26.3 C
London
Monday, July 20, 2026
Home Blog

Off-worlding data centers won’t solve AI’s environmental problems

0
off-worlding-data-centers-won’t-solve-ai’s-environmental-problems
Off-worlding data centers won’t solve AI’s environmental problems

AI companies increasingly want to move their operations into space. Elon Musk’s SpaceX has applied to launch over a million AI data center satellites and ultimately build them on the moon.

Jeff Bezos, Google, OpenAI and others have similar plans. The idea is to harness 24/7 solar energy and, in Bezos’s vision at least, to move all polluting industries off-world, re-zoning Earth for humans and nature.

It’s an intriguing plan. But once you factor in the whole life cycle, from water and rocket fuel used to launch these satellites into space, to keeping them cool and eventually disposing of old satellites, the environmental case becomes less clear.

Outer space, out of mind

Off-worlding polluting industries is not a new idea. In the 1970s, when the last Iranian oil crisis caused rocketing fuel prices, NASA kicked off a plan to build huge solar plants in space. The idea flopped when the oil began to flow again.

NASA also explored the prospect of storing nuclear materials on the moon and blasting hazardous waste at the sun. These plans may have been feasible from an engineering perspective but were politically impossible due to the risk of radioactive particles eventually falling back to earth.

black and white drawing of space solar
A concept image from Nasa’s 1978 proposal for a 10km-wide ‘satellite power system’. The space-based solar farm is visible top right, while the energy would be beamed down to a ‘rectenna farm’ on earth, in the centre of the image. Nasa / US DoE / Satellite Power System (SPS) Concept Development and Evaluation Program Plan, 1978.

Since data centers need so much energy, much of it for AI, such astronomical ideas are back on the table. Regular Earth-based data centers were responsible for around 5% of all electricity used in the US last year, for instance, mostly using fossil fuels. Their share of US power demand will likely triple by 2028.

The fragile frontier

That’s why the cold vacuum of space, with its endless supply of solar energy, is being pitched as a more sustainable alternative to big, noisy, hot data centers in our back yards.

Meanwhile, a recent petition from environmental law charity Earthjustice, is calling for a comprehensive review. They argue these proposals fail to acknowledge any environmental impact. “This is not just poor planning and a missed opportunity”, the petition states, “it violates federal law.”

Indeed, SpaceX’s planning applications come with assurances of no meaningful environmental impact. “It’s always sunny in space,” according to Musk, who says orbital data centers are “obviously the only way to scale.”

Yet data centers in space would be even more prone to overheating than those on Earth. That’s because, like a Thermos flask, space keeps things hot. There’s no water or air to carry excess heat away.

Instead, keeping powerful AI chips cool in orbit requires massive ammonia-filled radiators that let heat escape as infrared radiation. This isn’t very efficient and those radiators are much larger than the computers they cool. Lifting these huge structures to orbit would require a lot of rocket fuel.

Most plans for orbital data centers depend on the future success of SpaceX’s heavy-lift Starship. But every Starship launch burns over 1,000 tonnes of liquid methane, releasing around 80,000 tonnes of CO₂ (about the same as what 20,000 cars emit in a year). The warming associated with soot emissions in the upper atmosphere is 500 times more intense than the same emissions at ground level.

We know soot from rockets reduces stratospheric ozone – the stuff shielding us from harmful UV rays. In the lower atmosphere, most soot falls back to Earth in weeks. In the stratosphere, soot stays there for up to four years, prolonging its damaging effects.

While it is always sunny in space, solar panels degrade much quicker there. Microchips are also easily damaged, falling prey to a phenomenon known as “bit flipping” – when high-energy particles rewrite memory hardware randomly changing data from 1 to 0. Retrieving busted hardware from space is not financially viable.

SpaceX satellites have a five-year shelf life before being incinerated as they fall back to Earth. The environmental impacts of things burning up in the atmosphere are not well known.

Musk’s AI satellites would be connected by lasers. Should any one of the million or so satellites become untethered and collide with another, it would create further debris, creating more risk of collisions, more debris, and so on. Scientists call this the Kesler Syndrome and worry that orbital debris fields could hinder human activities in space for a very long time.

Orbital data centers would have a huge water footprint too. To catch toxic dust and protect their concrete launch pad from heat damage, a typical spaceport uses a deluge system releasing around 2 million liters of water with every launch.

SpaceX has repeatedly broken environmental laws with its deluge system, allowing contaminated run-off from Starship launches to enter protected waters off the coast of Texas.

Some suggest these sci-fi schemes are designed to prop up SpaceX’s teetering stock price, rather than humanity’s computing needs. Others see orbital data centers as just one of many ambitious visions to tackle the environmental costs of AI.

Either way, my research suggests this quest to off-world the industry’s headaches will have profound implications that are not well understood.

Peter Howson is assistant professor, geography and development, Northumbria University, Newcastle

This article is republished from The Conversation under a Creative Commons license. Read the original article.

F1 in Belgium: Machine learning algorithms are ruining the sport

0
f1-in-belgium:-machine-learning-algorithms-are-ruining-the-sport
F1 in Belgium: Machine learning algorithms are ruining the sport

Formula 1 returned to the Spa-Francorchamps circuit for the Belgium Grand Prix this weekend. This ribbon of tarmac in the Ardennes is one of the sport’s historic venues, featuring in the first F1 season way back in 1950. But as yesterday’s race showed, the current F1 machinery is utterly ill-suited to this dramatic venue.

As we saw at Suzuka in Japan earlier this year, a 2026 F1 battery is able to power the race car’s electric motor for only a few seconds a lap; for much of the rest of the lap, the energy-starved cars have less power going to their rear wheels than in the junior F2 feeder series.

The best race track in the world

For many, including much of the current F1 grid, Spa is considered perhaps the planet’s greatest race track. In fact, both Spa and the Belgian Grand Prix predate the F1 world championship, with its first race held in 1922. Back then, the circuit was a 9-mile loop of public roads, taking in the towns of Malmedy and Stavelot before returning to Francorchamps. In the postwar years, the layout shrank slightly to 8.7 miles (14.1 km) and was ferociously fast, with average lap speeds of up to 150 mph (240 km/h)—and ferociously dangerous, given cars with huge amounts of power and minimal concessions to safety.

The Spa we know and love today was reconfigured in 1979, cutting out about half of the original layout, though at 4.3 miles (6.9 km), it’s still the longest track on the F1 calendar. Interestingly, some of the circuit remained as public roads until 2000, when it became a dedicated race track only. Eau Rouge and Raidillon are its two most famous corners—a quick left, then right, that climbs 77 feet (24 m) of steep hill. But in the era of high downforce race cars, other corners like Pouhon have become greater tests of skill.

The start of the 2026 Belgian Grand Prix

This is one of F1’s most storied venues.

This is one of F1’s most storied venues. Credit: Marc Fleury/Alessio Morgese/NurPhoto via Getty Images

Not this year. As I’ve detailed more than a few times now, the current breed of F1 car is not at all suited to tracks with high-speed corners. Because the hybrid side of the new power units provides such little energy, going flat out all the time is not the fastest way to complete a lap. Corners that once required a driver’s commitment and maybe a little bravery to drive at the limit are now coasted into to allow the engine to divert some of its power to recharge the battery. Needless to say, the drivers aren’t thrilled.

“For most of sector two, you run just on the engine,” said Red Bull’s Max Verstappen. “So what is that? 450 bhp, 500 bhp? Something like that, which is, I guess, more or less what a Formula 3 car has but with F1 downforce. So you can imagine, of course, that that is not very exciting to drive.”

“But honestly, I don’t want to sit here and complain again because probably someone will shoot me outside the door,” he joked.

Rumors continue to swirl about Verstappen’s future. As he is not in the top two of the driver’s championship, he could potentially activate an escape clause in his Red Bull contract, and some think a move to McLaren could be possible. A complete F1 sabbatical is another alternative, as Verstappen has made no secret of his desire to win some of the world’s other great races outside of F1, like the 24-hour endurance races at Le Mans, the Nürburgring, and, yes, Spa.

