⤴⤵ Up Wing/Down Wing #38

A curated selection of pro-progress and anti-progress news items from the week that was

In case you missed it .. .

✨ A twist on the notion that big ideas are getting harder to find (Monday)

⚛️ Our nuclear problem: A Quick Q&A with … economist Eli Dourado on NRC overreach (Tuesday)

🧬 MAGA vs. the transhumanists of Silicon Valley (Wednesday)

⤴ Up Wing politics: A chat with sociologist Steve Fuller (Throwback Thursday)

(Thursday)

✨🤖 What jobs today are really most AI vulnerable? (Friday)

---

Up Wing Things

⚛ World Bank reconsiders its ban on funding nuclear power. In a major policy shift, World Bank President Ajay Banga has proposed lifting the institution’s longstanding ban on funding nuclear energy projects. Speaking in Washington, Banga said the board is open to discussion, with a broader energy strategy expected by June 2026. The move comes as developing nations from the Philippines to Vietnam seek clean, reliable alternatives to coal. Nuclear energy, including emerging small modular reactors, could offer a transformative solution. With global electricity demand rising due to AI and data center expansion, the World Bank’s support for nuclear power could unlock vital low-carbon energy options for the world’s poorest nations. (Financial Post)

⚡ Researchers are finding smarter ways to power AI. As demand for electricity grows, driven by powerful AI models like ChatGPT, researchers and engineers are rising to the challenge. From smarter software to cutting-edge chips, innovations are making data centers dramatically more energy-efficient. One new method, developed by Jae-Won Chung, targets GPU coordination:

. . . Chung built a software tool called Perseus that identified the scope of the workloads assigned to each GPU in a cluster. Perseus takes the estimated time needed to complete the largest workload on a GPU running at full. It then estimates how much computation must be done on each of the remaining GPUs and determines what speed to run them so they finish at the same time.

While exact power costs can be hard to pinpoint, the momentum toward efficiency is strong—and with breakthroughs like photonic computing on the horizon, the future of AI looks bright and sustainable. (Ars)

🚀 Germany’s Spectrum rocket nears historic launch. Europe is on the brink of a space milestone as German startup Isar Aerospace prepares to launch Spectrum, the largest German rocket since the V-2. Spectrum is set to lift off from Norway’s Andøya Spaceport and marks Europe’s first commercial orbital launch from the continent. Powered by nine Aquila engines in the first stage and one engine in the second stage, the 92-foot-tall rocket can carry one ton to low-Earth orbit. Isar is backed by over €400 million in funding, including support from ESA and NATO, and is leading a new wave of private European launchers. Amid growing geopolitical uncertainty, the mission represents a significant move toward establishing independent and reliable European access to space. (Ars)

🎨 ChatGPT now creates complex, custom images. OpenAI has unveiled a powerful upgrade to ChatGPT, enabling the chatbot to generate detailed, imaginative images from even the most complex prompts. Built on the new GPT-4o model, the system now combines text and image generation, transforming a single description into a four-panel comic strip, surreal illustration, or creative scene. Unlike earlier versions, which separated image generation via DALL·E from ChatGPT, this unified model uses vast internet training data to respond to voice, text, image, and even video inputs. (NYT)

🏙️ Saudi Arabia plans the world’s tallest skyscraper. Saudi Arabia is setting its sights sky-high with a proposed 2-kilometer (1.2-mile) tall skyscraper that would become the tallest man-made structure ever built, towering over Dubai’s Burj Khalifa by more than double and reaching 3.5 times the height of the USA's tallest building, the One World Trade Center. Designed by Foster + Partners, the record-breaking building is planned as part of a futuristic business hub near Riyadh called the North Pole. The $5 billion project is in early stages, with firms recently invited to bid on project management. If completed, it would demonstrate Saudi Arabia’s focus on engineering and rapid modernization. (NA)

🌊 An ocean-friendly plastic is making waves. Japanese researchers at national research institute RIKEN have developed a promising new plastic that stays strong during use but dissolves harmlessly in saltwater. The strength of conventional plastics’ bonds are what makes them durable, but not without a cost:

That cup you used once and threw away will sit in landfill for decades, even centuries, before it fully breaks down. And when it does, it forms microplastic pieces that are turning up in all corners of the natural world, including our own bodies, where they wreak havoc on our health in ways we’re only just beginning to understand.

