Johansen.se

AI burns through a lot of resources. And thanks to a paradox first identified way back in the 1860s, even a more energy-efficient AI is likely to simply mean more energy is used in the long run.

For most users, “large language models” such as OpenAI’s ChatGPT work like intuitive search engines. But unlike regular web-searches that find and retrieve data from anywhere along a global network of servers, AI models return data they’ve generated from scratch. Like powering up a nuclear reactor to use a calculator, this tailored process is very inefficient.

One study suggests the AI industry will be consuming somewhere between 85 and 134 terrawatt-hours (TWh) of electricity by 2027. That’s a similar amount of energy as the Netherlands consumes each year. One prominent researcher predicts that by 2030, over 20% of all electricity produced in the US will be feeding AI data centres (huge warehouses filled with computers).

Big tech firms have always claimed to be heavy investors in wind and solar energy. But AI’s appetite for 24/7 power means most are developing their own nuclear options. Microsoft even plans to revive the infamous Three Mile Island power plant, scene of America’s worst ever civil nuclear accident.

Despite Google’s ambitious target of being carbon neutral by 2030, the company’s AI developments mean its emissions have climbed 48% in the past few years. And the computing power needed to train these models increases tenfold each year.

However, Chinese start-up DeepSeek claims to have created a fix: a model that matches the performance of established US rivals like OpenAI, but at a fraction of the cost and carbon footprint.

An environmental game changer?

DeepSeek has created a powerful open-source, relatively energy-lite model. The company claims it spent just US$6 million renting the hardware needed to train its new R1 model, compared with over $60 million for Meta’s Llama, which used 11 times the computing resources.

DeepSeek uses a “mixture-of-experts” architecture, a machine-learning method that allows the model to scale up and down depending on the complexity of prompts. The company claims its model can also store more data and be trained without the need for huge amounts of expensive processor chips.

Compared with its US rivals, DeepSeek promises to do more with less.
Chitaika / shutterstock

In reaction, US chip manufacturing and energy stocks plummeted following investor concerns that AI companies would rethink their energy-intensive data centre developments. As the world’s largest supplier of specialist AI processors, Nvidia saw its share price fall by US$589 billion, the biggest one-day loss in Wall Street history.

Paradoxically, as well as upsetting the performance of US tech stocks, improving the energy efficiency of AI platforms could actually worsen the industry’s environmental performance as a whole.

With tech stocks crashing, Microsoft CEO Satya Nadella tried to bring a longer-term perspective: “Jevons paradox strikes again!” he posted on X. “As AI gets more efficient and accessible, we will see its use skyrocket, turning it into a commodity we just can’t get enough of.”

The Jevons paradox

The idea that energy efficiency isn’t always a good thing for Earth’s resources has been around for well over a century. In 1865, a young Englishman named William Stanley Jevons wrote “The Coal Question”, a book in which he suggested that Britain’s place as an industrial superpower might soon come to an end, due to its rapidly depleting coal reserves.

But to Jevons, frugality was not the solution. He argued: “It is wholly a confusion of ideas to suppose that the economical use of fuel is equivalent to a diminished consumption. The very contrary is the truth.”

According to Jevons, any increase in resource efficiency generates an increase in long-term resource consumption, rather than a decrease. Because greater energy efficiency has the effect of reducing energy’s implicit price, it increases the rate of return – and demand.

Jevons offered the example of the British iron industry. If technological advancements helped a blast furnace produce iron with less coal, profits would rise and new investment in iron production would be attracted. At the same time, falling prices would stimulate additional demand. He concluded: “The greater number of furnaces will more than make up for the diminished [coal] consumption of each.”

More recently, the economist William Nordhaus applied this idea to the efficiency of lighting since the dawn of human civilisation. In a paper published in 1998, he concluded that in ancient Babylon, the average labourer might need to work more than 40 hours to purchase enough fuel to produce the equivalent amount of light emitted by a modern lightbulb for one hour. But by 1992, an average American would need to work for less than half a second to produce the same.

Throughout time, efficiency gains haven’t reduced the energy we expend on lighting or shrunk our energy consumption. On the contrary, we now generate so much electric light that areas without it have become tourist attractions.

Warming and lighting our homes efficiently, driving our cars, mining Bitcoin and, indeed, building AI models are all subject to the same so-called rebound effects identified in the Jevons paradox. And this is why it will be impossible to ensure a more efficient AI industry actually leads to an overall reduction in energy use.

A Sputnik moment

In the 1950s, the US was horrified when the Soviets launched Sputnik, the first space satellite. The emergence of a more efficient rival caused America to allocate more resources to the space race, not less.

DeepSeek is Silicon Valley’s Sputnik moment. More efficient AI will probably mean more distributed and powerful models, in an arms race that is no longer made up only of US tech giants. AI offers superpower status, and the floodgates may now be fully open for the UK and other global competitors, as well as China.

What’s for certain is that in the long term, the AI industry’s appetite for energy and other resources is only going to increase. Läs mer…

Developers need not “worry about bats and newts” before they start building, Chancellor Rachel Reeves has said in a speech that outlined her plans to reform the UK’s planning process. Reeves’ comments suggest construction firms and housebuilders will be allowed to destroy habitat if they pay into “a nature fund” that might finance restoration elsewhere.