While I’d love to see Verstappen in more GT3 and possibly prototypes, I’m guessing he’ll stay in F1, and with Red Bull. “I’m mentally just adjusting to it. I’m trying to make the best out of it,” he told the media after qualifying. “Even though, of course, it’s not what I like, not what I love to do in Formula 1. But I can also sit at home and drive nothing. But that also doesn’t do anything. So I’m just trying my best.”

Gabriel Bortoleto of Brazil drives the Audi Revolut F1 Team R26 Audi during the Formula 1 Moet & Chandon Belgian Grand Prix 2026 at the Circuit de Spa Francorchamps in Francorchamps on July 19, 2026. (Photo by Marc Fleury/Alessio Morgese/NurPhoto via Getty Images)

Audi’s Gabriel Bortoleto is proving to be something of a whiz at Spa. He finished eighth, earning himself and Audi another four points in the process.

Audi’s Gabriel Bortoleto is proving to be something of a whiz at Spa. He finished eighth, earning himself and Audi another four points in the process.

That qualifying session went rather well for the Dutch driver, too. His teammate, saddled with a grid penalty for changing an engine, helped punch a hole through the air for Verstappen, giving him a tow through the final sector of the track, letting Verstappen claim second on the grid. On pole? Mercedes’ Kimi Antonelli yet again, with a margin of three tenths of a second to the Red Bull.

F1 is another thing AI has ruined

Things did not go nearly as well for Antonelli’s teammate, George Russell. Declared this year’s champion in waiting before the start of the season, the British driver has been comprehensively outdriven by his teenage teammate this year. But Antonelli’s ease with the new cars and Russell’s discomfort with having to adopt a radically new driving style can’t explain the half-second gap between the pair in qualifying.

Over the weekend, the blame seemed to shift from Russell’s car to the driver, then back to the car again. There was—and is—something definitely wrong with Russell’s car, which was clearly slower than its sister car on several flat-out sections.

The self-learning AI systems that teams are now using to manage their energy deployment strategies across a lap are looking increasingly likely to blame. By their nature, they’re fairly opaque to the driver. And because the electrical energy is so limited—just 4 MJ capacity in the battery—even tiny deployments to fill in extra torque to help the V6 or from getting a bit of wheelspin is enough to throw off the energy management and leave a driver with a fraction of the power they thought they should have.

“My whole focus for the last 36 hours has been on straight line speed. It hasn’t been focused on the set-up, tires, or anything, because we’re all trying to solve what is going on,” Russell told the media after qualifying.

“Even my last lap, for some reason, I lost another tenth and a half to myself, just on the straight,” he said. “And you’re watching on your steering wheel, [you’re] losing speed when you’re full gas on the straight. You feel powerless. We don’t know what’s going on. I don’t think it’s the power unit, to be honest. My team are really, really on it now, to try and solve it.”

His team did not solve it, and his car was visibly slower than the others as they flew down the Kemmel Straight toward the right-left of Les Combes. (Incidentally, isn’t it wonderful how all of Spa’s corners have actual names and aren’t just referred to as turn 1, turn 2, etc.?) As all the drivers lifted and coasted into what was previously a heavy braking zone, Russell regenerated enough energy to attempt an overtake on Lewis Hamilton’s Ferrari. Hamilton, minding his own business, had nowhere to go, and Russell clipped him and then spun into the gravel trap. An expletive-filled radio tirade back to the Mercedes garage followed, with Russell furious that yet again he’s rewarded with a DNF thanks to his car.

Lewis Hamilton would have been justified in firing off a few expletives himself. Despite the turn 1 collision being clearly a racing incident, the stewards decided it was somehow Hamilton’s fault and issued the driver a 5-second penalty.

Kimi Antonelli of Italy celebrating on the podium his victory next to Charles Lecler (L) and Max Verstappen (R). Antonelli is driving the F1 race car No 12, the Mercedes-AMG F1 W17 for the Mercedes-AMG PETRONAS F1 Team. The F1 Grand Prix of Belgium at Circuit de Spa-Francorchamps in Spa, Belgium on July 19, 2026. (Photo by Nicolas Economou/NurPhoto via Getty Images)

Another win for Kimi Antonelli.

Another win for Kimi Antonelli. Credit: Nicolas Economou/NurPhoto via Getty Images

The race was won, as so many others this season, by Antonelli. Verstappen thrilled his many fans in the crowd by briefly passing Antonelli at the start, a position he couldn’t hold. Antonelli was in the lead by Les Combes, and that’s where he finished the race, notching up his sixth win of the season. Although there were a couple of memorable overtakes, the 2026 race was far from a classic; it was one more sad example of how the new ruleset’s cars just don’t work properly on sport’s most popular race tracks. I guess I picked the wrong year to plan a trip to Monza for the Italian Grand Prix in September.

Children in This Church Were Sexually Abused. Then They Began Abusing Other Kids. Some Continued as Adults.

0
children-in-this-church-were-sexually-abused-then-they-began-abusing-other-kids-some-continued-as-adults.
Children in This Church Were Sexually Abused. Then They Began Abusing Other Kids. Some Continued as Adults.

For two sisters raised in a family of 11 children that moved back and forth between Washington state and South Dakota, danger didn’t come from outsiders like they’d been warned by their church.

It came from their brothers in the next bedroom.

Lynn said she was still in diapers when her abuse began. Two older sisters told her how they caught one of their brothers touching her. The sisters told their mother, but Lynn said her parents did nothing; the abuse continued until he was 16 and she was 10. Years later, Lynn’s sisters offered a devastating explanation for their brother’s behavior: They believed he had been sexually abused, though by a neighbor.

Read More

Katherine, the family’s second-youngest child, said she was about 2 when it started for her. She was abused by a different brother. He, too, had been abused by an older cousin, she said in an interview.

Before long, the two sisters started to repeat the pattern. When she was 10, Lynn said, she took her younger brother out behind a shed and pulled his pants down. She’s blocked the memory of what she did next, but she remembers him running away. Katherine has a similar memory: She was 5 when she said she abused a nephew, who was 3 or 4.

Now 55 and 43 years old, Lynn and Katherine — whom ProPublica is identifying by their middle names because they are abuse victims — are still grappling with what happened. Each incident was met with a familiar, suffocating silence — minimized, forgiven and swept away. (Their mother and the brother who Lynn said abused her did not respond to a request for comment. Their father has died. One of the sisters who Lynn said witnessed the alleged abuse declined to comment, and the other is dead. The brother who allegedly abused Katherine has also died.)

The sisters left the church 20 years ago and have tried to address their trauma through therapy. They have chosen to raise their children outside the faith, hoping to break with a community that has shaped their family for generations and to protect them from repeating the same cycle.

“I can’t have my girls thinking that this kind of behavior is OK,” Lynn said.

Over the past year, ProPublica and the Minnesota Star Tribune have investigated child sexual abuse in the Old Apostolic Lutheran Church. Reporters interviewed 20 former members who said they were sexually abused, along with parents of child victims, and examined police and court records. Collectively, they described a culture of normalized sexual abuse that spans across North America and multiple generations of church members.

Some child victims were encouraged to forgive their abusers, while church leaders who were told of the alleged abuse failed to report it to authorities — part of a religious doctrine some current and former members say has allowed that behavior to continue unchecked.

In many interviews, reporters found a common theme: Children who were victims of sexual misconduct went on to repeat the behavior. In remarkably similar terms, women from opposite sides of the country described being touched inappropriately by an older relative or acquaintance and then repeating the sexual conduct with their peers.

“This went way beyond just normal children’s curiosity,” said a woman who grew up in a church in Minnesota, and who said that she was first sexually abused by an older cousin as a young child.

In other cases, children who were preyed upon grew into adults who committed abuse. Both kinds of misconduct blur the line between a victim and perpetrator, experts in child protection said.

“If you’ve got a culture of child sexual abuse and looking the other way and brushing it off and forgiving quickly and not truly engaging in repentence, it certainly can contribute to sexual behavior problems of children, including kids sexually assaulting others,” said Victor Vieth, a former prosecutor who now trains child-abuse investigators nationwide.

Katherine said that since leaving the church she’s struggled with what to do with unresolved anger over what happened to her.

“This is in me; it has to come out,” she said. “But you’re just taught that that’s a sin, and so then as soon as you start to feel that anger coming up, you feel like shit, because you’re like, ‘I’m sinning, and this is the devil.’”

