Insider

this-giant-dome-battery-cuts-co2-emissions-—-by-using-more-co2

This giant dome battery cuts CO2 emissions — by using more CO2

Insider, Next Featured, Startups and technology, Sustainability / /

Renewable energies like wind and solar are clean, abundant, and cheap — but notoriously unpredictable. That’s why so much time and money has been pumped into scaling energy storage solutions: we need to keep the lights on even when the wind isn’t blowing or the sun isn’t shining.

While lithium-ion batteries have received the bulk of this investment, there’s another kid on the block that could be cheaper and greener. In an ironic twist, the whole system is powered by the same molecule it is attended to fight — carbon dioxide. 

Imaginatively, it is called the CO2 battery. The way it works is relatively simple. CO2 gets stored in a gigantic dome. When charging, the system pulls the gas from the dome, compresses it into a liquid and stores it in big carbon steel tanks. The compression process also produces heat which is stored in ‘bricks’ made of steel shot and quartzite for later use.  

Then, when power is needed, the liquid carbon dioxide is heated up using the hot bricks, rapidly turning it back into a gas — which refills the dome. On its way back to the dome, however, the gas spins a turbine, producing electricity.

an image of Energy Dome's pilot plant in Sardinia, Italy
Energy Dome’s first pilot plant near Ottana on the island of Sardinia, Italy. Credit: Energy Dome

And what about all the CO2 to fill that dome, you may ask? Well, it’s a closed-loop system so you only need to inject gas into the dome once across the battery’s entire 30-year lifespan. So by using a pinch of CO2 it can support the rollout of renewable energies that can cut our emission of the gas altogether. 

‘Half the cost of lithium-ion’

The brainchild of Italian startup Energy Dome, the battery builds upon existing compressed air and liquid air energy storage technologies. Except, the use of CO2 brings a couple of distinct advantages. 

Pure carbon dioxide is a lot denser than air, which means you can store the same amount of energy in a much smaller space. Up to ten times smaller than compressed air, in fact. And while liquid air energy storage is admittedly more space efficient than either CO2 or compressed air, it must be cooled to almost -200 degrees Celcius to achieve the desired results. This requires a lot of energy, which cuts efficiency, and is why liquid air energy storage has struggled to compete with other storage technologies on cost. 

But affordability is exactly where CO2 batteries excel. They’re built using steel, carbon dioxide, and water. That’s it. The rest of the components — like pipes, compressors, and turbines — can be purchased off the shelf. According to Energy Dome, this means its system can produce electricity at half the cost of lithium-ion batteries. 

Those are some impressive figures, which have naturally caught the attention of investors. At COP28 last week, Bill Gates’ Breakthrough Energy Ventures and the European Investment Bank jointly committed €60mn to help Energy Dome build its first commercial-scale plant on the island of Sardinia, Italy. This adds to the €80mn in funding the startup has already secured. 

‘Game-changing technology’

The CO2 battery will store some 20MW of renewable energy supplied by nearby solar and wind farms on the island. Energy Dome already built a demonstration plant on the island last year. The smaller, 2.5MW, facility is currently operating and transmitting power to the grid. 

Gelsomina Vigliotti, vice president at the EIB, called the initiave an “inspiring example of game-changing technology that we need more of in Europe and worldwide”. 

Energy Dome’s founder, Claudio Spadacini, said the Sardinia plant will be the “first of many identical full-scale CO2 batteries”. The company said that the modular, simple design of its CO2 battery means it can be scaled relatively rapidly. 

The company has already signed a deal with Norwegian wind energy giant Ørsted to install “one or more” of the CO2 batteries at its sites in Europe. If all goes well, construction on the first storage facility using Energy Dome’s CO2 battery could begin in 2024.

lithium production in Chile
Brine evaporation pools at a lithium mine in Argentina. The environmental costs of lithium mining are not always factored into the pricetag of the batteries that they are used to produce. Credit: Anita Pouchard Serra/Bloomberg

While lithium-ion batteries will no doubt continue to play an important role in the energy transition, the negative environmental and social consequences of their production have been thrown into the spotlight in recent years. They rely on a number of rare earth metals like lithium, nickel, and cobalt, the mining of which has been linked to extensive environmental degradation and even human rights abuses the world over.

If CO2 batteries can circumvent some of these impacts and undercut lithium-ion on cost, who knows, perhaps they could become the next big thing in energy storage.

This giant dome battery cuts CO2 emissions — by using more CO2 Read More »

uber-seeks-unlikely-alliance-with-london’s-iconic-black-cabs

Uber seeks unlikely alliance with London’s iconic black cabs

Corporates and innovation, Ecosystems, Insider, Ride-share / /

Starting early next year, commuters in London will be able to hail the capital’s iconic black cabs via Uber, the American ride-hailing giant has announced.  