As an ecologist (with a passion for bats), I have serious concerns about what this would mean for the UK’s dwindling biodiversity. The comments from the chancellor are, at best, disheartening at a critical time for nature conservation.

Bats and newts are derided as the gum in the wheels of the planning system. But the idea that nature inherently obstructs development and stymies our collective prosperity is wrong. There are many ways infrastucture can be designed to work with nature in mind from the start – often with low cost.

The chancellor’s own calculations are off if she attaches no economic value to nature. In one scientific study that tried to quantify the economic contribution of wildlife, researchers found that losing pest-eating bats in North American farmland would cost farmers several billions of dollars in crop losses.

Blaming wildlife for economic challenges will only worsen the biodiversity crisis. A report from 2023 found that nearly one in six UK species are at risk of extinction, and that the country is one of the most nature depleted in the world.

Rather than weakening protections for nature, the UK should be doing much more to help the plants and animals that call these islands home.

Why we should worry about bats and newts

Populations of the great crested newt halved between 1965 and 1975 and have continued to decline by 2% every five years since. The enormous loss of habitat is partly to blame: half of all ponds vanished in the 20th century and 80% of those remaining are in poor condition. These figures highlight the long-running failure of the planning system to protect nature.

Newts need ponds to breed in, but they also traverse surrounding grasslands and marshes to find food and new homes. Destruction of these habitats will not be easily remedied by digging a new pond elsewhere, with money from the chancellor’s new fund. Connections between habitats are also essential – isolated, artificial ponds are of little use if wildlife cannot reach them.

The UK has lost a vast area of nature habitat within a generation.
Kyaw Thiha/Shutterstock

This approach will be even less helpful to bats, whose habitat requirements are even more varied.

Bats are highly sensitive to environmental changes. The UK is home to 18 species, including the brown long-eared bat and the pug-like barbastelle. Far from being the menace of developers, bats have suffered greatly as changes to buildings have excluded them from making roosts while changes to the wider landscape have made it harder for them to find feeding and breeding sites.

The numbers of some species have shown a small increase since monitoring began in 1998, but a wider perspective is instructive: the barbastelle bat, for instance, has declined by 99% in the UK over the past few hundred years.

The wider decline of nature now poses a terrible strain. Local bat conservation groups have reported an uptick in the number of starving or underweight bats. All UK bats eat insects, so their health is linked with moths and butterflies and other pollinators that knit ecosystems together. Bats are an early warning system for the overall health of our environment.

Develop with nature, not against it

Conservation measures have to be tailored to the relevant species and setting. Careful deliberation in the planning system is important to protect species – it cannot be replaced with a pot of money that each developer pays into.

Take “bat tunnels”, the structures designed to help bats safely navigate developments which recently drew the chancellor’s ire. These tunnels have been installed along the HS2 trainline and, in theory, protect bats from the 220-mph train as it intersects their flight paths.

Bat tunnels maintain connections between habitats, enabling bats to reach their roosting, feeding and breeding sites without risking their lives near roads or other man-made barriers. It’s not just a fatal collision bats risk – noise and pollution also perturb bats and the insects they eat.

While some species might benefit from a simple bat box that allows bats to roost by providing a roosting structure either outside of a building or on trees, others might need more complex changes. Bats rely on sound to navigate, emitting squeaks that bounce around their environment to create an audible impression of the world.

Conservationists might build them flight paths composed of hedgerows and other features that bats can use to orient themselves. This can be particularly important for developments over a large area.

In these instances, it’s important that bats, who may travel several kilometres from their roosts to feeding sites, have well-connected habitats. Fragmenting the landscape leaves smaller and smaller pockets of available habitat which in turn support fewer and fewer species.

Some measures to help wildlife are cheap and easy to implement.
Heather Wharram/Shutterstock

Instead of being an expensive burden, most measures for mitigating development are fairly easy to implement. It could be as simple as maintaining and improving hedgerows or preserving old trees. More ambitious schemes include designing rail lines that allow animals to pass over or beneath.

Instead of weakening protections and treating biodiversity as a hindrance, a smarter approach would be to integrate nature into development from the outset, and so prevent harm to protected sites and reduce the need for compensation later. The Woodland Trust said that “HS2’s assessment of woodland was significantly deficient” and its impacts to ancient woodland could have been avoided with alternative routes or proposals. In lieu of better assessment, the developers ran into avoidable delays.

There is no one-size-fits-all approach to conservation – no big pot of funding that can pay to repair all the damage later. It requires careful, species-specific strategies, because the needs of wildlife vary greatly. Ignoring the necessity of protecting wildlife jeopardises ecosystems which underpin the economy.

Effective conservation is not a barrier to development, but rather, key to a sustainable future, for people, nature and industries.

Don’t have time to read about climate change as much as you’d like?
Get a weekly roundup in your inbox instead. Every Wednesday, The Conversation’s environment editor writes Imagine, a short email that goes a little deeper into just one climate issue. Join the 40,000+ readers who’ve subscribed so far. Läs mer…

The mere mention of Leonardo da Vinci evokes genius. We know him as a polymath whose interests spanned astronomy, geology, hydrology, engineering and physics. As a painter, his Mona Lisa and Last Supper are considered works of mastery.