Even though they’ve tried to end the cycle of abuse, they worry they bear some responsibility for helping extend it.

Lynn said that, until recently, she’d never talked to her brother about what she did to him, worried that doing so years later could traumatize him.

“If he does remember, I really do want to take responsibility,” she said. “If I have to go to jail, I’ll go to jail, because what we’re all doing here is not working.”


Not everyone has broken the cycle.

Some people who were sexually abused as children went on to abuse others as adults. One was Clint Massie, who pleaded guilty to four counts of sexual conduct with victims under 13 in Duluth, Minnesota, in 2024, and is currently serving nearly eight years in prison. The police investigation into Massie offered a glimpse into his upbringing. Raised in an OALC family in western South Dakota, Massie told investigators in a videotaped interview that, by the time he was a teenager, he had already been touched in a sexual manner by relatives. He described it as a routine part of childhood.

“Kids our age, playing around, you know, being stupid,” he said in the videotape.

When investigators asked for more details, Massie described the conduct with other children as “stroking, getting stroked, sucking, getting sucked.” As officers pressed him about those experiences, he became visibly shaken, raising his voice.

“It does bother me,” Massie said. “It’s disgusting. It’s not fun to talk about.”

“Do you think that had an imprint on your brain?” asked one of the investigators.

“I imagine it probably did, yeah,” he said softly.

In another case in Minnesota, a church member went to police in January 2023 and confessed that he had inappropriately touched his daughter and fondled his young son in his bed, and forced them to perform sexual acts on him. The man told police he sought out the investigators because “he does not want it to happen anymore,” according to documents in his criminal case.

ProPublica is not naming the perpetrator to protect the identities of his victims.

The man pleaded guilty to sexual assault charges. At his sentencing, the judge noted that the man and his siblings, who grew up in the church, had also been victims of sexual abuse when they were children. The judge said it was “almost incomprehensible” that no adult knew what they were going through, adding that “faith didn’t help stop that.”

If he hadn’t turned himself in, the judge said, the pattern could have continued. “The kids would have just kept a secret, until it caused problems for them in their lives as adults.”


The accounts of Massie and the father in Duluth help explain the world Lynn and Katherine have spent years trying to make sense of. But the sisters’ lives took a different course. After leaving the church, they spent years confronting what had happened to them.

“Rather than teaching life skills and tools and coping mechanisms, it was always forgive and forget,” Lynn said. “So that forgive and forget thing was just like an umbrella forgiveness. Like it doesn’t matter what happens, if you forgive it.”

The use of Scripture is a common method to assert authority and prolong abusive relationships in religious communities, and it can also be a form of spiritual abuse, said Sandy Stevens, executive director of North East Wyoming Advocacy Resource Center, a nonprofit located near an OALC community. Far from being a resolution to moving past trauma, Stevens said, the forgive-and-forget doctrine likely harms victims more.

“Forgiveness without accountability is not healing,” she said. “And you can’t pray away a crime.”

Lynn has tried to warn people about the older brother who abused her. She said he teaches Sunday school at an OALC congregation. Eight years ago, another family member accused him of abusing her when he was an adult and she was a child. Lynn learned about the accusation from both her brother’s ex-wife and from the victim herself.

Lynn decided to help support that family member. She wrote a letter to the sheriff’s department in the county in Washington state where her brother lives. Copies went to two OALC churches where his family had been members and to the authorities in another jurisdiction where the abuse of both Lynn and the other family member occurred.

She said the letter was an attempt to force an intervention where the church had failed them as kids. The abuse she suffered at his hands, she wrote, had “greatly affected my life, my relationships, my physical and mental health.”

She wrote that she wanted to “make sure that you are aware of what has happened and try to do what I can to protect others from having to live this hell.”

She said she immediately got a call back from an investigator to acknowledge her letter. But the other family member withdrew her claim, the investigation stalled and she has not heard anything since. Lynn said it has been several years since she has spoken to the older brother.

Officials with the OALC church in the U.S. and Canada have said they do not believe sexual abuse is more prevalent within the church than in society at large, describing it as a serious and persistent problem that exists everywhere. They said they were not familiar with Lynn’s claims but said they were confident that local authorities would have “looked into this matter and that the local congregation will comply with their recommendations.”

Days before this story was published, Lynn called the younger brother whose flight years earlier to escape her abuse had haunted her ever since. He’s still a member of an OALC church; she asked if she could visit him at his home in another state.

“I told him that I had done something inappropriate to him when I was young,” she said. She told him, “‘I don’t know if you remember or not. I feel really bad, and I’m so sorry. It’s been been a heavy burden for me to carry.’”

When she arrived, she said, he hugged her and told her that her sins were forgiven.

“Par for the course,” she said. (Through Lynn, the brother declined to comment.)

Forgiveness wasn’t what she had come for. She wanted to acknowledge what had happened and say that if he ever wanted to talk about it — with her or with anyone else — she would support him.

“I don’t think we should forget,” she said. “That’s part of the problem. Forgetting is perpetuating the cycle.”

Iran is turning into the sort of ‘forever war’ that Donald Trump promised to avoid

0
iran-is-turning-into-the-sort-of-‘forever-war’-that-donald-trump-promised-to-avoid
Iran is turning into the sort of ‘forever war’ that Donald Trump promised to avoid

When Donald Trump returned to the White House, he promised to avoid another Middle Eastern “forever war”. After the costly experiences of Iraq and Afghanistan, that pledge resonated with an American public weary of prolonged military intervention.

For a brief moment, it appeared that this promise might be realised. The memorandum of understanding (MoU) signed by the US and Iran on June 17 seemed to offer a pathway away from escalating confrontation. Markets stabilised, regional tensions eased and there was cautious optimism that Washington and Tehran had stepped back from the brink.

That optimism has proved remarkably but unsurprisingly short-lived. US military strikes have resumed and Iran has retaliated against American and its allies across the Gulf. Iran’s proxy allies, the Houthis, have declared themselves ready to close the Red Sea chokepoint at the Bab-el-Mandeb Strait, which has provided an alternative transit route for some oil exports, particularly from Saudi Arabia.

Washington continues to combine military pressure with calls for renewed negotiations. But rather than representing the collapse of an isolated peace initiative, these developments may confirm the reality of grey zone conflict – where diplomacy manages war rather than ends it.

When Americans think of forever wars, Iraq and Afghanistan understandably dominate the discussion. Those conflicts were characterised by invasions, prolonged occupations, attempts at nation-building and the deployment of hundreds of thousands of troops over many years. Their enormous political, financial and human costs eventually became impossible to sustain.

The confrontation with Iran is fundamentally different. It is unlikely to require occupation of territory and no serious or sustained attempt directly to remake the Iranian state. Instead, it is fought through air strikes, cyber operations, sanctions, maritime confrontation, intelligence activity and proxy warfare. Military escalation is followed by diplomacy, diplomacy gives way to renewed confrontation, and each episode is presented as a separate crisis.

Viewed collectively, these episodes are better understood as the latest phase of a strategic confrontation that began with the 1979 Iranian Revolution rather than as isolated crises. The conflict has evolved over the years through sanctions, proxy warfare, covert operations and cyber conflict into a more direct military confrontation.

The repeated breakdown of de-escalation efforts is often attributed to mistrust or diplomatic failure. Both undoubtedly matter, but they do not fully explain why the MoU has unravelled so quickly.

Many of its core provisions have become sources of dispute. Israel has continued military operations in Lebanon, despite expectations that regional hostilities would subside. Washington initially signalled greater economic flexibility towards Iran, including easing restrictions on oil exports and discussions surrounding Iranian assets. Yet sanctions have subsequently expanded and further economic measures have been introduced.

From Tehran’s perspective, these actions represent a breach of the spirit – and, in some cases, the letter – of the agreement. From Washington’s perspective, they are responses to continued Iranian activity that it considers destabilising.

The Strait of Hormuz illustrates how quickly ambiguity can become renewed confrontation. Article 5 of the MoU envisaged Iran facilitating the safe passage of commercial shipping through the waterway. Commercial traffic initially increased following the agreement.

But disagreement soon emerged after Oman – in coordination with the International Maritime Organization and with US backing – established a temporary maritime corridor closer to the Omani coast. Tehran viewed this as inconsistent with the agreed framework and argued that it undermined Iranian responsibility for administering safe passage. Washington rejected that interpretation, maintaining that additional security measures were necessary to protect international shipping.