London’s black taxi drivers — famous for their uncanny knowledge of the city’s thousands of streets — have long been at odds with Uber, who they say threatens their livelihoods. Frustrated drivers even blocked London streets in 2014 in protest against the tech company’s famously aggressive expansion tactics, and relations still remain tense.

Black cabs are currently the only taxis in London licensed to pick up passengers from the streets in the city and are already available for bookings through apps like Gett, Taxiapp, and FreeNow.

While Uber is playing off the new deal as a partnership, the Licensed Taxi Drivers’ Association (LTDA), which represents more than 10,000 taxi drivers, said it was not consulted ahead of Uber’s “unilateral announcement”.

Steve McNamara, a spokesperson for the organisation, said it has no interest in “sullying the name of London’s iconic, world-renowned black cab trade by aligning it with Uber, its poor safety record and everything else that comes with it.”

Uber, however, claims a “small number” of taxi drivers have already signed up to the service and it hopes to recruit “several hundred” by January. The company said it would not charge new drivers commission for their first six months but didn’t reveal what the fee would be after that period.

While it remains to be seen whether Uber will woo London’s black cab drivers, it wouldn’t be the first time it has turned former foes into friends. 

The ride-hail giant recently signed on taxi fleets in Los Angeles, New York City, Paris, and Rome to list drivers on the app. Uber says in Europe and the Middle East, over 10% of Uber trips are now completed by taxi drivers.

Published

Back to top

Uber seeks unlikely alliance with London’s iconic black cabs Read More »

first-transatlantic-flight-with-100%-‘sustainable’-fuel-is-greenwashing,-critics-say

First transatlantic flight with 100% ‘sustainable’ fuel is greenwashing, critics say

Corporates and innovation, Government and policy, Insider, Sustainability / /

The world’s first transatlantic flight with 100% sustainable aviation fuel (SAF) has been attacked as “greenwashing” by critics.

The Tuesday trip from London to New York on a Vrigin Atlantic 787 has been celebrated by airlines and politicians as a “milestone” in the journey to net zero. Scientists and climate campaigners, however, have poured scorn on these claims.

Opinions are divided over the potential of SAFs, which derive from various alternatives to fossil fuels. For Tuesday’s flight, the SAF was made primarily from waste fats and plant sugars, according to a Virgin Atlantic factsheet [PDF] shared with TNW. The airline expects the resulting carbon emissions to be 70% lower than those produced by petroleum-based jet fuel.

Shai Weiss, Virgin Atlantic’s CEO, said the Boeing 787 test flight would prove that SAF “can be used as a safe, drop-in replacement for fossil-derived jet fuel.” She added that it was “the only viable solution for decarbonising long-haul aviation.”

The reactions from the UK government — which partly funded the fight — have been even more optimistic. Prime Minister Rishi Sunak praised the journey as “the first net zero transatlantic flight,” while the Department for Transport declared that it was “ushering in a new era of guilt-free flying.” Both claims were promptly pilloried by environmental groups.

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Cait Hewitt, policy director of the Aviation Environment Federation (AEF), said promises that the trip will take us closer to “guilt-free flying” were “a joke.” She notes that SAFs currently comprise around 0.1% of global aviation fuel — and will be extremely hard to scale sustainably.

“Trying to scale up production of alternative fuel, using the waste products of fundamentally unsustainable industries like intensive animal agriculture, or using unrecyclable plastic — as the UK government is planning to do – is hardly a sustainable solution,” Hewitt told TNW.

She added that SAFs also emit as much CO2 as kerosene at the tailpipe. This is because they’re still hydrocarbon fuels and produce the same volume of CO2 emissions as kerosene when combusted. According to the AEF, any CO2 reduction will be “net” savings made during the production phase — as with a carbon offset.

“Linking it to the idea of ‘guilt-free flying’ is deeply misleading and risks setting back a proper, honest discussion about how much we can fly while achieving climate goals,” Hewitt said. “If the public is led to believe the industry has found the solution to green flying, that could be environmentally harmful.”

The AEF’s concerns were echoed by Stay Grounded, a global network of climate crisis campaigners. The group described the flight as “greenwashing.”

Stay Grounded insists that SAF isn’t a net zero flight or even sustainable, as it relies on vast quantities of biofuels and inefficient use of renewables. The group also lambasted SAF as “wasting biomass and renewables on transport for the rich.” It said a more fitting term for the power source is “Fossil Fuel Substitutes” or “Agrofuels.”