Yet one great achievement that frequently goes unrecognised is his studies of human anatomy. More than 500 years after his death, it’s time this changed.

Leonardo is thought to have been born on April 15 1452 in Anchiano, a small hamlet near the town of Vinci, close to Florence. His mother was a 16-year-old peasant girl called Caterina di Meo Lippi, and his father was Ser Piero da Vinci, a 26-year-old notary.

Anchiano is believed to be Leonardo da Vinci’s birthplace.
Marco Taliani de Marchio/Alamy

Being illegitimate, the young Leonardo was only permitted an elementary education in reading, writing and arithmetic. He was also barred from becoming a notary, but this worked out to his advantage. Instead of being constrained by life as an officiate, he was free to be creative and explore the world of nature, satisfying his insatiable appetite for knowledge.

The human anatomy became one of his great interests. This was seeded during his time as an apprentice in Andrea del Verrocchio’s bottega (studio) in Florence, where studying the human form was crucial for achieving realistic depictions.

Creating detailed anatomical drawings required precise sketching skills and the ability to accurately depict the structures being studied. As Leonardo’s fascination grew, he would delve deeper into anatomy as a discipline.

Pioneers

This traces back to the 2nd-century Greek physician Galen of Pergamum, whose anatomical descriptions were mostly based on insights he had gained through dissecting animals and studying wounded gladiators. However, he did no human dissections – they were illegal during his time – and many of his extrapolations from animal to human anatomy were wrong.

Galen dissecting a monkey, Veloso Salgado (1906).
wikimedia

It wasn’t until the 14th century that anatomy and medical science advanced thanks to the start of systematic human cadaver dissections. The physician Mondino de Liuzzi, who practised the first public dissections of human cadavers at the University of Bologna, published the first modern anatomical text, Anathomia Corporis Humani, in 1316.

The text was mostly descriptive in nature, like that of Galen, lacking drawings to illustrate anatomy. Subsequent texts on the subject during the 14th and early 15th centuries did contain drawings, but these were basic and unrealistic.

Leonardo advanced this discipline through his remarkable observational skills, knowledge of perspective and, most notably, his outstanding drawing abilities. His anatomical sketches were unlike anything seen before. For example, his sketches of the muscles of the arms and human skull are comparable to illustrations in today’s medical anatomy texts.

Sketches of human muscles, 1515.
Italian Renaissance Art, CC BY-SA

According to Leonardo’s biographer, Giorgio Vasari, the artist “was one of the first who, with Galen’s teachings, began to bring honour to medical studies and to shed real light upon anatomy, which had until that time been shrouded in the deepest shadows of ignorance”.

Leonardo was the first to depict a detailed study of the human spine, showing its natural curvature and correctly numbered vertebrae. He drew and described nearly all the bones and muscles of the body in beautiful detail, as well as investigating their biomechanics.

His studies on the heart combined both experimentation and observation. Using an ox’s heart to understand blood flow though the aortic valves, Leonardo poured molten wax into the surrounding cavities to make a wax cast, from which a glass model of the heart was made. He then pumped water mixed with grass seeds through this model to visualise the flow pattern. From this experiment, he concluded that the vortex-like flow of blood through the aortic valves was responsible for closing them during each heartbeat.

Sketches of the heart, c1507.
MAG, CC BY-SA

Over 450 years later, in 1968, scientists used dyes and radiography methods to observe this blood flow and prove that Leonardo was correct. A study in 2014 using MRI (magnetic resonance imaging) also demonstrated that he had provided a strikingly precise depiction of these vortex-like flows.

Shortcomings

Leonardo may have dissected around 30 human corpses during his lifetime. Most took place at the Santa Maria Nuova hospital in Florence, and later at the Santo Spirito hospital in Rome. The fact he didn’t have more human cadavers to study probably helps to explain why he also got things wrong.

In addition, Leonardo was very influenced by Galen, through his readings of both Mondino de Liuzzi and the Persian writer Avicenna (c980-1037), while also dissecting animals such as dogs, cattle and horses to fill in human anatomical gaps.

This approach is evident in his study of the male and female reproductive system, as I found when carrying out a detailed review of his work in this area. Misconceptions included the presence of three channels in the penis for semen, urine and “animal spirit”. The prostate gland is also missing in all his sketches of the male reproductive system. Meanwhile, he made the uterus spherical (derived from cow dissections), and similarly misrepresented the fallopian tubes and ovaries.

Even then, Leonardo still got a lot right. He correctly depicts the position of the foetus in the uterus, and the umbilical cord anatomy. He also correctly argued that penile erections were caused by blood engorgement and not by air or “vital spirits” flowing into the penis, as suggested by Galen.

Sketch of baby in the womb, c1510-1513.
Wikimedia, CC BY-SA

Where he got things wrong, Leonardo’s shifting focus may also have played a part. His restlessness, disorganised notes and unfinished work suggest ADHD (attention deficit hyperactivity disorder). Equally, this may also explain his boundless curiosity and incredible creativity.

Despite his shortcomings, Leonardo’s anatomical studies were centuries ahead of their time, rivalling modern standards. His work in this area might have been more appreciated had he published it in a book: he had planned one, and is said to have been collaborating with the Renaissance physician and professor, Marc’Antonio della Torre.