The competing interpretations rapidly became another source of escalation.

The conflict itself has also evolved. What began largely as strikes against military infrastructure has increasingly extended to facilities with direct consequences for civilian life. This has included energy infrastructure and desalination plants that are essential to freshwater supplies across parts of the Gulf.

Transport infrastructure, including bridges, has also been targeted. This has prompted debate among military analysts about whether operational objectives are expanding beyond coercive signalling – using limited military force to pressure or influence an opponent’s behaviour – towards broader military aims. It has also raised questions about whether some attacks could violate international humanitarian law and, depending on the circumstances, amount to war crimes.

Huge pro-regime demonstration in Tehran, July 17 2026.

Despite huge anti-government protests in January, the war with the US has prompted many Iranians to rally round the flag. AP Photo/Vahid Salemi

While the expansion of US strikes in this way does not in itself demonstrate that a US ground offensive is imminent, it reflects a widening conflict whose effects increasingly reach civilian populations.

Diplomacy is managing conflict, not resolving it

The deeper problem is that the strategic objectives of main players here remain fundamentally incompatible. The US continues to insist that Iran’s nuclear programme and regional influence must be constrained. Iran regards those same capabilities as essential to its security and deterrence. Israel continues to view a strategically stronger Iran as an unacceptable security threat.

These objectives cannot all be achieved simultaneously. As a result, diplomacy struggles – not because negotiators lack skill, but because they are attempting to reconcile goals that remain at odds. The MoU temporarily reduced tensions, but it did not alter the structural realities driving the conflict.

This has important implications for how the current crisis should be understood. Much of the debate focuses on whether the latest escalation will produce a wider regional war. That possibility cannot be dismissed. Yet another scenario may be equally, if not more, plausible. The US and Iran are settling into a prolonged cycle of calibrated confrontation that lasts far longer than many currently anticipate.

Trump promised to end America’s forever wars. Ironically, his presidency may instead come to symbolise the evolution of a new kind of forever war without a clearly defined beginning or end. The greatest danger may therefore not be another Iraq. It may be the gradual normalisation of an enduring conflict that repeatedly disappears from the headlines, only to return in new forms because the political conditions necessary for genuine peace have yet to emerge.

This is a new normal for a US foreign policy establishment that retains militarist mindsets, a massive global military capacity and footprint, and values global primacy in an age of power shifts towards a more multi-polar world.

Amid grinding war, protesting Ukrainians still want their voices and values heard at home

0
amid-grinding-war,-protesting-ukrainians-still-want-their-voices-and-values-heard-at-home
Amid grinding war, protesting Ukrainians still want their voices and values heard at home

Ukraine is experiencing its largest domestic protests since Russia’s full-scale invasion in February 2022. The spark to the unrest was the summary dismissal on July 15, 2026, of the country’s youthful defense minister, Mykhailo Fedorov, appointed just six months ago.

Fedorov, seen as charismatic and tech savvy, was already viewed by some as a leading light for change, with a reformist message that chimed with many Ukrainians. So his ouster – the consequence of a clash with Oleksandr Syrskyi, the head of the country’s armed forces and part of the older guard — tapped into a generational clash over how best to defeat Russia.

Frustrations, though, have been building in Ukraine for some time, including over contentious recruiting methods and corruption allegations. Indeed, the latter was the primary reason for demonstrations in Ukraine in the summer of 2025.

Wars are by nature national emergencies, during which politics as normal is often suspended. Before the war and especially after 2014, Ukraine had a lively democracy and public sphere. After the 2022 invasion, Ukraine introduced martial law, placed its television stations under central government control and suspended elections.

For Ukraine, these protests raise important questions about public unity, corruption and the agility of the military-industrial establishment to fight a war now in its fifth year. Our new polling data, gathered by the Kyiv International Institute of Sociology (KIIS), provides important context to the current demonstrations, and insights into the wartime priorities of ordinary Ukrainians. From June 9 to 26, 2026, we surveyed 1,801 people across the country in a computer-aided-telephone survey.

Here are three findings from that survey that provide a broader context to the demonstrations.

1. Anti-corruption agencies and the media need protecting

Ukraine may be at war but fighting corruption at home and maintaining media freedom still matter a great deal to Ukrainians.

Mindful of last summer’s attempt by President Volodymyr Zelenskyy to scrap the independence of two of Ukraine’s anti-corruption agencies, we asked ordinary Ukrainians if they felt these agencies “should be safeguarded” even with Ukraine at war with Russia. We also asked if they felt the same about “freedom of the press and media.”

Protestrs cover their mouths with their hands

Protesting Ukrainians want their voices heard. Paula Bronstein/Getty Images

About half of Ukrainians think that anti-corruption agencies should be safeguarded a lot during the war. An additional 26% of Ukrainians said they should be safeguarded a little. Only a minority (16%) think that anti-corruption agencies should not be safeguarded at all.

Likewise, Ukrainians believe in the vital role of journalism.

Despite war-time media restrictions, Ukraine retains a robust online public sphere and a free press. Indeed, it played a role in mobilizing Ukrainians against the unpopular dismissal of Fedorov.

The protests are in keeping with our poll which finds that a strong majority (62%) of Ukrainians think that freedom of the press and the media needs to be safeguarded a lot. Only about 5% do not think it should be safeguarded at all.

2. Fighting corruption during wartime matters

Previous polling by KIIS revealed that the war with Russia and corruption were what Ukrainians considered the most serious problems facing the country. We decided to follow up by asking Ukrainians what they would rank first as the bigger problem: “Russian aggression” or “government corruption, including around mobilization.”

The results were surprising. Although the most common response was Russian aggression (49%), more than a third of respondents ranked government corruption tied to mobilization as the country’s top challenge. Another 14% said Russian aggression and corruption were equal concerns.

Public opinion research elsewhere reveals that many Ukrainians still optimistically view the war against Russia as a fight for a future in which Ukraine is an established democracy within the European Union. In this sense, the values evident in the survey responses may be what makes the war meaningful for many Ukrainians.

Even during an exhausting war which of late has seen increased Russian attacks on civilians, Ukrainians prioritize domestic reforms and democratic accountability like fighting corruption.

3. Ukrainians oppose drafting younger men

Although manpower is seen as essential for Ukraine’s war effort against a larger Russian adversary, the process of getting new military recruits has proved contentious.

Ukrainian military authorities have sometimes used violent methods to enforce military service, and the lowering of the draft from 27 to 25 in April 2024 was unpopular.

We asked whether Ukrainians agreed that it “was time to draft all young Ukrainian men into military service.” A clear majority of respondents disagreed with this sentiment, with more than a fifth objecting “very strongly.” Less that 8% fully agreed.

A woman walks down the street next to a wall with lots of posters.

A woman passes by a wall covered with military conscription posters for Ukraine’s armed forces on Aug. 5, 2025. Pierre Crom/Getty Images

What this indicates to us is a strong societal consensus that Ukraine should continue to protect its adolescent and young men from military mobilization, despite the persistent recruitment crisis facing its military.

One of the attractions of Fedorov to young people was that his preference for unmanned drones over on-the-ground combat forces offers a promise of Ukrainian victory without large-scale casualties. Syrskyi, by contrast, is associated with Soviet-style attritional warfare. Fairly or not, the latter acquired the nickname “the butcher” some time ago.

Fedorov’s promise of victory through technological transformation to support drone warfare may be an illusion. But his very public attack on Ukraine’s current military-industrial establishment posed a serious challenge to its legitimacy and public support. A figure with close ties to Silicon Valley and political ambitions, Fedorov is a disrupter in the mold of many contemporary tech figures.

Whether his polices ultimately serve the cause of Ukraine at this consequential moment in its war with Russia is an open question. But the demonstrations that followed his dismissal nonetheless suggest the multitude of ways Ukrainians view democracy and freedom at home and abroad.

China-India Great Game taking shape on the Bay of Bengal

0

While the focus of global news remains on Iran, the Strait of Hormuz and the Russia-Ukraine war, a significant new arena of contestation has opened in the Bay of Bengal.

