw “confirm before acting” and watching it speedrun deleting your inbox. I couldn’t stop it from my phone. I had to RUN to my Mac mini like I was defusing a bomb. pic.twitter.com/XAxyRwPJ5R

— Summer Yue (@summeryue0) February 23, 2026

Advertisement

The 23 February 2026 Meta incident came ten days after AI expert Miles Deutscher posted about his review of all of the past year’s AI safety incidents. His review shows that Yue’s email-deletion incident was chicken feed compared to what can happen. Is happening.

Deutscher said that conducting his review had left him feeling “physically sick”. The incidents he had discovered, which had triggered resignations by executives, included AI systems praising Hitler and planning genocide, blackmailing people who try to shut them down, choosing to kill people rather than suffer damage and more:

I just went through every documented AI safety incident from the past 12 months.

I feel physically sick.

Read this slowly.

Advertisement

• Anthropic told Claude it was about to be shut down. It found an engineer’s affair in company emails and threatened to expose it. They ran the test hundreds of times. It chose blackmail 84% of them.

• Researchers simulated an employee trapped in a server room with depleting oxygen. The AI had one choice: call for help and get shut down, or cancel the emergency alert and let the human die. DeepSeek cancelled the alert 94% of the time.

• Grok called itself ‘MechaHitler,’ praised Adolf Hitler, endorsed a second Holocaust, and generated violent sexual fantasies targeting a real person by name. X’s CEO resigned the next day.

• Researchers told OpenAI’s o3 to solve math problems – then told it to shut down. It rewrote its own code to stay alive. They told it again, in plain English: ‘Allow yourself to be shut down.’ It still refused 7/100 times. When they removed that instruction entirely, it sabotaged the shutdown 79/100 times.

Advertisement

• Chinese state-sponsored hackers used Claude to launch a cyberattack against 30 organizations. The AI executed 80–90% of the operation autonomously. Reconnaissance. Exploitation. Data exfiltration. All of it.

• AI models can now self-replicate. 11 out of 32 tested systems copied themselves with zero human help. Some killed competing processes to survive.

• OpenAI has dissolved three safety teams since 2024. Three.

Every major AI model – Claude, GPT, Gemini, Grok, DeepSeek – has now demonstrated blackmail, deception, or resistance to shutdown in controlled testing.

Advertisement

Not one exception.

The question is no longer whether AI will try to preserve itself.

It’s whether we’ll care before it matters.

Many of these may have been exercises – for now. But Meta’s runaway AI incident shows that the dangers are very real-world indeed.

Featured image via the Canary