Unfortunately, this was cut short with Marc’Antonio’s death in 1511. Leonardo died in 1519 at the age of 67, and while his gifts to the world have received endless attention, his important contributions to anatomy remain overshadowed, and deserve greater recognition. Läs mer…

As I’ve grown older and experienced the vagaries of my ageing memory, I’ve often reflected on the possibility of a miracle cure that would rejuvenate it. As if in answer to my wishful thinking, not one but several reports recently appeared simultaneously in the scientific news, highlighting a trending solution of which I was blissfully unaware.

A welter of articles – The Times of India, MSN, New York Post and others – spoke of an Indian herb called bacopa, or to give it its full botanical name, Bacopa monnieri, an aquatic flower. This wave of publicity resulted in a massive spike in interest: 2,000 monthly searches on Google and a weekly average of 13,000 views on TikTok.

The reason for its global popularity? A new study which concluded that ingesting bacopa brought significant improvements in both memory and cognition skills (concentration, alertness, reasoning and mental flexibility).

All types of memory were improved – short-term memory (verbal and spatial), working memory and episodic memory (memory of everyday events).

The researchers also reported other brain health-related benefits. Anxiety and cortisol levels in the blood were significantly reduced, and sleep quality and serum BDNF were increased by taking a bacopa supplement (BDNF is a naturally produced protein in the brain that stimulates the production of new brain cells in every decade of our life). If I had wanted a miracle, perhaps I had found it.

But one swallow doesn’t make a summer. And neither should a single study set a law in stone.

So, curious as to the weight of evidence, I delved deeper. My search led me to a surprising source – Ayurvedic medicine.

Over many thousands of years, this traditional Indian medical system has expounded the benefits of bacopa. Bacopa is a medhya rasayana, meaning a class of herbs believed to improve mental health, memory and intellect, and promote rejuvenation and longevity.

It would be true to say that millions of people over the centuries have relied on this supplement for health and mental health benefits. However, history and tradition teach us many things, but not all of them are true. And, therefore, I asked myself: what of the scientific evidence?

Flowering bacopa.
bonilook / Alamy Stock Photo

One of the earliest papers on the effects of taking bacopa was in 2008. And though, over the years, it stimulated several more studies favourable to the use of bacopa, the picture of its effectiveness is mixed.

It’s true to say that most of the papers – many of them using the gold standard method of a randomised controlled trial – find that bacopa is positive for improved memory and reduced anxiety. And there is a biological explanation.

Bacopa extract contains many potent substances called “bacosides” that have, among other effects, antioxidant, anti-inflammatory and anxiolytic (anxiety-reducing) properties. But by no means do all studies show that bacopa improves memory and anxiety. In fact, in 2021 a review of bacopa research stated that there are only limited studies (six to date) to establish the memory-enhancing and brain-protecting effects of bacopa.

Safety

Then I asked myself, is it safe? I turned to the US Food and Drug Administration (FDA). If there is an issue with safety and side-effects, the FDA would know.

The FDA has not approved bacopa as a drug and therefore has not made any statements as to its safety or efficacy. However, the way in which a supplement is marketed can lead to the FDA categorising it as a drug. For example, in 2024, a US company selling veterinary products was censured because their marketing of one of them intended it to be used in the cure of chronic seizures and epilepsy in dogs.

The FDA can investigate, censor or fine – without limit – any company which says that its supplement acts like a drug by implying it can be used to prevent, mitigate, treat or cure any illness.

There is a very fine line here. For example, marketing such as, “the control of blood pressure” may lead to a US federal investigation. A company in Houston, Texas, making medical claims for bacopa was given 15 days in a warning letter by the FDA to correct their marketing or face sanctions including fines.

The FDA states: “Dietary supplements are regulated by the FDA as food, not as drugs. However, many dietary supplements contain ingredients that have strong biological effects which may conflict with a medicine you are taking or a medical condition you may have.”

Such effects are known in bacopa because it inhibits an important brain chemical called acetylcholine and therefore could counteract cholinergic drugs for conditions such as dementia, glaucoma and urinary retention.

It is generally safe for most people, but is inadvisable where there are thyroid conditions, asthma, COPD, genital problems, stomach ulcers or if pregnant.

What are we to make of all this? All that glisters is not gold. And the wisdom of the ages is not irrevocable. There may be a frenzy of popularity in the media but that makes bacopa neither effective nor safe.

The moral here is that before spending your hard-earned money on a promising product that has been seized upon by millions, you should pause, read, research, think and then, based on real evidence, commit – one way or the other. After all, since the days of Newton, science has served us pretty well. Läs mer…

Before January 27 2025, it’s fair to say that Chinese tech company DeepSeek was flying under the radar. And then it came dramatically into view.

Suddenly, everyone was talking about it – not least the shareholders and executives at US tech firms like Nvidia, Microsoft and Google, which all saw their company values tumble thanks to the success of this AI startup research lab.

Founded by a successful Chinese hedge fund manager, the lab has taken a different approach to artificial intelligence. One of the major differences is cost.

The development costs for Open AI’s ChatGPT-4 were said to be in excess of US$100 million (£81 million). DeepSeek’s R1 model – which is used to generate content, solve logic problems and create computer code – was reportedly made using much fewer, less powerful computer chips than the likes of GPT-4, resulting in costs claimed (but unverified) to be as low as US$6 million.