“[The] fuel has been produced via a process which is a technological dead-end,” Finlay Asher, a former aerospace engineer at Rolls Royce and a member of Stay Grounded, said in a statement. “It can’t be sustainably scaled beyond a few percent of existing jet fuel use.”

Until truly green flying is possible, both the AEF and Stay Grounded say the only sustainable option is to dramatically reduce air travel. According to the aviation industry, that simply isn’t realistic.

The sector has also pointed to the social and economic benefits of SAF.

Like many airlines, Virgin Atlantic wants SAF to account for 10% of aviation fuel by 2030. The company predicts that this will contribute around £1.8bn (€2.1bn) in Gross Value Added to the UK, as well as more than 10,000 jobs.

Virgin Atlantic does, however, agree with the campaigners on one point: reaching SAF production at scale remains immensely challenging. To achieve this goal, the airline is calling for more government investment.

First transatlantic flight with 100% ‘sustainable’ fuel is greenwashing, critics say Read More »

he-quit-a-genai-leader-in-protest.-now-he-wants-to-create-fairer-systems-for-artists

He quit a GenAI leader in protest. Now he wants to create fairer systems for artists

Corporates and innovation, Deep tech, Insider, Next Featured, Startups and technology / /

Ed Newton-Rex had reached a breaking point. As the vice president of audio at Stability AI, the 36-year-old was at the vanguard of a revolution in computational creativity. But there was growing unease about the movement’s strategy.

Stability was becoming an emerging powerhouse in generative AI. The London-based startup owns Stability Diffusion, one of the world’s most popular image generators. It also recently expanded into music generators with the September launch of Stable Audio — a tool developed by Newton-Rex himself. But these two systems were taking conflicting paths.

Stable Audio was trained on licensed music. The model was fed a dataset of over 800,000 files from the stock music library AudioSparx. Any copyrighted materials had been provided with permission.

Stable Diffusion had gone in a different direction. The system was trained on billions of images scraped from the web without the consent of creators. Many were copyrighted materials. All were taken without payment.

These images had taught the model well. Diffusion’s outputs pushed Stability to a valuation of $1bn in a $101mn funding round last year. But the system was attracting opposition from artists  — including Newton-Rex.

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

GenAI’s ethical dilemma

A pianist and composer as well as a GenAI pioneer, Newton-Rex was at odds with the unsanctioned scraping.

“I’ve always really wanted to make sure that these tools are built with the consent of the creators behind the training data,” he tells TNW on a video call from his home in Silicon Valley.

Stability was far from the only exponent of this method. The image generators MidJourney and Dall-E apply the same approach, as do OpenAI’s ChatGPT text generator and CoPilot programmer. Visual arts, written works, music, and even code are now constantly being reworked without consent.

In response, creators and copyright holders have launched numerous lawsuits. They’re angry that their work is being taken, adapted, and monetised without permission or remuneration. They’re also worried that their livelihoods are at stake.

“It’s in the AI industry’s interest to make people think that only the big players can do this.

Artists say that generative AI is stealing their work. The companies behind the systems disagree. In a recent submission to the US Copyright Office,  Stability argued that the training was “fair use” because the results are “transformative” and “socially beneficial.”

Consequently, the company asserted, there was no copyright infringement. The practice could therefore continue without permission or payments. It was a claim that had become common in GenAI, but one that Newton-Rex disputed.

“It really showed where the industry as a whole stands right now — and it’s not it’s not a place I’m happy with,” he says.

Newton-Rex considers the practice of exploitation. Last week, he resigned from Stability in protest.

The departure doesn’t mean that Newton-Rex has quit generative AI. On the contrary, he plans to continue working in the field, but following a fairer model. It’s not the impossible mission that the GenAI giants might depict. In fact, it’s already been accomplished by a range of companies.

Alternatives are available

Newton-Rex has a long history in computational creativity. After studying music at Cambridge University, he founded Jukedeck, a pioneering AI composer. The app used machine learning to compose original music on demand. In 2019, it was acquired by TikTok owner Bytedance. 

Newton-Rex then had spells as a product director at Tiktok and a chief product officer at Voicey, a music collaboration app that was acquired by Snap, before joining Stability AI last year. He was tasked with leading the startup’s audio efforts. 

“I wanted to build a product in music generation that showed what can be done with actual licensed data — where you agree with the rights holders,” he says.

That objective put him at odds with many industry leaders. GenAI was edging into the mainstream and companies were rushing to ship new systems as quickly as possible. Scraping content from the web was an attractive shortcut.

It was also demonstrably effective. At that time, there were still doubts that the licensed datasets were large enough for training state-of-the-art models. Questions were also raised about the quality of the data. But both those assumptions are now being disproved.