The material of the new Japanese plastic is made from supramolecular polymers and forms durable “salt bridges” that can be broken down when exposed to seawater — disappearing in under nine hours and leaving behind plant-friendly compounds. A hydrophobic coating prevents premature breakdown, and a simple scratch activates the process when disposal is desired. Unlike traditional plastics, which linger for centuries and create microplastics, this innovation could reduce pollution and even aid recycling efforts. (NA)

🏠 Upgraded windows could slash energy waste. A new generation of ultra-thin, high-performance windows may soon reshape home energy efficiency. Using the same technology behind smartphone screens, Corning has adapted its ultra-tough glass to create multi-pane windows that insulate better than most walls. These lightweight windows can reduce heating and cooling costs while meeting hurricane-level impact standards. With triple- and quad-pane designs, homes stay warmer in winter and cooler in summer—potentially saving Americans up to $25 billion in annual energy waste. If widely adopted, this innovation could transform residential construction and reduce household carbon emissions. (WSJ)

⚛️ Big tech bets on nuclear power. A clean energy revolution is underway as tech giants like Microsoft, Google, and Amazon invest heavily in nuclear power to fuel AI’s massive electricity demands. Microsoft is backing a $1.6 billion revival of Pennsylvania’s iconic Three Mile Island, now dubbed the Crane Clean Energy Center, aiming for zero-carbon power within three years. Meanwhile, Google is funding next-gen small modular reactors, with plans to bring the first online by 2030. Industry leaders say the key to speed and cost is repetition, building identical reactor designs one after another in a “cookie-cutter” approach that could finally make nuclear scalable and affordable. (CBS)

🛰️ Gravitics to launch space-based satellite depot. Gravitics, a Seattle-based startup, has secured a $60 million Space Force STRATFI award to develop an “orbital carrier,” acting as a satellite depot that could revolutionize tactically responsive space operations. The concept draws inspiration from naval aircraft carriers, offering protection and concealment for satellites already positioned in orbit:

Such a module would isolate the satellites from the space environment, sparing their batteries and sensitive electronics from harsh thermal cycles every 90 minutes, and provide some shielding from radiation. In addition, the orbital carrier would obfuscate the satellites inside from observation by other nations or hostile actors in space.

When needed, satellites can be rapidly deployed into multiple orbits. Gravitics is also advancing StarMax, a 7.6-meter-wide module with 400 cubic meters of pressurized volume. A demonstration mission for the orbital carrier could launch as early as 2026. (Ars)

🚗 Waymos are way better drivers than we are. In January, the first fatal crash involving a fully driverless vehicle occurred in San Francisco — but the Waymo car wasn’t at fault. It was stopped at a red light when a speeding SUV triggered a six-car pileup. Waymo’s broader safety record remains strong: Over more than 50 million driverless miles, Waymo vehicles have been involved in far fewer serious crashes than human drivers — 83 percent fewer airbag-deploying incidents and 81 percent fewer injury-causing crashes. Insurance data backs this up too, showing over 90 percent fewer successful injury claims. As driverless tech expands, the data continues to favor automation over human error. (Ars)

⛏️ AI unearths hidden mineral riches in Australia. Earth AI, a mining startup founded by Ukrainian-born geologist Roman Teslyuk, is reimagining mineral exploration with artificial intelligence. By training algorithms on decades of underused geological data from Australia’s national archives, the company has uncovered promising copper, cobalt, and gold deposits in the Northern Territory, as well as silver, tin, and molybdenum in New South Wales. To prove its tech, Earth AI built its own drilling hardware and raised $20 million in Series B funding. The company’s approach allows it to scan wide areas quickly, cutting mineral discovery timelines from decades to just a few years. (Tech Crunch)

💊 AI accelerates Alzheimer’s research. At the Oxford Drug Discovery Institute, artificial intelligence and knowledge graphs are transforming the search for Alzheimer’s treatments. Researchers recently used these tools to analyze 54 immune-related genes from a genome-wide study. Work that once took weeks is now completed in days:

The Oxford Drug Discovery Institute’s biologists worked with Graphwise to customize a large-scale knowledge graph of their life sciences research information. That process helped the institute’s biologists speed up the time it took to evaluate 54 genes from a few weeks to a few days, it said, and is helping them identify biomarkers that may be linked to the genes.