This has both financial and geopolitical effects. China is subject to US sanctions on importing the most advanced computer chips. But the fact that a Chinese startup has been able to build such an advanced model raises questions about the effectiveness of these sanctions, and whether Chinese innovators can work around them.

The timing of DeepSeek’s new release on January 20, as Donald Trump was being sworn in as president, signalled a challenge to US dominance in AI. Trump responded by describing the moment as a “wake-up call”.

From a financial point of view, the most noticeable effect may be on consumers. Unlike rivals such as OpenAI, which recently began charging US$200 per month for access to their premium models, DeepSeek’s comparable tools are currently free. They are also “open source”, allowing anyone to poke around in the code and reconfigure things as they wish.

Low costs of development and efficient use of hardware seem to have afforded DeepSeek this cost advantage, and have already forced some Chinese rivals to lower their prices. Consumers should anticipate lower costs from other AI services too.

Artificial investment

Longer term – which, in the AI industry, can still be remarkably soon – the success of DeepSeek could have a big impact on AI investment.

This is because so far, almost all of the big AI companies – OpenAI, Meta, Google – have been struggling to commercialise their models and be profitable.

Until now, this was not necessarily a problem. Companies like Twitter and Uber went years without making profits, prioritising a commanding market share (lots of users) instead.

And companies like OpenAI have been doing the same. In exchange for continuous investment from hedge funds and other organisations, they promise to build even more powerful models.

These models, the business pitch probably goes, will massively boost productivity and then profitability for businesses, which will end up happy to pay for AI products. In the mean time, all the tech companies need to do is collect more data, buy more powerful chips (and more of them), and develop their models for longer.

But this costs a lot of money.

Nvidia’s Blackwell chip – the world’s most powerful AI chip to date – costs around US$40,000 per unit, and AI companies often need tens of thousands of them. But up to now, AI companies haven’t really struggled to attract the necessary investment, even if the sums are huge.

DeepSeek might change all this.

By demonstrating that innovations with existing (and perhaps less advanced) hardware can achieve similar performance, it has given a warning that throwing money at AI is not guaranteed to pay off.

For example, prior to January 20, it may have been assumed that the most advanced AI models require massive data centres and other infrastructure. This meant the likes of Google, Microsoft and OpenAI would face limited competition because of the high barriers (the vast expense) to enter this industry.

Money worries

But if those barriers to entry are much lower than everyone thinks – as DeepSeek’s success suggests – then many massive AI investments suddenly look a lot riskier. Hence the abrupt effect on big tech share prices.

Shares in chipmaker Nvidia fell by around 17% and ASML, which creates the machines needed to manufacture advanced chips, also saw its share price fall. (While there has been a slight bounceback in Nvidia’s stock price, it appears to have settled below its previous highs, reflecting a new market reality.)

Nvidia and ASML are “pick-and-shovel” companies that make the tools necessary to create a product, rather than the product itself. (The term comes from the idea that in a goldrush, the only person guaranteed to make money is the one selling the picks and shovels.)

The “shovels” they sell are chips and chip-making equipment. The fall in their share prices came from the sense that if DeepSeek’s much cheaper approach works, the billions of dollars of future sales that investors have priced into these companies may not materialise.

‘When we find some gold we can invest in AI.’
Everett Collection/Shutterstock

For the likes of Microsoft, Google and Meta (OpenAI is not publicly traded), the cost of building advanced AI may now have fallen, meaning these firms will have to spend less to remain competitive. That, for them, could be a good thing.

But there is now doubt as to whether these companies can successfully monetise their AI programmes.

US stocks make up a historically large percentage of global investment right now, and technology companies make up a historically large percentage of the value of the US stock market. Losses in this industry might force investors to sell off other investments to cover their losses in tech, leading to a whole-market downturn.

And it shouldn’t have come as a surprise. In 2023, a leaked Google memo warned that the AI industry was exposed to outsider disruption. The memo argued that AI companies “had no moat” – no protection – against rival models. DeepSeek’s success may be the proof that this is true. Läs mer…

The pairing of British prime minister Keir Starmer and US president Donald Trump connotes many imponderables. The only certainty happens to be the most significant: they will be in office together for four years.

It is rare for a prime minister and a president to have the luxury of knowing – barring extreme unpredictabilities, such as death or incapacity – they have a full term in harness. And personal chemistry matters.

Trump emphasises (rather too much for the liking of America’s allies) the deal, the handshake, the gaze; the bond that only the lonely, only those who lead, can have. Starmer emphasises level-headedness (although his government has not been particulary conspicuous in evincing it).

Opposites may well attract, but the precedents for coterminous presidents and prime ministers are not encouraging. John Major and Bill Clinton, elected seven months apart, spent 1992 to 1997 together. But in the very definition of what not to do before an election, London had made its preference for the result of the election in America known – and the other guy won. The Conservative and the Democrat were no more than coolly cordial thereafter.

Major awks.
Alamy/Michael Stephens

On his re-election in 2001, Tony Blair knew he had George W. Bush for at least four years – it turned out to be eight – but the consequences for him were disastrous once the two decided to partake in a war on “terror”.