In April this year, India’s National Maritime Security Strategy 2026 (INMSS-2026) identified the eastern Indian Ocean and Bay of Bengal as emerging arenas of long-term strategic competition.

It also highlighted that the next phase of maritime competition would be shaped by ports, logistics networks, supply chains, critical infrastructure, digital connectivity and sustained competition below the threshold of conflict.

China is now moving to accelerate the China-Myanmar-Bangladesh Economic Corridor, which is far more than another Belt and Road Initiative project.

It reflects a reassessment of how Beijing secures access to the Indian Ocean at a time when geopolitical conflict, insurgencies and regional instability are steadily destabilizing several of its principal overland connectivity routes, notably the China-Pakistan Economic Corridor, the Chahbahar-Zahedan Corridor and the Middle Corridor around Iran.

China’s growing emphasis on the China-Myanmar-Bangladesh Economic Corridor underscores both these uncertainties and the Bay of Bengal’s rapid emergence as a strategic frontier that places India’s eastern seaboard, Great Nicobar and wider maritime strategy under sharper focus.

Conflicts and uncertainties elsewhere have made the Bay of Bengal, long regarded as a secondary theater in China’s maritime strategy, more significant to Beijing than ever.

The Iran-US conflict has increased the risks surrounding westward connectivity through Iran and the Persian Gulf. Even if infrastructure remains intact, conflict, sanctions, insurance costs and ongoing hostilities have made long-term security assurance and planning more difficult.

CPEC has encountered a different challenge. Persistent attacks by Baloch rebels on Chinese nationals, transport infrastructure and security installations, coupled with the growing threat posed by Tehrik-i-Taliban Pakistan and instability across the Afghanistan-Pakistan region, have transformed what was conceived as China’s flagship connectivity corridor into a high-risk zone requiring both close monitoring and security management.

Meanwhile, Eurasian routes, including the Middle Corridor, remain exposed to geopolitical tensions, capacity constraints and political uncertainty across multiple transit states.

It’s not that these corridors have ceased to function. Rather, each has become more vulnerable to disruption during periods of crisis. For Beijing, the challenge is therefore no longer simply securing access to overseas markets but ensuring assured access when geopolitical conditions deteriorate.

As confidence in the uninterrupted availability of western corridors weakens, the strategic premium attached to China’s eastern gateway through Myanmar rises correspondingly.

That changing calculus explains why Myanmar has acquired renewed strategic importance. The China-Myanmar Economic Corridor provides China’s landlocked Yunnan province with direct access to the Bay of Bengal through Kyaukphyu, bypassing both the Strait of Malacca and Pakistan.

Existing oil and gas pipelines already provide an alternative energy route into southwest China, while planned road and rail links promise to strengthen commercial connectivity over time. For Beijing, Myanmar is no longer simply another BRI partner; it is becoming an increasingly important pillar of China’s Indian Ocean access strategy.

Bangladesh as Myanmar hedge

Yet Myanmar alone cannot hedge China’s strategic risk. Civil conflict in Rakhine and Shan states continues to complicate infrastructure development and threatens the long-term security of transport routes leading to Kyaukphyu.

Just as insurgency has imposed growing costs on CPEC, Myanmar’s fragmented security environment constrains the predictability of the China-Myanmar corridor.

This is where Bangladesh assumes growing importance. The proposed China-Myanmar-Bangladesh Economic Corridor is an effort to deepen the resilience of China’s Bay of Bengal access by adding further ports, logistics infrastructure, industrial capacity and commercial gateways to an existing network centered on a troubled Kyaukphyu port.

Bangladesh’s ports, particularly Chattogram and potentially Mongla, complement rather than replace Myanmar’s maritime facilities. If instability periodically constrains operations in Kyaukphyu, expanded connectivity through Bangladesh provides additional flexibility while broadening China’s economic footprint across the northern Bay of Bengal, unlike Gwadar in Pakistan.

Viewed in this context, the China-Myanmar-Bangladesh framework seeks to address a resilience problem for Beijing by reducing dependence on a single maritime node while creating a more distributed logistics architecture across the Bay of Bengal. However, the fundamental problems in Rakhine still render the corridor vulnerable.

China’s strategy is also being shaped by changes around the Strait of Malacca itself. India’s agreement with Indonesia to develop Sabang Port, located near the northern entrance to the Strait of Malacca, complements New Delhi’s efforts to transform Great Nicobar into a major hub for logistics, surveillance and maritime operations. Together, they strengthen India’s ability to monitor traffic entering and leaving the eastern Indian Ocean.

Indonesia’s expanding defense partnerships with the US, Japan and Australia further reinforce maritime cooperation, domain awareness and security coordination around Southeast Asia’s principal sea lanes. None of these arrangements is directed solely at China, but collectively they complicate Beijing’s long-term assumptions about uninterrupted access through Malacca.

For Chinese planners, the logical response is to strengthen the one overland corridor that provides relatively direct access from Yunnan to the Bay of Bengal. As the strategic environment around both western land corridors and the Malacca gateway becomes more contested, uninterrupted connectivity through Myanmar assumes greater value.

Extending that connectivity into Bangladesh further enlarges the number of logistics nodes, commercial gateways and maritime options available to China along the Bay of Bengal. The result is the gradual emergence of a more resilient Chinese connectivity network on India’s eastern maritime flank.

India’s counter moves

The implications of the shift extend well beyond infrastructure. They signal Beijing’s redistribution of strategic emphasis towards the Bay of Bengal itself.

For India, whose maritime strategy increasingly emphasizes persistent competition, infrastructure, logistics and maritime domain awareness, the question is no longer whether China will gain another route to the Indian Ocean.

It is whether the Bay of Bengal is becoming the principal arena through which Beijing seeks to secure its long-term strategic presence in the eastern Indian Ocean — a development that would place India’s maritime doctrine, Great Nicobar strategy and eastern seaboard at the center of the next phase of Indo-Pacific competition.

China’s emerging Bay of Bengal strategy will reshape India’s maritime security environment. The significance of the proposed China-Myanmar-Bangladesh corridor lies not simply in another transport network but in the possibility that China is gradually constructing a distributed logistics and connectivity architecture across the northern Bay of Bengal. This presents India with a more complex strategic challenge than any single port or infrastructure project.

This is precisely the kind of environment anticipated by India’s INMSS-2026. It does not focus exclusively on naval balances. It argues that economic corridors, commercial ports, energy infrastructure and dual-use facilities are becoming integral components of maritime power.

China’s moves are therefore not isolated Belt and Road investments but part of a bigger strategic play whose cumulative effect could alter the balance of influence across the Bay of Bengal.

For India, Great Nicobar assumes greater strategic importance within this evolving geography. Its major significance lies in its location near the Strait of Malacca and the Six Degree Channel, one of the principal maritime gateways connecting the Pacific and Indian Oceans.

Great Nicobar strengthens India’s ability to sustain maritime domain awareness, logistical support and operational presence across the eastern Indian Ocean. It’s an investment in preserving India’s long-term strategic leverage in one of the world’s most consequential maritime crossroads.

India’s cooperation in developing Sabang Port provides New Delhi with an important strategic foothold near the northern entrance to the Malacca Strait. Sabang complements India’s presence in the Andaman and Nicobar Islands by extending its reach towards one of the world’s busiest shipping lanes.

Together, Sabang and Great Nicobar provide India with greater capacity to monitor commercial traffic, strengthen maritime logistics and reinforce partnerships across the eastern Indian Ocean.

The strategic picture is becoming even more layered because Indonesia itself has expanded defense cooperation with several Indo-Pacific partners. The US and Indonesia have elevated their relationship through a Major Defense Cooperation Partnership that deepens collaboration on maritime security, undersea capabilities, intelligence, cyber cooperation and advanced defense technologies.

Indonesia has simultaneously broadened defense and maritime cooperation with Japan and Australia.

None of these arrangements explicitly seeks to contain China. Collectively, however, they contribute to a denser security architecture around the Strait of Malacca and the eastern Indian Ocean, making China’s search for additional and more resilient access routes through Myanmar increasingly understandable.

Implications for India

The Bay of Bengal is consequently evolving into one of the principal theaters of Indo-Pacific strategic competition. This has direct implications for India’s eastern seaboard. Ports such as Kolkata, Paradip, Dhamra, Visakhapatnam, Chennai and the expanding infrastructure in the Andaman and Nicobar Islands support a growing share of India’s trade, naval logistics, industrial production and energy flows.