This AI-powered system not only accelerates target identification but also helps trace the origin of findings, boosting transparency. (WSJ)

---

---

On sale everywhere ⏩ The Conservative Futurist: How To Create the Sci-Fi World We Were Promised

---

---

Down Wing Things

🤖 Our current AI models probably won’t get us to AGI. A new survey of 475 AI experts finds growing skepticism that current large language models (LLMs) will ever achieve artificial general intelligence (AGI). About 76 percent say simply scaling today’s models — by adding more data, hardware, and money — is unlikely to get us there. Experts cite fundamental architectural limits and diminishing returns, with rising costs, energy use, and stagnating benchmarks. Despite the hype, many believe the industry has hit a wall. Still, promising paths remain, such as reasoning models, new designs like DeepSeek, and probabilistic programming that could offer smarter, more efficient ways forward. (Live Science)

🦠 The Trump administration has pulled funding for dozens of research projects aimed at developing vaccines and treatments for future pandemics in addition to Covid-19:

The government’s rationale is that the Covid pandemic has ended, which “provides cause to terminate Covid-related grant funds,” according to an internal N.I.H. document viewed by The New York Times.

But scientists say this abrupt halt undermines promising work on broad-spectrum antivirals and vaccines for emerging viruses like Ebola and Nipah virus. Researchers had identified compounds nearing clinical trials and made strides toward universal coronavirus vaccines. Now, with funding gone, much of that progress is at risk. Many in the scientific community are frustrated and concerned about the US stepping back from pandemic preparedness. (NYT)

💥 Global trade war reaches dangerous new heights. A wave of tariffs and import restrictions is sweeping the globe, triggering the most rapid rise in protectionism since the 1930s. Trump’s second-term tariffs, including 25 percent on metals like steel and aluminum, 20 percent on all Chinese imports, and new levies on Mexico and Canada, have unleashed a cascade of retaliatory measures from the EU, China, and beyond. As the WTO fades in influence, countries like Vietnam, Russia, and South Korea are targeting Chinese goods, too. With US tariff rates back to 1946 levels, economists warn of lasting economic and diplomatic damage. A return to global trade cooperation now appears increasingly unlikely. (WSJ)

🤖 International trust in American AI is falling. At this year’s RightsCon in Taiwan, global digital rights groups addressed the US's diminished support during Trump’s second term. This change affects how American tech companies operate internationally, raising concerns over content moderation failures in non-English-speaking regions. Experts fear situations may worsen as reliance on LLMs increases:

If platforms begin to rely even more on LLMs for content moderation, this problem will likely get worse, says Marlena Wisniak, a human rights lawyer who focuses on AI governance at the European Center for Not-for-Profit Law. “The LLMs are moderated poorly, and the poorly moderated LLMs are then also used to moderate other content,” she tells me. “It’s so circular, and the errors just keep repeating and amplifying.”

Momentum is growing in Europe to develop localized AI and digital infrastructure. There is a push for "AI sovereignty” and culturally-aware language models to lessen reliance on US tech and create inclusive, community-driven systems. (MIT)

📉 Trump’s golden age turns to economic whiplash. What began as Wall Street excitement over Trump’s promise of a “golden age” has rapidly turned to anxiety. The S&P 500 shed $4 trillion in value amid erratic tariff policies and sagging consumer sentiment. CEOs who once welcomed tax cuts are now bracing for April’s “liberation day” tariffs on key allies, while inflation worries mount and dealmaking stalls. Fed Chair Jerome Powell cited “tariff inflation” as a new concern, and business investment is slumping. Executives describe a “sense of helplessness” as unpredictability derails hiring and growth. If this continues, recession risks could become a self-fulfilling prophecy. (WSJ)

🧠 Q-Day looms as quantum threat grows fast. Experts warn of “Q-Day,” when quantum computers could break current encryption, exposing emails, financial records, military secrets, and Bitcoin wallets. Though theoretical, analysts estimate a one-in-three chance it arrives before 2035, with some believing it may already be unfolding in secret:

From a typical person’s vantage point, maybe Q-Day wouldn’t be recognizable as Q-Day at all. Maybe it would look like a series of strange and apparently unconnected news stories spread out over months or years. London’s energy grid goes down on election day

Nations like China and the US are racing to develop quantum systems, and hackers are already harvesting encrypted data to decrypt later. If encryption fails, protecting the digital infrastructure for authentication and critical systems becomes impossible. (Wired)