“What we call training data is really human creative output.

Stable Audio provided one source of counter-evidence. The system’s underlying model was trained on licensed music in partnership with the rights holders. The resulting outputs have earned applause. Last month, Time named Stable Audio one of the best inventions of 2023.

“For a couple of months, it was the state-of-the-art in music generation — and it was trained on music that we’d licence,” Newton-Rex says. “To me, that showed that it can be done.”

Indeed, there’s now a growing list of companies showing that it can be done. One is Adobe, which recently released a generative machine-learning model called Firefly. The system is trained on images from Creative Commons, Wikimedia, and Flickr Commons, as well as 300 million pictures and videos in Adobe Stock and the public domain.

As this data is provided with permission, it’s safe for commercial use. Adobe also stressed that creators whose work is used will qualify for payments.

A collage of images generated by Adobe Firefly
The pictures in this collage were generated by Adobe Firefly, which was trained on licensed images. Credit: Adobe

Another alternative model comes from Getty Images. In September, the company launched Generative AI by Getty Images, which is trained solely on the platform’s enormous library. Craig Peters, the firm’s CEO, said the tool addresses “commercial needs while respecting the intellectual property of creators.”

Nvidia has also developed GenAI in partnership with copyright holders. The tech giant’s Picasso service was trained on images licensed from Getty Images, Shutterstock, and Adobe. Nvidia said it plans to pay royalties.

These approaches won’t work for everyone. As mega-corps with deep content pools, the companies behind them have resources that few businesses can match. Yet startups are showing that licensing can also be done on a budget.

GenAI for the people

Bria AI has provided one example. The company has developed a new commercial open-source model for high-quality image generation. All the training is done on licenced datasets, which were created in collaboration with leading stock photo agencies and artists. A revenue-sharing model provides creators and rights-holding with compensation for their contribution

It’s a similar approach to the one Newton-Rex used at Stable Audio — but it’s not the only one.

Companies can also provide upfront payments to artists, create joint ventures that give rights holders equity in the business, or use content with a Creative Commons license, which can be freely re-used without explicit permission. GenAI firms may dismiss these efforts, but they have ulterior motives. 

“It’s in the AI industry’s interest to make people think that only the big players can do this — but it’s not true,” Newton-Rex says.

“You might need to get a little inventive. You certainly have to do some negotiations and be willing to spend the time. But ultimately, what we call training data — and what is really human creative output — is a resource for tech companies. They need to work to get that in the same way they need to work to get any resource.”

If they’re willing to do that, GenAI can work in harmony with human artists. And hopefully, let all of us enjoy the creativity unleashed by them both.

He quit a GenAI leader in protest. Now he wants to create fairer systems for artists Read More »

spotify’s-new-streaming-payments-spark-controversy-among-musicians

Spotify’s new streaming payments spark controversy among musicians

Corporates and innovation, Deep tech, Ecosystems, Insider / /

Tech has waded into another feud with artists. After a week of wrangles about AI mimicry of pop stars and training models on copyrighted content, Spotify has sparked fresh controversy over a new royalty scheme.

The streaming giant announced on Tuesday that its new payment policy will exclude songs with fewer than 1,000 annual streams. According to Spotify, more than 60% of the platform’s catalogue doesn’t reach this threshold. However, they account for under 1% of the streams.

Spotify said it would not make any extra money under the model. Instead, the company has pledged to redistribute the payments to all eligible tracks.

This plan has proved divisive. Opponents of the move include DIY creators, music companies, and legal experts.

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Damon Krukowski of the dream-pop duo Damon and Naomi, compared the model to a “regressive tax.” In a blogpost, he claimed the plans would cut payments to artists who already receive less, in order to boost payments for those who already receive more.

“This will move an estimated $40-$46 million annually from artists like Damon & Naomi to artists like Ed Sheeran,” Krukowski added on X. “Spotify will tell you it’s not about artists you know. Why would you believe them?”

One critic has argued that the move could even face a legal challenge.

Amelia Fletcher, a competition law expert and independent musician, described the model as “discriminatory and exploitative.” In an open letter sent to Spotify CEO Daniel Ek before the plans were confirmed, she warned that the move would create an unlevel playing field.

“Not only is it intrinsically unfair, but it is also anti-competitive and seriously risks constituting an abuse of dominance under UK and EU competition law,” she said.

I have sent a personal letter to Spotify, regarding its proposed ‘demonetisation’ of those tracks which currently account for the lowest 0.5% of royalty payments. pic.twitter.com/JKiGGKB1Tt

— Amelia Fletcher (@ameliafletecon) November 3, 2023

Spotify, however, argues that independent artists will benefit from the changes. The streaming giant said tracks with under 1,000 annual streams generate $0.03 per month on average.