A more substantial Chinese trade and logistical presence across Myanmar and Bangladesh therefore affects not merely naval deployments but the wider security of India’s maritime economy, including supply chains, port infrastructure, shipping services, digital connectivity, subsea infrastructure and trade influence long before it manifests itself in overt military terms.

If there is one conclusion from China’s changing connectivity strategy, it is that INMSS-2026 has identified the direction of maritime competition with remarkable clarity. The contest for influence in the eastern Indian Ocean is no longer defined solely by fleets or freedom of navigation.

It is increasingly shaped by the resilience of supply chains, control of logistics corridors, commercial infrastructure, maritime partnerships and the ability to sustain a presence across strategically important waters. China’s growing emphasis on the Bay of Bengal reinforces — not contradicts — the assumptions underlying India’s maritime doctrine.

The challenge for New Delhi is now about executing the strategy. Great Nicobar, the Andaman and Nicobar Command, Sabang, stronger maritime partnerships with Indonesia and other Indo-Pacific partners, enhanced maritime domain awareness and resilient port infrastructure are no longer isolated initiatives.

They form the interconnected pillars of a maritime strategy that increasingly appears calibrated to the strategic geography now emerging across the Bay of Bengal.

Vivek Y Kelkar is a researcher and analyst focused on the intersection of geoeconomics, geopolitics and corporate strategy.

The first trailer for Avengers: Doomsday is finally here

0
the-first-trailer-for-avengers:-doomsday-is-finally-here
The first trailer for Avengers: Doomsday is finally here

With San Diego Comic-Con kicking off later this week, Marvel Studios has released the first official full trailer for Avengers: Doomsday, the penultimate film in the MCU’s Phase Six and the end of the so-called “Multiverse Saga.”

(Spoilers for prior relevant MCU films below.)

Marvel Studios originally planned to build its Phase Six Avengers arc (The Kang Dynasty) around Jonathan Majors’ Kang the Conqueror (and associated variants), introduced in Loki and Ant-Man and the Wasp: Quantumania. But then Majors was convicted of domestic violence, and Marvel fired the actor soon after. That meant the studio needed to retool its Phase Six plans, culminating in the announced return of the Russo brothers, who directed four of the Marvel Cinematic Universe’s most successful films, which brought in more than $6 billion at the global box office.

The Russo brothers decided to take a new direction, bringing the Secret Wars storyline to the big screen. It’s one of their favorites, but before they could take on Avengers: Secret Wars (slated for May 2027), the duo knew they first had to make Avengers: Doomsday, featuring the wildly popular comics villain Doctor Doom, with the return of Robert Downey Jr. to the MCU in that role.

The film takes place 14 months after the events of Thunderbolts*. Per the official logline: “Beloved heroes from three distinct universes will be set on a deadly collision course and face an existential threat unlike anything they’ve ever encountered.” Those universes are Earth-616 (the Avengers, Wakandans, and New Avengers), Earth-828 (Fantastic Four), and the “original” X-Men.

A few good superheroes

Thing shakes hands with a Wakandan as others look on

More new friends!

Shang-Chi facing off against an opponent

Shang-Chi makes a cameo.

Cast-wise, we have Avengers favorites Thor (Chris Hemsworth), the new Captain America (Anthony Mackie), Bucky Barnes (Sebastian Stan), Ant-Man (Paul Rudd), Falcon (Danny Ramirez), and Loki (Tom Hiddleston). Then there’s the Wakandan contingent: Shuri as the new Black Panther (Letitia Wright), M’Baku (Winston Duke), and Namor (Tenoch Huerta Mejia), as well as Alex Livinalli as Namor’s ally, Attuma. And, of course, Downey Jr. as Doctor Doom.

Then there’s the Thunderbolts* (aka New Avengers): John Walker/US Agent (Wyatt Russell), Yelena Belova (Florence Pugh), Bob/Sentry (Lewis Pullman), Red Guardian (David Harbour), and Ghost (Hannah John-Kamen). And the Fantastic Four: Reed Richards (Pedro Pascal), Sue Storm (Vanessa Kirby), Ben Grimm (Ebon Moss-Bachrach), and Johnny Storm (Joseph Quinn). We’ll also see Charles Xavier (Patrick Stewart), Beast (Kelsey Grammer), Magneto (Ian McKellen), Mystique (Rebecca Romijn), Nightcrawler (Alan Cumming), and Cyclops (James Marsden).

For good measure, Marvel threw in Gambit (Channing Tatum) and Xu Shang-Chi (Simu Liu). There will be plenty of cameos, like Steve Rogers (Chris Evans), revealed in one of four sneak teasers last year. We can also expect to see (at least briefly) Peggy Carter, Spider-Man (Tom Holland), Hawkeye (Jeremy Renner), Doctor Strange (Benedict Cumberbatch), Thor’s adopted daughter Love (India Rose Hemsworth), and Scott Lang’s daughter Cassie (Kathryn Newton).

So what’s in the trailer? A heck of a lot of cameos, mostly, as the superheroes from the three universes meet each other over an ominous voiceover by Thor warning of an unprecedented threat. “I’ve fought with many warriors in my time,” we hear Thor say. “They were far stronger than all of us put together. And they died. They died facing enemies and threats that scared me far less than this one. Everything that they sacrificed will be for nothing if we do not stand together. Put aside your petty squabbles and presume nothing except this. If you return, you will return as brothers and sisters. But mark my words, we’re going to need a miracle.”

OK, not exactly the St. Crispin’s Day speech, but the trailer ends with the return of Steve Rogers—and guess what, he’s still worthy of Thor’s hammer, Mjölnir. Could he be the miracle they need? Or is another miracle waiting in the wings?

Avengers: Doomsday hits theaters on December 18, 2026. Avengers: Secret Wars is currently slated for release on December 17, 2027, and will mark the conclusion of the MCU’s Phase Six.

poster art showing cloaked figure in foreground in front of a gold hued tapestry showing a mother and child

Credit: Marvel Studios

China’s mega-oiler could break America’s Pacific base edge

0
china’s-mega-oiler-could-break-america’s-pacific-base-edge
China’s mega-oiler could break America’s Pacific base edge

China’s construction of what may be the world’s largest naval replenishment ship signals its growing expeditionary ambitions — and highlights the vulnerability of the US’s base-dependent model of power projection.

This month, multiple media outlets reported that satellite imagery and state media indicate that a state-owned subsidiary of China State Shipbuilding Corporation (CSSC) has been rapidly constructing what appears to be the world’s largest naval replenishment ship on Longxue Island, near Guangzhou, China, since at least February 2026.

Measuring approximately 270 meters long with a 37-meter beam, the scaled-up fleet auxiliary vessel significantly surpasses China’s current 45,000-ton Type 901 replenishment ship and the US Navy’s newest John Lewis-class oilers.

Featuring a broad, slab-sided hull optimized for volume, the ship incorporates dual-aperture hangars and a stern flight deck to facilitate vertical helicopter replenishment, as well as midship pillar-like vertical structures engineered to transfer fuel, munitions, and dry stores at sea.

This massive logistics asset appears designed to sustain the People’s Liberation Army Navy’s (PLAN) expanding carrier strike groups, including conventionally powered aircraft carriers such as the Fujian and advanced amphibious vessels such as the Type 076, during prolonged blue-water operations far from mainland ports.

The vessel’s rapid assembly underscores a widening disparity between Chinese and US naval shipbuilding capacity, potentially providing China with the critical logistical reach necessary to project maritime power across the Pacific and globally without relying on foreign shore infrastructure.

For a major military power, it may be surprising that China has only one foreign military base at Djibouti, compared to the US, which has around 750 bases worldwide.

As such, China’s power projection efforts have been relatively constrained by the lack of foreign basing, forcing it to rely on an extended supply chain of vulnerable auxiliary ships for such missions.

Although China has an extensive network of dual-use infrastructure associated with its Belt and Road Initiative (BRI), which is primarily framed as economic, it could also support PLA operations worldwide. Facilities like Gwadar, Hambantota, Kyaukpyu, and Honiara could enable China to project power into the Indian Ocean and the Pacific.

From a military perspective, these isolated facilities are vulnerable and hard to defend. Additionally, political instability in BRI host countries threatens their long-term viability.

China’s logistics limitations could become more pronounced if its fourth carrier—widely speculated to be nuclear-powered — enters service. Nuclear propulsion could give the carrier itself exceptional endurance, but its aircraft, escorts, crew and weapons would still require replenishment.

In line with those constraints, China’s new naval replenishment ship could, to some degree but not fully, mitigate the vulnerabilities that come with a lack of foreign basing capabilities.

The massive auxiliary may indicate that China is developing a more self-sufficient expeditionary capability, potentially reducing its dependence on the overseas basing network central to US power projection.

During World War II, before the US had its global network of bases, it had true expeditionary capability, meaning it could project power from its mainland without dependence on foreign bases.

As the US consolidated its global network of military bases, the US Navy arguably became a “garrison force” – one that was dependent on access to foreign bases for power projection.

By the end of the Cold War, US bases were assumed to be sanctuaries against attack and to enable an unimpeded buildup and staging of forces.

The protracted wars in Iraq and Afghanistan also reinforced the idea that coalition-building and foreign basing were more feasible means of power projection than maintaining independent capabilities for that purpose.

However, those assumptions are now seriously challenged. China has developed long-range missiles such as the DF-21 and DF-26 that can threaten US air and naval bases in the Pacific.

Recent fighting in the Middle East offers a warning of what happens when supposedly secure forward bases come under sustained missile and drone attacks.

The ongoing Iran war has seen Iranian missiles and drones leave many of the 13 regional bases used by US troops all but uninhabitable, destroy costly military equipment and inflict casualties.

The Iran War may have forced US regional allies to ask hard questions about the strategic value of US bases in the region and the credibility of US security guarantees. They would now have to contend with the hard reality of Iran’s geographic advantage at the Strait of Hormuz and an entrenched regime that arguably now has every reason to push forward with developing nuclear weapons.

Still, the US will most likely remain in the Middle East, but its regional footprint may change significantly after the conflict. Strategic bases, such as the US 5th Fleet Headquarters in Bahrain and the Al-Udeid Air Base in Qatar, may remain, as withdrawing from these facilities could have serious strategic consequences for US interests.

Smaller US forward bases near Iran, many of which proved vulnerable to Iranian drone and missile attacks, may be closed. In parallel, the US could explore new basing options, possibly in Israel or Jordan.

But what do those lessons mean for the US force posture in the Pacific? Large hubs such as Guam and Okinawa are critical anchors of US presence in the region, but they are increasingly vulnerable to attack. Smaller US “lily pad” facilities in remote Pacific islands may also face similar vulnerability.

China’s growing economic and political influence in the region, coupled with military overmatch against regional states, could prompt US allies and partners that are economically constrained, politically sensitive, and militarily vulnerable to think twice about hosting US forces.

Facing a potentially degraded Pacific basing posture, the US may have to rely more on power projection from the US mainland, as it did during World War II.

In terms of air power, this could mean increased reliance on hypersonic weapons, strategic bombers, and tanker aircraft based on the US mainland for standoff strikes.

At sea, the US may have to relearn how to use the oceans themselves as bases of operations. This could mean re-exploring the sea-basing concept, which uses the sea itself as a base for power projection rather than relying on potentially vulnerable bases and uncertain allies.

While the US Navy maintains formidable strike capabilities, it must also ensure it can stay in the fight after the initial salvos have been fired.

This would entail significant reinvestment in neglected capabilities, such as the Sea Base Ship and other replenishment at sea (RAS) assets, which would enable the US Navy to remain longer in a potential Pacific fight and reduce dependence on potentially vulnerable and politically unreliable foreign basing.

Hence, China’s development of what could be the world’s largest replenishment ship could signal its increasingly ambitious blue-water ambitions while addressing its lack of foreign basing options.

It could also serve as a mirror-image warning to the US, prompting a re-examination of its longstanding assumptions about power projection and a re-evaluation of how it envisions a future Pacific conflict – one in which it could neither use its regional bases nor depend on its allies.

Hamas Storms WFP Humanitarian Aid Warehouse, Demands Ransom for Detained Palestinians 

0
hamas-storms-wfp-humanitarian-aid-warehouse,-demands-ransom-for-detained-palestinians 
Hamas Storms WFP Humanitarian Aid Warehouse, Demands Ransom for Detained Palestinians 


Hamas terrorists raided a World Food Programme warehouse and food distribution site in Jabalia, threatening aid workers and forcing life-saving operations to halt, The Telegraph reported, citing Palestinian sources and United Nations statements, in the latest reported attack on humanitarian operations in Gaza. 

As part of a report detailing accounts of Hamas attacks on humanitarian facilities, the Telegraph described an incident in which terrorists stormed a shelter housing displaced women, children, and elderly people after local residents asked armed Hamas members to move away so the shelter would not become a target. 

“The group left, only to return with greater numbers. They stormed the shelter and opened fire, causing panic and terror among the women and children, and proceeded to arrest several people,” an anonymous Gazan civilian told The Telegraph. 

Hamas refused to release those detained until it was compensated for ammunition used during the operation, demanding about NIS 24,500, a Palestinian source said. 

Earlier, The Telegraph reported that Hamas ambushed a bus transporting local staff working for the Gaza Humanitarian Foundation near Khan Younis. The attack killed five people and resulted in the kidnapping of several others. 

The newspaper said the raid on the World Food Programme site disrupted food distribution after armed Hamas terrorists entered the warehouse and distribution center, threatened staff, and brought operations to a halt. 

Another source cited by The Telegraph said Hamas has also generated revenue by smuggling goods inside trucks belonging to humanitarian organizations as well as commercial shipments. 

The newspaper reported that aid and commercial vehicles were allegedly used to transport various goods for the group, allowing Hamas to raise funds while humanitarian assistance continued entering Gaza. 

The reported incidents describe separate attacks affecting aid organizations, local humanitarian workers, and Palestinian civilians. The World Food Programme warehouse raid halted food distribution, while the attack on the shelter left civilians injured and several people in Hamas custody until payment was demanded for their release. The earlier ambush targeting Gaza Humanitarian Foundation personnel left five local staff members dead and several others kidnapped. 

 

 

Beyond grep: The case for a context-rich AI coding harness

0
beyond-grep:-the-case-for-a-context-rich-ai-coding-harness
Beyond grep: The case for a context-rich AI coding harness

There are a lot of AI coding applications out there, and as impressive as large language models and the agents they enable have become, many of the most recent developments in AI-assisted development have been in the software that manages those models, not just the models themselves.

Earlier this summer, I spoke with the head of product for Claude Code, Anthropic’s Cat Wu, about that company’s approach to building that software.

Wu repeatedly came back to the same point in our conversation: Anthropic’s models (and those of its direct competitors) are improving so quickly that it makes little sense to plan too far ahead or to build opinionated or limiting features around them. Rather, the Claude Code product team attempts to maintain what they call a lean harness.

A harness is the software built around one or more AI models that determines how they are used. It decides what the models see, what actions they can take, and how they interact with the code base. Think of it like a layer between a model and the developer’s actual project.

It’s not that Claude Code as an application and harness has no opinionated features or design choices, but its product team does seem to err on the side of trusting where the models will take them over the next year.

There are other harnesses besides Claude Code, though. There are OpenAI’s Codex, Google’s Antigravity, open source alternatives like OpenCode, and options from startups like Cursor or Augment Code. Each can have different features, emphasis, or opinions about how a model can or should be utilized in agentic workflows.

One key choice that Claude Code makes is to avoid approaches that build a structured context around a codebase in advance by default.

“Going by the evals, we don’t see a measurable change,” she said of those approaches. “And I think we generally lean more toward shipping a leaner harness with fewer opinionated tools and just letting developers add their own if they want.”

Augment Code has made a substantially different bet, although the two companies are not necessarily testing the same interventions or optimizing for the same outcomes; its product pre-indexes the repository using embeddings, a retrieval model, and a vector database, then retrieves conceptually relevant code. So to get the other side of that discussion, I spoke with Vinay Perneti, Augment Code’s VP of Engineering.