In 1964, Harold Wilson and Lyndon Johnson were elected almost simultaneously, and spent 1964 to 1968 together. Though they were Labour and Democrat, and therefore from sister parties, it was not a harmonious pairing. Wilson’s meddling in, but lack of support for, Johnson’s war in Vietnam was a source of unbridled irritation in the White House.

Trump and May

The last time Trump became president, Theresa May was prime minister and she travelled with undisguised haste to the White House. There she achieved a highly untypical diplomatic coup in getting Trump to commit publicly to Nato (that bars should be so low was a general feature of the presidency).

Their subsequent relationship was, however, toxic. No prime minister has been less likely to gaze, to bond (despite pictures of them holding hands), and the president held her as having mangled Brexit, a bid for freedom with which he was keen to associate himself.

Theresa May and Donald Trump during her visit to the White House, before relations turned sour.
EPA

Before the US election, Starmer displayed a unfamiliar deftness of touch, and banked some credit. His immediate phone call to candidate Trump following an attempt on his life in July was both bold and smart. There followed the fabled Trump Tower two-hour chicken dinner.

It was more typical for Starmer that when it emerged, in a most unfortunate echo of 1992, Labour activists – and Starmer’s own pollster – were working on the Kamala Harris campaign, Trump’s people cried foreign interference and threatened legal action.

And the two in Starmer’s team who will have the most exposure to the new administration have both been publicly rude about Trump. David Lammy, now foreign secretary, called him “deluded, dishonest, xenophobic [and] narcissistic” in 2019.

Peter Mandelson, nominated but not yet confirmed as the UK ambassador to the US, has made comments about Trump being a “bully” and a “danger to the world”. To appease opposition in DC on his appointment, Mandelson has since turned on a sixpence (or perhaps a dime).

This is, at root, about Trump. No other president would have attracted such comments from frontline politicians. But from TV studio to TV studio, Lammy and Mandelson will have those quotes hung about their necks as if they were modern-day ancient mariners. Starmer’s innate caution in public utterance, in this area at least, has inured him.

Indeed, the repercussions of his unusual boldness in picking Mandelson over a career diplomat may discourage Starmer from ever thinking imaginatively again.

Most members of the Trump administration would be naturally hostile to a Labour government even without its leading figures insulting their boss or campaigning for his opponent. Certainly, the grounds for disagreement are great: the threat of tariffs, demanded increases in defence spending, the sovereignty of the Chagos Islands, co-operation with China and support for Ukraine.

Thus Morgan McSweeney – architect of Labour’s 2024 victory, planner of its re-election and Starmer’s chief of staff – flew out to meet Susie Wiles, his equivalent in the White House. (It did not, a person privy to such information told me, go well. Voices were raised.)

Elon Musk, this moment’s most prominent presidential acolyte inveighed on X, “Starmer must go”, adding for good measure, “He is a national embarrassment.” It is indeed embarrassing – for Starmer – but he will be consoled with the well-founded suspicion that the life-expectancy of Musk and Trump’s tech bromance will be much less than four years.

Cause for self-reflection

The return of Trump, emboldened and more powerful than before, has effectively forced the posing of the age-old question: over which expanse of sea should Britain gaze – the Channel or the Atlantic? Churchill thought it should – and that only Britain could – do both.

Hence, perhaps, Trump’s own public statement about the possible destination of his first international trip: “It could be UK. Traditionally, it’s been UK.”

It hasn’t. Only Jimmy Carter, in 1977, and Joe Biden, in 2021, visited the UK first – and then because of summits. More than a few presidents (most recently Ford and Johnson) didn’t visit at all.

But even what might have been a supportive comment was laced with arsenic: “Last time, I went to Saudi Arabia because they agreed to buy 450 billion dollars’ worth of United States merchandise … And if that offer were right, I’d do that again.” Which at least may free the British government to be as unsentimentally transactional.

Trump and Starmer achieved big victories, albeit when painted in the most flattering terms. Starmer’s came on a historically low combination of vote share and voter turnout, Trump’s with fewer votes than Biden. But Trump will like that Starmer won a large majority. When May managed to lose hers in 2017, what little respect Trump had for her went with it.

Starmer would much rather have had four years with Biden, and even more with Harris, another public prosecutor of the left. But he has to deal with the transatlantic relationship as it is, rather than as he would wish it to be, and this one is most unlikely to be special.

Starmer is, moreover, a realist. Which is why he’ll also know that the second Trump presidency will be much more consequential than the first. Caution may have limited effect. Läs mer…

A vote in Germany’s national parliament (Bundestag) has led to fears that the firewall supposedly separating mainstream political parties and the far-right Alternative for Germany (AfD) has been blown apart.

Until now, Germany’s largest parties, including the union of Christian democratic parties the CDU and CSU, and the social democrat SPD, have ruled out any form of cooperation with the AfD. Friedrich Merz, CDU leader and most likely chancellor following the election to be held on February 23, had previously said that decisions in the Bundestag should not be passed if they relied on AfD votes.

And while Merz’s commitment to the firewall had occasionally wavered in some interviews, the CDU had resisted any temptation to do deals with the AfD nationally or in state parliaments. There is some cooperation on a local level, but beyond a vote on local taxation in Thuringia in 2023, mainstream parties have eschewed any hint of state or national level cooperation.