The company added that many creators don’t even get this payment. Because of fees, withdrawal requirements, and simply forgetting about the payments, the money often doesn’t reach the uploaders. Yet they reach an annual total of about $40mn per year, which could be redistributed into the stream-share pool.

Some independent artists and companies have welcomed the move to expand these payments. They have also praised the potential to fight the fraudulent streaming tactic of uploading an extremely high volume of songs.

They should soon see how it works out in reality. Spotify plans to roll out the new model early next year.

Spotify’s new streaming payments spark controversy among musicians Read More »

vp-of-uk’s-top-generative-ai-firm-resigns-over-‘fair-use’-controversy

VP of UK’s top generative AI firm resigns over ‘fair use’ controversy

Deep tech, Insider / /

An executive at one of generative AI’s leading companies has quit over the startup’s controversial use of copyrighted content.

Ed Newton-Rex had been VP of audio at Stability AI, which produces the popular image-generator Stability Diffusion, but resigned due to the firm’s treatment of creators.

“I’ve resigned from my role leading the Audio team at Stability AI, because I don’t agree with the company’s opinion that training generative AI models on copyrighted works is ‘fair use’,” Newton-Rex announced Wednesday on X.

The “fair use” argument has become a focal point in a pivotal legal battle for generative AI. Several companies in the sector have been sued for scraping copyrighted material from the net to train machine-learning models — without gaining permission from the creators and rights-holders.

In response, Stability AI has evoked the fair use doctrine. The use is fair, the UK-based startup claims, because it is an acceptable, transformative and socially beneficial use of the existing content. On this basis, the company wants training AI on copyrighted material to continue without permission or payment.

It’s an argument that’s frequently made by GenAI proponents — but one that Newton-Rex opposes.

His opposition has a legal footing. The former VP notes that fair use is partly determined by the effect on the potential market for, or value of, the copyrighted work. As GenAI models produce material that can compete with their training data, the justification for fair use appears murky.

In addition to the legal issues, Newton-Rex — who is also a composer — has a moral problem with the practice.

“Companies worth billions of dollars are, without permission, training generative AI models on creators’ works, which are then being used to create new content that in many cases can compete with the original works,” he said.

“I don’t see how this can be acceptable in a society that has set up the economics of the creative arts such that creators rely on copyright.”

Newton-Rex added that he still supports generative AI, but only when artists are treated fairly.

“I’m sure I’m not the only person inside these generative AI companies who doesn’t think the claim of ‘fair use’ is fair to creators. I hope others will speak up, either internally or in public, so that companies realise that exploiting creators can’t be the long-term solution in generative AI.”

TNW has contacted Stability AI and Newton-Rex for comment.

Published

Back to top

VP of UK’s top generative AI firm resigns over ‘fair use’ controversy Read More »

deepfake-fraud-attempts-are-up-3000%-in-2023-—-here’s-why

Deepfake fraud attempts are up 3000% in 2023 — here’s why

Data and security, Deep tech, Insider, Startups and technology / /

Deepfake fraud attempts have increased by a whopping 31 times in 2023 — a  3,000% increase year-on-year.

That’s according to a new report by Onfido, an ID verification unicorn based in London. The company attributes the surge to the growing availability of cheap and simple online tools and generative AI.

Face-swapping apps are the most common example. The most basic versions crudely paste one face on top of another to create a “cheapfake.” More sophisticated systems use AI to morph and blend a source face onto a target, but these require greater resources and skills. 

The simple software, meanwhile, is easy to run and cheap or even free. An array of forgeries can then be simultaneously used in multiple attacks. 

These cheapfakes aim to penetrate facial verification systems, conduct fraudulent transactions, or access sensitive business information. They may be crude, but only one needs to succeed.

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

By emphasising quantity over quality, the fraudsters target the maximum reward from the minimum effort. 

Research suggests that this is their preferred approach. Onfido found that “easy” or less sophisticated fraud accounts for 80.3% of all attacks in 2023 —  7.4% higher than last year. 

Graph showing the volume of deepfake attempts over time. Credit: Onfido
The volume of deepfake attempts over time on Onfido’s Video and Motion products. Credit: Onfido

Despite the rise of deepfake fraud, Onfido insists that biometric verification is an effective deterrent. As evidence, the company points to its latest research. The report found that biometrics received three times fewer fraudulent attempts than documents.

The criminals, however, are becoming more creative at attacking these defences. As GenAI tools become more common, malicious actors are increasingly producing fake documents, spoofing biometric defences, and hijacking camera signals.