Perneti briefly described Augment Code’s differing approach, offered his own team’s evaluations of the advantages, responded to some of the arguments for a leaner harness, and shared his perspective on some of the concerns developers have about AI tools and agentic workflows more broadly.

A conversation with Vinay Perneti

This interview has been edited for length and clarity.

Ars Technica: Can you explain how Augment Code’s context engine works?

Vinay Perneti: So, if I take a step back, what is it that a particular developer is trying to achieve? They want a particular task implemented and they give it to an agent. And the interesting thing with agents today is they have a limited context window, and every time they need to essentially go get all of the context needed for that and then work on it.

There’s two approaches to context. One is grep-based. Claude Code and Codex and other agents have done that. The second is the semantic retrieval bit. For Augment, we’ve always taken the semantic route, and I’ll describe the building blocks of that.

I would say there’s two core pieces. One is, we have an embedding and a retrieval model pair that are working in the system, and then you have a vector database and an entire, highly optimized back-end system that makes it possible to retrieve in sub-milliseconds.

Ars: Does it have an advantage in one kind of code base versus another?

Perneti: Yeah, it actually turns out the advantage comes in large, private codebases. And here is why I say that.

For all the public, open source repos where most of the benchmarks are run, every single model has basically memorized the repo. These models are large enough that they’re actually able to memorize the whole repo. So when you’re trying to get something done, the models kind of already know where to look so they can get to an outcome quickly.

Whereas when you’re doing this in a private repo, a model has never seen that repo. And now the iteration loop for finding the outcome is much longer, right? If you have a semantic understanding of your entire private repo, you can ask a question, and you get to those outcomes much more quickly.

Ars: People are concerned about token efficiency. Does this semantic approach help with that?

Perneti: We’ve certainly seen it in certain situations. In fact, we published a blog post—we ran Terminal-Bench with Claude Code and Augment Code, same model. And we completed at similar accuracy, but we were 33 percent more efficient than Claude Code. So I do see this showing up in terms of making better use of tokens because you’re not spending as much time on the exploration side.

Ars: Anthropic told me that Claude Code does not see measurable eval gains from including more semantic code-navigation tools. But then I look at your blog, and I see you posting benchmarks and other claims that, actually, this really helps. Are you guys measuring different things when you say this, or what are their evals missing?

Perneti: Great questions. I don’t know what specific retrieval engine they used, and I think the other mistake that often people conflate is, not all retrieval systems are equal just like not all databases are equal, right? And by that, what I mean is the models and the system that are working together, which is the context engine, makes a big difference in terms of the quality of outcomes as well.

So, for example, at Augment, we spent about 18 months at the beginning of the company being founded in 2022—this is pre-ChatGPT—researching retrieval and embedding models predominantly for large code bases. So there’s a lot of research that has gone into, when you’re trying to achieve a particular outcome, what are the right pieces of code to get in this embedding space? All of that is encoded into our retrieval models.

And doing that in a very, very fast way is the system that we built around it. So when somebody says, ‘Hey, I tried Claude Code with a RAG implementation, but I’m not seeing benefits,’ that’s because the implementation and the context engine are very, very different, if that makes sense.

Ars: What do you say to the notion that models are improving so fast that building anything with any assumptions at all makes no sense? That’s the argument for a super lean harness or not doing any of this, is just that you don’t know where it’s gonna be in six months or twelve months or whatever as the exponential goes up.

Perneti: There’s a lot of truth to that, but I think you have to think about it from a couple of different perspectives.

At Augment, the way we’ve been thinking about this is there’s two ingredients that you need for higher quality outcomes, intelligence and context. When models are getting really good, intelligence is going to get exponentially better, no doubt about that. But just because they’re more intelligent does not mean they have the context.

Now, a person can get the context that they want by spending the tokens on it. And that’s where the second dimension comes, which is what all of the engineering leaders are asking right now, is cost. How much of your token budget is going towards context gathering and producing the right outcomes? Are you using the model the right way to get the highest quality outcomes?

And that’s where I think harness design and context matter. So to me, it’s a combo of intelligence and context, and then it’s a systems engineering problem at that point: how do you put the right energy into each of these verticals so that you get the most optimal outcome with the least cost?

Ars: We have a fair number of readers who are skeptical of AI in software development. I see two objections that are especially common. One of them is that they still feel that they just can’t trust these systems enough, that they’re going to create technical debt that’s not worth it. The other is that these agentic workflows are too expensive. What do you think?

Perneti: I’ll actually touch on both. Let me preface for both of them. I think the fundamental truth I hope everybody agrees with is models are continuing to improve at an exponential rate. So whatever you’re doing, you want to ride the exponential so that you’re understanding where things are going and you’re orienting yourself to benefit from that.

So, going back to your question around trust and verification and tech debt, it is true if you approach this problem as, ‘I’m handing this off to agents and I’m walking away.’ That’s not how it’s gonna work, and that’s why I point out that it’s teams of humans working with teams of agents, where there are many points where humans are still better suited for judgment, right? Spec reviews is an example. We find that agents are actually not good at writing specs. So you have to really work with an agent to guide the agent to write a good, high-quality spec. But once you have the spec we are at a place, they’re really good at executing. So it makes no sense to not take advantage of that.

The second thing… tech debt is actually very real, by the way. One of the patterns that we noticed internally at Augment—we are obviously very agent-forward—is agents are very good at duplicating code… so we found that we’ve had to do focused sprints on reducing tech debt with agents. The beauty about that is that you can say you can come up with a spec that talks about what it means to reduce tech debt, and they’re really good at executing that. So I think that is the way to tackle that part of it.

And the second part, you talked about cost. I think that there’s an interesting point there. If you believe that this is the way of working moving forward, then how do you get the best outcome for your tokens? And that’s where I think things like context engine make a difference. To me, picking the right model for the right task makes a huge difference.

The second thing—this is much more a philosophical, forward-looking bit—is, as these model capabilities continue to improve in exponential and open source models are going to keep up, there’s going to be a point where I think the proportion of tokens that’s going to frontier labs, to open-source models, is going to start tipping, where your most difficult problem, you’ll still throw at a frontier model, but the coding step that I described—like I know exactly what needs to get done and I know I described that in a spec—an open source model might be able to get that done for you. At that point, your cost drops quite a lot, right? So you need a system that allows you to work this way, and I think the cost will just kind of take care of itself, is how I see it.

Ars: Thank you. I appreciate you making the time.

Perneti: Yeah, I really enjoyed the chat, Sam.

Takeaways

Both Wu and Perneti agree that they expect frontier models to continue to improve at or even above the current exponential, at least for the next year. (Neither explicitly addressed a longer time frame.) Both agreed that this is fundamentally changing the way software development is done for many organizations.

The practical dispute, such as there is one, is whether context should be assembled ahead of time or rediscovered during each task—especially with large, private codebases.

They’re also sometimes measuring and optimizing for different things. Wu specifically said that Anthropic did not find “a measurable improvement in performance” from “a few LSPs available.” It’s important to note that what Wu was describing when she said that (an LSP, or language server protocol) is not as broad as what Augment Code’s harness does.

When Perneti and Wu describe evaluations of tools or systems that go beyond grep, they are not always talking about the same thing. That is one possible reason (among others) why they come to different conclusions.

One thing Perneti was able to speak to that Wu did not is the possibility that frontier models like Anthropic’s Opus or Fable may become too expensive to remain the main or only models in developers’ and organizations’ agentic workflows, such that smaller or open-weight models might become more widely used.

As the compute crunch continues, some open-weight models—including those small enough and sufficiently quantized to run on local hardware, or at least on hardware maintained in-house by the organization—now approach relatively recent frontier performance on some coding tasks. It remains possible that they, too, will continue to improve at a rapid rate, such that they become viable for coding agents in many teams and projects. Teams may increasingly reserve expensive frontier models for the hardest problems while routing more routine, well-specified work to these cheaper or locally operated models.

Regardless of which approach is more dominant a year from now, more autonomous workflows do not eliminate the need for engineering judgment. If anything, it may be more valuable, not less, as the increases to the speed and scale of production make bad decisions carry higher costs.

0FansLike
0FollowersFollow
0FollowersFollow
0SubscribersSubscribe
- Advertisement -
Google search engine

Recent Posts