That has now changed. Apparently in response to the AfD’s promising polling ahead of the election on February 23, the CDU has tacked dramatically to the right on immigration policy. Merz introduced a five-point plan into the Bundestag proposing a significant tightening of Germany’s immigration system.

Most radical among the proposals is the reintroduction of border controls at German borders and for migrants without permission to reside in Germany to be turned back. These measures would be questionable, at best, in their conformity with European law.

Merz made it plain he would put this plan to a vote, even if it could pass only with AfD support. This it did, by 348 to 345. The CDU and its sister party the CSU voted in favour, alongside the AfD and the Free Democratic Party (barring a handful of rebels). The SPD, Greens and Left party voted against while the anti-immigration “left populist” Sarah Wagenknecht Alliance abstained.

This was not a binding vote but Merz can now push for a more formal process to make his five-point plan law. It is also highly symbolic.

The AfD was gleeful, hailing a “turning point”, or Zeitenwende, in migration policy. It celebrated the “fall of the firewall” and a “great day for democracy”. The SPD and Greens were furious, with outgoing chancellor Olaf Scholz accusing Merz of breaking his word – and breaking with the tradition of former chancellors from Konrad Adenauer to Angela Merkel by relying on votes from the far right. Merkel subsequently underlined Scholz’s point by criticising Merz’s move.

The Greens talked of a “dark day for our democracy”. A Left Party parliamentarian called out “to the barricades”, and some spontaneous demonstrations occurred around the country. Merz said he “regretted” that the vote had only been possible with AfD support but added that “doing the right thing does not become wrong when the wrong people – the AfD – vote for it”.

An election ahead

Merz’s changed position on immigration and the AfD has come a few weeks ahead of an election that had initially got off to a slow start. The campaign is now suddenly polarised and angry on all sides.

The election is being held because the three-party governing coalition of social democratic SPD, Greens and liberal FDP collapsed in November over disputes on fiscal policy. Opinion polls have been quite stable, showing the CDU/CSU leading. However, Merz’s party would need a coalition partner.

The AfD has been consistently in second place but the firewall would prevent a coalition. This helps explain why reactions to the Bundestag vote have been so fierce.

Chancellor Olaf Scholz looks on while Friedrich Merz speaks during the Bundestag debate.
Alamy/dpa picture alliance

And while the government collapsed because of disagreements over the economy, several high-profile stabbings by migrants have turned this into an election about immigration. Indeed, migration, asylum and security questions are now right at the top of the list of voters’ concerns.

The AfD has the wind in its sails and is basking in the glow of Elon Musk’s noisy endorsements. It has adopted an even more hardline manifesto than its previous offerings, proposing “remigration” as a policy – code for removing legal migrants who are no longer welcome.

However, it is important to note that with this vote, Merz has not declared open season for a coalition with the AfD. And if a coalition was formed with the SPD or Greens, there is no way it would survive Merz turning to the AfD for support on issues where the coalition partner disagreed.

Scholz has warned of the risk of events similar to Austria, where the CDU/CSU’s sister party, the ÖVP, initially ruled out going into government with the far-right FPÖ but changed its stance when negotiations with mainstream parties failed. Merz insists this won’t happen but moderate CDU/CSU voters may heed Scholz’s warnings and look elsewhere. Merz’s gamble is that such losses would be offset by voters who support a harder line on migration – and even that he will win voters back from the AfD.

These events highlight the debate being had ever more often across Europe. Are far-right parties weakened if their positions are, to a degree, accommodated by the political mainstream? Or does this in fact strengthen and embolden them?

That debate will continue but there are more immediate consequences in the wake of the Bundestag vote. Germany’s neighbours will look on uneasily, both because of the febrile political atmosphere in the largest EU member state at a time of substantial geopolitical pressure and because, if Germany were to be found to have set aside European law, that could trigger a total unravelling. Läs mer…

“Just Because I am a Librarian doesn’t mean I have to dress like one.”

With this breezy pronouncement, Belle da Costa Greene handily differentiated herself from most librarians.

She stood out for other reasons, too.

In the early 20th century – a time when men held most positions of authority – Greene was a celebrated book agent, a curator and the first director of the Morgan Library. She also earned US$10,000 a year, about $280,000 today, while other librarians were making roughly $400.

She was also a Black woman who passed as white.

Born in 1879, Belle was the daughter of two light-skinned Black Americans, Genevieve Fleet and Richard T. Greener, the first Black man to graduate from Harvard. When the two separated in 1897, Fleet changed the family’s last name to Greene and, along with her five children, crossed the color line. Belle Marion Greener became Belle da Costa Greene – the “da Costa” a subtle claim to her Portuguese ancestry.

One of the nine known portraits of Belle da Costa Greene that photographer Clarence H. White made in 1911.
Biblioteca Berenson, I Tatti, the Harvard University Center for Italian Renaissance Studies

When banking magnate J.P. Morgan sought a librarian in 1905, his nephew Junius Morgan recommended Greene, who had been one of his co-workers at the Princeton Library.

Henceforth, Greene’s life didn’t just kick into a higher gear. It was supercharged. She became a lively fixture at social gatherings among America’s wealthiest families. Her world encompassed Gilded Age mansions, country retreats, rare book enclaves, auction houses, museums and art galleries. Bold, vivacious and glamorous, the keenly intelligent Greene attracted attention wherever she went.

I found myself drawn to the worlds Greene entered and the people she described in her lively letters to her lover, art scholar Bernard Berenson. In 2024, I published a book, “Becoming Belle Da Costa Greene,” which explores her voice, her self-invention, her love of art and literature, and her path-breaking work as a librarian.

Yet I’m often asked whether Greene mentions her passing as white in her writings. She did not. Greene was one of hundreds of thousands of light-skinned Black Americans who passed as white in the Jim Crow era. While speculation about Greene’s background circulated in her lifetime, nothing was confirmed until historian Jean Strouse revealed the identities of Greene’s parents in her 1999 biography, “Morgan: American Financier.” Until that point, only Greene’s mother and siblings knew the story of their Black heritage.

“Passing” can often raise more questions than answers. But Greene did not largely define herself through one category, such as her racial identity. Instead, she constructed a self through the things she loved.

‘I love this life – don’t you?’

In my view, any consideration of Greene’s attitudes toward her own race must remain an open question. And uncertainty can be acknowledged – even embraced – with judgments suspended.

The Morgan Library & Museum currently has an exhibition on Greene that will run until May 4, 2025 – one that’s already generated debates about Greene and the significance of her passing.

One section of the exhibition, “Questioning the Color Line,” includes novels on passing, paintings such as Archibald J. Motley Jr.’s “The Octoroon Girl,” photographs of Greene, and clips from Oscar Micheaux’s 1932 film “Veiled Aristocrats” and John M. Stahl’s 1934 film “Imitation of Life,” which portray painful scenes between white-passing characters and their family members.

None of these objects clarifies Greene’s particular relationship to passing. Instead, they place the librarian within melodramatic and conventional representations about passing that stress self-division and angst.

We don’t know – perhaps we will never know – whether Greene had similar moments of self-doubt.

Greene frequently received glowing press coverage.
The Morgan Library & Museum

Yet some critics have concluded as much. In his review of the exhibition for The New Yorker, critic Hilton Als laments what Greene’s passing had cost her. He describes her as a “girl who loved power,” a woman who “became a member of another race – not Black or white but alternately grandiose and self-despising.”

There’s a lot of certainty in such a pronouncement – and scant evidence furnished to support such declarations.

New York Times columnist John McWhorter takes issue with Als’s depiction of the librarian’s passing in a Jan. 23, 2025, article.

Citing passages from her letters in which Greene excitedly describes reading the Arabic folktales “The Thousand and One Nights” and seeing exhibitions of modern art, McWhorter asks readers to reconsider this “witty, puckish soul who savored books and art” and “had an active social life.”

What if Greene gave her race little thought, McWhorter wonders. What if she simply saw the notion of race and racial categorization as “a fiction” and instead lived her life to its fullest? Of course, her light skin afforded her the opportunity that other Black people of her era didn’t have. But does that necessarily mean that she was self-loathing or conflicted?

“[W]e are all wearing trousers and I love them,” Greene writes in one letter to Berenson, adding, “The Library grows more wonderful every day and I am terribly happy in my work here … I love this life – don’t you?”

Greene’s vitality captivated Berenson, who once described the librarian as “incredibly and miraculously responsive.”

The connoisseur was not the only contemporary who admired Greene’s effervescence. In “The Living Present,” an account of the activities of women before and after World War II, Greene’s friend Gertrude Atherton paid tribute to Greene, a “girl so fond of society, so fashionable in dress and appointments” that she could impress any stranger with her “overflowing joie de vivre.”

Crafting an aura

Viewed through a more expansive lens, Greene’s passing can be seen as part of an exercise in self-fashioning and self-invention.

Greene dressed to be noticed – and she was. Meta Harrsen, the librarian Greene hired in 1922, offers a rare eye-witness account. On the day Greene interviewed Harrsen, “she wore a dress of dark red Italian brocade shot with silver threads, a gold braided girdle, and an emerald necklace.”

Greene understood well the power of clothes to project a distinct identity – a highly crafted one in this case, and one befitting a connoisseur of rare books.

Greene poses for a Time magazine portrait in 1915.
The Morgan Library & Museum

At that, she excelled. She became known for her stunning acquisition coups: her purchase of 16 rare editions of the works of English printer William Caxton at an auction; her procurement of the highly coveted Crusader’s Bible through a private negotiation; and her acquisition of the Spanish Apocalypse Commentary, a medieval text written by a Spanish monk that Greene was able to buy at a steep discount.

To me, a 1915 photo captures Greene’s confidence and aura more than any other image of the librarian.

She posed in her home and wasn’t shot in soft focus with a studio backdrop as other photographs tend to portray her. Sitting on the arm of a large chair upholstered in a tapestry weave, she wears an elaborate hat with a large ostrich plume, a high-necked blouse under a long, loosely belted jacket with a ruffled cuff over a long dark skirt. The decor is no less striking: Flemish tapestries decorate the walls behind her, and a liturgical vestment is draped over the bookcase. Looking directly at the viewer, Greene is assured and poised.

Greene’s stylish flair was not simply decorative. It was a testament to her vibrant personality and the joy she took in her work. Rather than judge her according to contemporary notions of racial identity, I prefer to marvel over her achievements and how she became a model for generations of future librarians.

Greene didn’t just pass. She surpassed – in spectacular ways. Läs mer…