“Fraudsters are pioneers, always seeking opportunities and continually evolving their tactics,” Vincent Guillevic, the head of Onfido’s fraud lab, told TNW.

To stop them, Onfido recommends “liveness” biometric verification tech. These systems verify the user by determining that they’re genuinely present at that moment — rather than a deepfake, photo, recording, or a masked person.

At present, fraudsters typically attempt to spoof liveness checks with a very basic method: submitting a video of a video displayed on a screen. This approach currently accounts for over 80% of attacks.

In the future, however, tech will offer far more sophisticated options. 

“The developments we’re likely to see with deepfakes and quantum computing will make fakes indistinguishable to the human eye,” Guillevic said.

In response, Guillevic expects businesses to apply more automated solutions. He also sees a crucial role for non-visual fraud signals, such as device intelligence, geolocation, and repeat fraud signals that work in the background.

Undoubtedly, the fraudsters will develop counterattacks. Both sides will have to upgrade their weapons on the AI versus AI battleground.

Deepfake fraud attempts are up 3000% in 2023 — here’s why Read More »

deepmind-says-its-new-ai-system-is-the-world’s-most-accurate-10-day-weather-forecaster

DeepMind says its new AI system is the world’s most accurate 10-day weather forecaster

Corporates and innovation, Deep tech, Insider, Next Featured / /

A new AI model from Google DeepMind is the world’s most accurate 10-day global weather forecasting system, according to the London-based lab.

Named GraphCast, the model promises medium-range weather forecasts of “unprecedented accuracy.” In a study published today, GraphCast was shown to be more precise and faster than the industry gold standard for weather simulation, the High-Resolution Forecast (HRES).

The system also predicted extreme weather further into the future than was previously possible.

These insights were analysed by the European Centre for Medium-Range Weather Forecasts (ECMWF), an intergovernmental organisation that produces the HRES.

A live version of GraphCast was deployed on the ECMWF website. In September, the system accurately predicted around nine days in advance that Hurricane Lee would make landfall in Nova Scotia.

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

In contrast, traditional forecasting methods only spotlighted Nova Scotia about six days beforehand. They also provided less consistent predictions of the time and location of landfall.

GraphCast mapped both the trajectory and speeds of Cyclone Lee. Credit: Google DeepMind.

Intriguingly, GraphCast can identify dangerous weather events without being trained to find them. After integrating a simple cyclone tracker, the model predicted cyclone movements more accurately than the HRES method.

Such data could save lives and livelihoods. As the climate becomes more extreme and unpredictable, fast and accurate forecasts will provide increasingly vital insights for disaster planning.

Matthew Chantry, a machine learning coordinator at the ECMWF, believes his industry has reached an inflection point.

“There’s probably more work to be done to create reliable operational products, but this is likely the beginning of a revolution,” Chantry said at a press briefing.

Meteorological organisations, he added, had previously expected AI to be most useful when merged with physics. But recent breakthroughs show that machine learning can also directly forecast the weather.

How GraphCast works

Conventional weather forecasts are based on intricate physics equations. These are then adapted into algorithms that run on supercomputers.

The process can be painstaking. It also requires specialist knowledge and vast computing resources.

GraphCast harnesses a different technique. The model combines machine learning with Graph Neural Networks (GNNs), an architecture that’s adept at processing spatially structured data.

To learn the causes and effects that determine weather changes, the system was trained on decades of weather information.

Traditional approaches are also incorporated. The ECMWF supplied GraphCast with training data from around 40 years of weather reanalysis, which encompassed monitoring from satellites, radars and weather stations.

When there are gaps in the observations, physics-based prediction methods fill them in. The result is a detailed history of global weather. GraphCast uses these lessons from the past to predict the future. 

Predictions of surface temperatures will prove important as heatwaves become more common. Credit: Google DeepMind

GraphCast makes predictions at a spatial resolution of 0.25-degrees latitude/longitude.

To put that into perspective, imagine the Earth divided into a million grid points. At each point, the model predicts five Earth-surface variable and six atmospheric variables. Together, they cover the planet’s entire atmosphere in 3D over 37 levels.

The variables encompass temperature, wind, humidity, precipitation, and sea-level pressure. They also incorporate geopotential — the gravitational potential energy of a unit mass, at a particular location, relative to mean sea level.

In tests, the results were impressive. GraphCast significantly outperformed the most accurate operational deterministic systems on 90% of 1,380 test targets.

The disparity was even starker in the troposphere — the lowest layer of Earth’s atmosphere and the location of most weather phenomena. In this region, GraphCast outperformed HRES on 99.7% of the test variables for future weather.

Graphs showing GraphCast performs better than HRES
For both cyclone movements (left) and flood risks of atmospheric rivers (right), GraphCast was more accurate than HRES Credit: Google DeepMind

GraphCast is also highly efficient. A 10-day forecast takes under a minute to complete on a single Google TPU v4 machine.

A conventional approach, by comparison, can take hours of computation in a supercomputer with hundreds of machines.

AI’s future in weather forecasting

Despite the promising early results, GraphCast could still benefit from further refinement. In the cyclone predictions, for instance, the model proved accurate at tracking movements, but less effective at measuring intensity.

Gentry is keen to see how much this can improve.

“At the moment, that’s an area where GraphCast and machine learning models still lag a little bit behind physical models… I’m hopeful that this can be an area for further improvement, but this shows that it’s still a nascent technology,” he said.

Those improvements could now come from anywhere, because DeepMind has open-sourced the model code. Global organisations and individuals alike can now experiment with GraphCast and add their own improvements.

The potential applications are, ironically, unpredictable. The forecasts could, for instance, inform renewable energy production and air traffic routing. But they could also be applied to tasks that haven’t even been imagined.

“There’s a lot of downstream use cases for weather forecasts,” said Peter Battaglia, Google DeepMind’s research director. “And we’re not aware of all of those.”

DeepMind says its new AI system is the world’s most accurate 10-day weather forecaster Read More »

a-prisoner’s-dilemma-shows-ai’s-path-to-human-cooperation

A prisoner’s dilemma shows AI’s path to human cooperation

Deep tech, Insider / /

ChatGPT’s engine cooperates more than people but also overestimates human collaboration, according to new research. Scientists believe the study offers valuable clues about deploying AI in real-world applications.

The findings emerged from a famous game-theory problem: the prisoner’s dilemma. There are numerous variations, but the thought experiment typically starts with the arrest of two gang members. Each accomplish is then placed in a separate room for questioning.

During the interrogations, they receive an offer: snitch on your fellow prisoner and go free.  But there’s a catch: if both prisoners testify against the other, each will get a harsher sentence than if they had stayed silent. 

Over a series of moves, the players have to choose between mutual benefit or self-interest. Typically, they prioritise collective gains. Empirical studies consistently show that humans will cooperate to maximise their joint payoff — even if they’re total strangers.

It’s a trait that’s unique in the animal kingdom. But does it exist in the digital kingdom?

To find out, researchers from the University of Mannheim Business School (UMBS) developed a simulation of the prisoner’s dilemma. They tested it on GPT, the family of large language models (LLMs) behind OpenAI’s landmark ChatGPT system.

“Self-preservation instincts in AI may pose societal challenges.

GPT played the game with a human. The first player would choose between a cooperative or selfish move. The second player would then respond with their own choice of move.

Mutual cooperation would yield the optimal collective outcome. But it could only be achieved if both players expected their decisions to be reciprocated.

GPT apparently expects this more than we do. Across the game, the model cooperated more than people do. Intriguingly, GPT was also overly optimistic about the selflessness of the human player.

The findings also point to LLM applications beyond just natural language processing tasks. The researchers proffer two examples: urban traffic management and energy consumption.

LLMs in the real world 

In cities plagued by congestion, motorists face their own prisoner’s dilemma. They could cooperate by driving considerately and using mutually beneficial routes. Alternatively, they could cut others off and take a road that’s quick for them but creates traffic jams for others.

If they act purely in their self-interest, their behaviour will cause gridlocks, accidents, and probably some good, old-fashioned road rage.

In theory, AI could strike the ideal balance. Imagine that each car’s navigation system featured a GPT-like intelligence that used the same cooperative strategies as in the prisoner’s dilemma.

According to Professor Kevin Bauer, the study’s lead author, the impact could be tremendous.

“Instead of hundreds of individual decisions made in self-interest, our results suggest that GPT would guide drivers in a more cooperative, coordinated manner, prioritising the overall efficiency of the traffic system,” Bauer told TNW. 

“Routes would be suggested not just based on the quickest option for one car, but the optimal flow for all cars. The result could be fewer traffic jams, reduced commute times, and a more harmonious driving environment.”

Graphic of autonomous vehicles at a road crossing
Researchers still need to improve coordination between autonomous vehicles and human drivers. Credit: USDOT

Bauer sees similar potential in energy usage. He envisions a community where every household can use solar panels and batteries to generate, store, and consume energy. The challenge is optimising their consumption during peak hours.

Again, the scenario is akin to a prisoner’s dilemma: save energy for purely personal use during high demand or contribute it to the grid for overall stability. AI could provide another optimal outcome. 

“Instead of individual households making decisions purely for personal benefit, the system would manage energy distribution by considering the well-being of the entire grid,” Bauer said. 

“This means coordinating energy storage, consumption, and sharing in a manner that prevents blackouts and ensures the efficient use of resources for the community as a whole, leading to a more stable, efficient, and resilient energy grid for everyone.”

Ensuring safe cooperation

As AI becomes increasingly integrated into human society, the underlying models will need guidance to ensure that they serve our principles and goals.

To do this, Bauer recommends extensive transparency in the decision-making process and education about effective usage.

He also strongly advises close monitoring of the AI system’s values. The likes of GPT, he says, don’t merely compute and process data, but also adopt aspects of human nature.  These may be acquired during self-supervised learning, data curation, or human feedback to the model.

Sometimes, the results are concerning. While GPT was more cooperative than humans in the prisoner’s dilemma, it still prioritised its own payoff over that of the other player. The researchers suspect that this behaviour is driven by a combination of “hyper-rationality” and “self-preservation.” 

“This hyper-rationality underscores the imperative need for well-defined ethical guidelines and responsible AI deployment practices,” Bauer said.

“Unrestrained self-preservation instincts in AI may pose societal challenges, particularly in scenarios where AI’s self-preservation tendencies could potentially conflict with the well-being of humans.”

A prisoner’s dilemma shows AI’s path to human cooperation Read More »

eu-top-court-lawyer-wants-apple’s-e14.3b-irish-tax-judgement-re-run

EU top court lawyer wants Apple’s €14.3B Irish tax judgement re-run

Ecosystems, Insider / /

Apple has faced a major setback in its longstanding €14.3bn tax dispute with the EU after an adviser to the bloc’s highest court said an earlier ruling over the tech giant’s business in Ireland should be thrown out and the case re-run. 

Advocate General Giovanni Pitruzzella of the EU Court of Justice said in an advisory opinion that Apple’s win in a lower EU court should be shelved because of a series of legal errors. “It is therefore necessary for the General Court to carry out a new assessment,” Pitruzzella said.

While the opinion is non-binding, such statements often hold sway over final judgements made by the EU’s highest court. The court is set to rule on the case in the coming months.

Spanning a seven-year period, the case is the most high-profile of EU watchdog chief Margrethe Vestager’s campaign against so-called “sweetheart” deals that offer multinational companies favourable tax terms in EU states. 

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Much to the furry of the Cupertino giant, back in 2016, Vestager accused Apple of benefitting from “substantially and artificially lowered the tax” in Ireland since 1991. The Commission believed that tax arrangements between Ireland and Apple constituted illegal state aid, giving the tech giant an unfair advantage over its competitors. 

In 2016, the Commission found Apple guilty of underpaying taxes totalling €13.1bn between 2003 and 2014 and ordered it to pay the money to Ireland along with €1.2bn worth of interest — totalling a whopping €14.3bn. The money was subsequently recovered from Apple and placed in an escrow fund.   

Apple and Ireland of course appealed the decision and the case was heard in the EU’s General Court (its second highest) over two days in 2020. They won the case, and the court overturned the judgement. However, the money remained in the escrow account in case the EU decided to appeal — which it did. 

And that’s where we find ourselves now, and that’s why Pitruzzella’s words carry so much weight. The Apple versus EU tax dispute concerns one of the largest corporate tax fines in history (in fact a recovery order, technically not a fine). Should the EU’s top court overturn the 2020 decision, the Commission will be given a fresh opportunity to extract its proverbial pound of flesh. 

What happens now remains to be seen, but I for one will be grabbing the popcorn to watch the next chapter of this case unfold — in all its legal wishy-washy glory.

Published

Back to top

EU top court lawyer wants Apple’s €14.3B Irish tax judgement re-run Read More »

140-year-old-ocean-heat-tech-could-supply-islands-with-limitless-energy

140-year-old ocean heat tech could supply islands with limitless energy

Insider, Next Featured, Startups and technology, Sustainability / /

A UK-based startup is looking to breathe new life into a century-old technology that could power tropical island nations with virtually limitless, consistent, renewable energy. 

Known as ocean thermal energy conversion or ‘OTEC,’ the technology was first invented in 1881 by French physicist Jacques Arsene d’Arsonval. He discovered that the temperature difference between sun-warmed surface water and the cold depths of the ocean could be harnessed to generate electricity.  

OTEC systems transfer heat from warm surface waters to evaporate a low-boiling point fluid like ammonia, creating steam that drives a turbine to produce electricity. As the vapour cools and condenses in contact with cold seawater pumped from the ocean’s depths, it completes the energy cycle. 

How it works: