美國對石油的依賴遠低於以往

Recently the New York Times reported the following:

America Depends Less on Oil Than Ever

Gasoline still drives household budgets, but energy efficiency and renewables have reduced the economy’s overall reliance on petroleum.

The New York Times - By Lydia DePillis, Karl Russell contributed reporting.

March 14, 2026, 5:02 a.m. ET

War with Iran has frozen commerce in the Persian Gulf and boosted oil prices by more than 50 percent worldwide, translating almost immediately into higher gasoline costs. It’s the largest global oil disruption ever and is likely to accelerate inflation throughout this year.

And yet, in the United States, the impact is much more muted than it would have been a few decades ago.

That’s in part because America uses less energy per unit of economic output than it used to. In economist-speak, the U.S. economy is less “energy intensive,” for a few reasons.

One, the U.S. economy now depends largely on services like health care, retail and entertainment, which require much less energy than manufacturing industries. There are only about 21 million jobs in goods-producing sectors, while private services employ 114 million people.

And two, the machines that Americans do use are now much more efficient, a trend that started in earnest after the oil price shocks of the 1970s. According to the Department of Transportation, the average new light-duty vehicle gets 28 miles per gallon of gas, up from 13 in 1975. Gasoline consumption rose until 2007, then leveled off as electric vehicles gained traction. As a result, consumer spending on gas as a share of discretionary income has fallen.

Economists at Wells Fargo estimate that a sustained 50 percent rise in oil prices — similar to the current situation — would have had about twice the effect in the 1980s as it would today, when it’s expected to trim about one percentage point from annual consumer spending growth.

The United States has also become the world’s largest oil and gas producer. Rather than depend on supply from the Middle East, the rest of the world now consumes petroleum products that are fracked from North Dakota and West Texas.

All that new supply helped bring down prices globally in the 2010s, especially after Congress lifted a ban on natural gas exports in 2015. Theoretically, it means that profits from oil production stay in the United States, and can be redeployed for other investment. According to research from the Dallas and Kansas Federal Reserve branches, the shale boom added 1 percent to gross domestic product.

However, it is not clear that U.S. drillers are inclined to play the role of “swing producer” again this time. Price competition during the fracking boom was ruinous for investors. Many companies went bankrupt, unable to pay back all the money they had borrowed for expensive extraction infrastructure. They learned a lesson: Don’t invest a lot of money in pumping more just because prices rise. Especially if you’re not sure whether those prices will sink back down.

“I don’t see much happening from the production side in order to mitigate the effects for the U.S. economy,” said Christiane Baumeister, an economics professor at the University of Notre Dame who studies oil markets. “Companies just prioritize delivering returns to shareholders,” she said. “I think they would rather take advantage of the current situation to increase profits rather than investing that back into expanding output.”

Another deterrent to ramping up production: Steel and aluminum tariffs have raised the cost of the pipes and valves needed for it. The number of oil rigs that are actively pumping in the United States is down 7 percent from this time last year.

Even with the U.S. oil industry operating at full steam, it doesn’t generate many benefits for U.S. workers, as oil companies have learned how to operate with fewer people. The United States is pumping more oil and gas than ever, but the extraction, drilling and oil field services industries have been shedding jobs. The sector employs about 363,000 people, which is about 0.2 percent of all employment.

The Shale Jobs Boom Is Over

And despite the huge boom in U.S. oil production over the past 15 years, it has not become a more meaningful part of Americans’ stock portfolios. Exxon Mobil and Chevron had long been among the most valuable companies in the stock market. But the entire oil and gas sector now makes up just 3.2 percent of the S&P 500 index, down from 5.5 percent a decade ago. Their stocks had consistently underperformed the broader index, until soaring oil prices lifted their fortunes after the United States and Israel attacked Iran.

“The oil and gas industry’s financial strategy has been ‘pray for war,’ because those are the conditions under which they make money,” said Clark Williams-Derry, an oil industry financial analyst at the Institute for Energy Economics and Financial Analysis. “They have to have a big price spike every few years literally to make ends meet.”

The Americans most vulnerable to an oil-price shock are people with lower incomes who haven’t been able to buy their way out of gasoline dependence. Electric vehicles are more expensive than gas-powered cars, and their owners are disproportionately affluent and well educated, according to a recent paper by researchers at Carnegie Mellon University. As the Trump administration rolls back energy efficiency standards for appliances and fuel economy rules for cars, oil price increases could start to have a bigger impact.

Lower-Income Americans Spend More on Fuel

Electricity bills are another substantial cost, absorbing 3.6 percent of budgets for households in the lowest 20 percent of earners, according to the Bureau of Labor Statistics. Although renewable energy sources have been supplying more of that electricity, fossil fuels still account for 60 percent.

“Talk to someone who doesn’t make a lot of money, and see if they’re resilient to oil and gas,” Mr. Williams-Derry said. “In terms of a variable expenditure that you can’t control, it’s up there.”

Translation

美國對石油的依賴遠低於以往

汽油仍然是家庭預算的主要支出，但能源效率和再生能源的發展降低了美國經濟整體地對石油的依賴

美國與伊朗的戰爭導致波斯灣地區的貿易停滯，並使全球油價上漲超過50%，幾乎立即推高了汽油價格。這是有史以​​來規模最大的全球石油供應中斷事件，並可能在今年加速通膨。

然而，在美國，戰爭的影響遠比幾十年前小得多。

部分原因是美國每單位經濟產出的能源消耗量比以往低。用經濟學家的話來說，由於以下幾個原因，美國經濟的「能源密集度」降低了。

首先，美國經濟如今主要依賴醫療保健、零售和娛樂等服務業，這些產業所需的能源遠低於製造業。商品生產業的就業人數僅約2,100萬個，而私人服務業則僱用了1.14億人。

其次，美國人使用的機器效率大幅提升，趨勢始於1970年代的石油危機之後。根據美國交通部統計，目前新型輕型車輛的平均油耗為每加侖汽油行駛28英里，高於1975年的13英里。汽油消耗量在2007年之前持續曾長，之後隨著電動車的普及而趨於平穩。因此，消費者在汽油上的支出佔可支配收入所得的比例下降。

富國銀行的經濟學家估計，如果油價持續上漲50% - 類似於目前的情況 - 其影響在1980年代將是現在的兩倍左右，現預計只會使年度消費者支出成長率下降約1個百分點。

美國也成為全球最大的石油和天然氣生產國。如今，世界其他地區不再依賴中東的供應，而是消費來自北達科他州和西德克薩斯州頁岩氣開採的石油產品。

所有這些新增供應在2010年代幫助全球油價下降，尤其是在國會於2015年解除天然氣出口禁令之後。理論上，這意味著石油生產的利潤留在美國，可以重新用於其他投資。根據達拉斯和堪薩斯聯邦儲備銀行的研究，頁岩氣繁榮使美國國內生產毛額增加了1%。

然而，目前尚不清楚美國鑽井公司是否願意再次扮演「搖擺生產商」的角色。頁岩氣繁榮時期的價格競爭給投資者帶來了毀滅性的打擊。許多公司破產，無力償還為昂貴的開採基礎設施而貸來的所有資金。他們吸取了一個教訓：不要因為價格上漲就投入大量資金增加產量。尤其是在不確定油價是否會回落的情況下。

聖母大學研究石油市場的經濟學教授 Christian Baumeister 表示:「我認為生產方面不會採取太多措施來緩解對美國經濟的影響」。她說: 「公司優先考慮的是為股東創造回報」; 「我認為他們寧願利用當前形勢增加利潤，而不是將利潤再投資於擴大產量」。

另一個阻礙增產的因素是：鋼鐵和鋁關稅提高了石油生產所需管線和閥門的成本。目前美國正在作業的鑽井平台數量比去年同期下降了7%。

即使美國石油業全力運作，也未能為美國工人帶來多少好處，因為石油公司已經學會如何以更少的人力去運作。美國的石油和天然氣產量比以往任何時候都高，但開採、鑽井和油田服務業卻一直在裁員。該行業僱用了約36.3萬人，約佔全部就業人數的0.2%。

頁岩油就業繁榮已經結束

儘管過去15年美國石油產量大幅成長，但它並未在美國人的股票投資組合中佔有更重要的地位。埃克森美孚和雪佛龍長期以來一直是股市中最有價值的公司之一。但如今，整個石油和天然氣產業在標普500指數中的佔比僅3.2%，低於十年前的5.5%。在美和以攻擊伊朗後，油價飆升才提振了它們的業績，此前它們的股票表現一直遜於大盤。

能源經濟與金融分析研究所的石油行業金融分析師Clark Williams-Derry表示:“石油和天然氣行業的財務策略一直是 ‘祈求戰爭’，因為只有在戰爭時期他們才能賺錢” ;“他們必須每隔幾年就經歷一次油價大幅上漲才能維持收支平衡。”

最容易受到油價衝擊的美國人是低收入者難以擺脫對汽油的依賴。卡內基美隆大學研究人員最近發表的一篇論文指出，電動車比汽油車更貴，而且電動車的車主往往更富裕、受過良好教育。隨著特朗普政府逐步取消家用電器能源效率標準和汽車燃油效率規定，油價上漲的影響可能會越來越大。

低收入美國人用較多錢在燃料上

電費是另一項重大開支，根據美國勞工統計局的數據顯示，電費佔收入最低的20%家庭預算的3.6%。儘管再生能源的電力供應佔比不斷提高，但化石燃料仍佔60%。

Williams-Derry先生說: 「和那些收入不高的人聊聊，看看他們是否能承受得住石油和天然氣的衝擊」; 「就你無法控制的可變動支出而言，這筆開支是相當可觀」。

        So, the war with Iran has boosted oil prices by more than 50 percent worldwide. But in the United States, the impact is much lighter than it would have been a few decades ago. In economic terms, the U.S. economy is less “energy intensive.” U.S. economy now depends largely on services like health care, retail and entertainment, which require much less energy than manufacturing industries. The Americans most vulnerable to an oil-price shock are those with lower incomes who haven’t been able to reduce their gasoline dependence.

科學家窺見新疫情的起源（2/2）

Recently the New York Times reported the following:

Scientists Get a Glimpse of How New Pandemics Are Made (2/2)

Researchers have devised a new tool for discerning between naturally occurring viral outbreaks and those resulting from lab accidents.

The NYT - By Carl Zimmer

March 9, 2026

Updated 9:43 a.m. ET

(continue)

Since then, researchers have not found direct evidence to test this scenario or others like it. But the new study by Dr. Wertheim and his colleagues concluded that the 1977 virus underwent some odd evolution before the pandemic — and that the mutations it gained bear patterns identical to those found in viruses that are grown in labs.

Gigi Gronvall, a biosecurity expert at Johns Hopkins University, said that the new study suggests that the 1977 pandemic began as a vaccine trial. “It’s more evidence that they were trying to create an attenuated vaccine and failed spectacularly,” she said.

Spyros Lytras, a virologist at the University of Tokyo who was not involved in the new study, said the study’s methodology offered a new tool for tracing the origin of outbreaks. “The genetic data holds a robust signal that can distinguish between lab passaging and manipulations versus natural spillovers,” he said.

James Lloyd-Smith, a disease ecologist at the University of California, Los Angeles, said that detecting the genetic signature of lab evolution in the Russian flu was “very cool.” But he cautioned that scientists typically only capture a fraction of the evolution of viruses in their samples. “This approach isn’t a magic wand, and the study still faces the same challenges of limited data that have dogged the field,” he said.

Of the several outbreaks that Dr. Wertheim and his colleagues analyzed, only the Russian flu proved to be an exception to the rule. The virus that caused Covid — SARS-CoV-2 — was not.

The researchers found no peculiar changes in SARS-CoV-2 before it jumped into humans. It gained mutations as it spread from bat to bat, just like other bat coronaviruses did; only after the virus emerged in humans did it undergo a marked shift. Within a year, radically new variants were evolving, with mutations that made them exquisitely well-adapted to humans.

The new study adds to the ongoing debate over Covid’s origins. In a January interview with The New York Times, Dr. Jay Bhattacharya, the director of the National Institutes of Health, asserted that SARS-CoV-2 emerged from a lab. “I think if you just focus on the scientific evidence alone, I would say it’s certain,” he said.

Some experts who favor a lab-leak origin point to just how well prepared SARS-CoV-2 was to spread among humans from the start of the pandemic. In November, Dr. Robert Redfield, former director of the Centers for Disease Control and Prevention, claimed that the Chinese military had created SARS-CoV-2 as an inhaled vaccine. “I think that was all engineered,” Dr. Redfield said in a podcast interview.

But last month, a group of experts assigned by the World Health Organization to examine Covid’s origins came down in favor of SARS-CoV-2 having originated in bats, which then passed it to animals sold in a market in Wuhan.

“Most of the peer-reviewed scientific evidence supports this hypothesis,” the scientists noted.

Dr. Wertheim has helped collect some of that peer-reviewed scientific evidence. The new study adds more evidence in favor of an animal origin, he said.

If SARS-CoV-2 was reared in a lab, its mutations would unfold in a pattern like that of the Russian flu. Instead, Dr. Wertheim and his colleagues found, its mutation pattern matched the five naturally occurring outbreaks they studied.

Instead, Dr. Wertheim said, Covid appears to have arisen from some really bad luck. As the precursor virus was adapting to infect bats, it ended up ready to start a pandemic among people. “It’s coincidentally exceptionally good at being a human virus,” Dr. Wertheim said.

David Robertson, a virologist at the University of Glasgow who was not involved in the new research, said that the study offered insights not just for Covid but for any zoonotic virus — a virus that spreads from animals to humans. “It’s a key point for understanding zoonotic risk,” he said. “Viruses can be circulating in nature without requiring adaptations to infect or transmit successfully in humans.”

If that’s the case across a wide range of zoonotic viruses, Dr. Wertheim said, we can expect more pandemics in the future. “It’s what we don’t know that’s going to get us,” he said. “They’re out there, and they’re ready to go.”

Translation

科學家窺見新疫情的起源（2/2）

研究人員開發出一種新工具，用於區分由自然發生的病毒爆發疫情和由實驗室事故導致的爆發

（繼續）

此後，研究人員尚未找到直接證據來驗證此假設或其他類似假設。但Wertheim博士及其同事的一項新研究得出結論：1977年的病毒在大流行之前經歷了某種奇特的進化 - 而且它獲得的突變模式與實驗室培養的病毒的突變模式完全相同。

約翰霍普金斯大學的生物安全專家Gigi Gronvall表示，這項新研究表明，1977年的疫情始於一次疫苗試驗。 她說:「這進一步證明，他們當時試圖研發減毒疫苗，結果卻慘敗了」。

並未參與這項新研究的東京大學病毒學家Spyros Lytras表示，該研究的方法為追蹤疫情起源提供了一種新工具。他說: 「基因數據包含一個可靠的訊號，可以區分實驗室傳代和人為操作，相對於自然溢出」。

加州大學洛杉磯分校疾病生態學家James Lloyd-Smith表示，在俄羅斯流感中檢測到實驗室進化的基因特徵是「非常酷」的事。但他提醒說，科學家通常只能捕捉到病毒樣本中演化過程的一小部分。他說: 「這種方法並非萬能，這項研究仍然面臨著數據有限的挑戰，而數據有限一直是該領域長期存在的難題」。

在Wertheim博士及其同事分析的幾宗疫情中，只有俄羅斯流感被證明是例外。導致新冠肺炎的病毒—SARS-CoV-2—則不是。

研究人員發現，SARS-CoV-2 在感染人類之前並沒有發生任何異常變化。它像其他蝙蝠冠狀病毒一樣，在蝙蝠間傳播的過程中不斷發生變異；只有在病毒出現在人類身上之後，才出現了顯著的變化。不到一年，全新的變種病毒就不斷湧現，這些變異使其能夠完美地適應人類。

這項新研究為新冠病毒起源的持續爭論增添了新的內容。美國國立衛生研究院院長Jay Bhattacharya博士在1月接受《紐約時報》採訪時斷言，SARS-CoV-2 源自於實驗室。他說: 「我認為，如果只看科學證據，我會說這是肯定的」。

一些支持實驗室洩漏起源論的專家指出，SARS-CoV-2 從疫情爆發之初就充份预備好在人群中傳播。去年11月，美國疾病管制與預防中心前主任Robert Redfield博士聲稱，中國軍方發展了SARS-CoV-2病毒作為吸入式疫苗。Redfield博士在一次播客採訪中說道: 「我認為這一切都是設計出來的」。

但上個月，世界衛生組織指派的一個專家小組對新冠病毒的起源進行了調查，最終得出結論：SARS-CoV-2病毒起源於蝙蝠，然後蝙蝠將其傳播給了武漢一家市場出售的動物。

科學家指出:「大多數經過同行評審的科學證據都支持這個假設」。

Wertheim博士曾參與那個收集一部分曾經過同儕審查的科學證據。他表示，這項新研究為病毒起源於動物的觀點提供了更多證據。

如果SARS-CoV-2病毒是在實驗室中培育的，那麼它的變異模式應該類似於俄羅斯流感病毒。然而，Wertheim博士及其同事發現，它的變異模式與他們研究的五宗自然發生的疫情相吻合。

Wertheim博士表示，新冠病毒的出現似乎純粹是運氣不佳。其前身病毒在適應感染蝙蝠的過程中，最終卻意外地具備了在人類中引發大流行的能力。 Wertheim博士說: 「它巧合地非常擅長成為一種人類病毒」。

並未參與這項新研究的格拉斯哥大學的病毒學家David Robertson，他表示，這項研究不僅為新冠病毒，也為所有人畜共通病毒 - 即從動物傳播給人類的病毒 - 提供了新的見解。 “這是理解人畜共患風險的關鍵，”他說道:“病毒可以在自然界中傳播，而無需進行適應性改造就能成功感染或傳播給人類。”

Wertheim博士表示，如果這種情況普遍存在於多種人畜共通病毒中，那麼未來我們可能會面臨更多疫情。 他說道:「讓我們陷入困境的，是我們不知道的事」，「它們就在那裡，隨時準備爆發」。

So, a new study finds that Covid is pretty ordinary. Scientists have compared seven viral outbreaks that occurred in recent decades. For the most part, the researchers find that the outbreaks have not been preceded by any unusual genetic changes in the viruses. In all but one case, in 1977, the viruses circulated in animals and gained the ability to spread to and among people. Covid appears to have arisen from some really bad luck of human. Apparently, we can expect more pandemics in the future. 

科學家窺見新疫情的起源（1/2）

Recently the New York Times reported the following:

Scientists Get a Glimpse of How New Pandemics Are Made (1/2)

Researchers have devised a new tool for discerning between naturally occurring viral outbreaks and those resulting from lab accidents.

The NYT - By Carl Zimmer

March 9, 2026

Updated 9:43 a.m. ET

The Covid pandemic was an extraordinary moment in history. Starting at the end of 2019, a virus new to science swept across the planet, killed more than 25 million people and caused trillion of dollars in economic damage.

But as outbreaks go, Covid was pretty ordinary, a new study finds.

Scientists compared seven viral outbreaks that occurred in recent decades, including epidemics of Covid, Ebola and influenza. For the most part, the researchers found, the outbreaks were not preceded by any unusual genetic changes in the viruses. In all but one case, in 1977, the viruses circulated in animals and gained the ability to spread to and among people only by unfortunate coincidence.

“We see that time and again,” said Joel Wertheim, a virologist at the University of California San Diego. He and his colleagues published the study on Friday in the journal Cell.

Dr. Wertheim and his colleagues reconstructed the evolutionary history of these viruses by looking at their genes. They tracked how viruses gained different kinds of mutations before causing outbreaks and looked at that pattern after the viruses jumped into the human population.

In one line of research, the scientists examined the influenza pandemic of 2009. In that year, a new strain of influenza emerged in North America and went on to infect one-quarter of all people on Earth and kill 230,000.

Other studies of the virus have revealed that it came from pigs, where influenza viruses gain mutations on a regular basis. Some mutations made it harder for the viruses to spread to other pigs. Others provided an evolutionary edge; still others had no effect.

The viral strain that jumped into humans in 2009 had split off as its own evolutionary branch at least a decade earlier. Until it reached humans, its evolution looked ordinary; the pattern of mutations gained and lost by the virus resembled what scientists would expect to see in any flu virus thriving in pigs.

Only after the virus jumped to humans did things change, and dramatically.

After infecting humans in 2009, the flu virus gained many new mutations. In pigs, those mutations would probably have hindered the virus’s ability to multiply and caused it to be outcompeted by other viruses infecting the animals.

But once the virus settled in a new host, those old constraints were gone. It began adapting to spread more successfully in humans.

Dr. Wertheim and his colleagues ran the same analysis on other outbreaks, including the Ebola epidemic that swept across West Africa in 2013 and is thought to have originated in a bat, and the 2022 outbreak of mpox, a virus that causes painful blisters and is believed to be harbored by African squirrels. Again and again, the researchers saw the same pattern: The viruses that eventually jumped into humans did not evolve in an unusual way beforehand but did change dramatically afterward.

“Once it gets into humans, it’s a new day,” Dr. Wertheim said.

But one virus turned out to be a major exception to that rule, the new study found. Its unique mutations suggest that it may have been set loose by a scientific accident.

In 1977, the world was hit by a pandemic that came to be known as Russian flu, because the first cases were reported by the Soviet Union. Scientists were baffled by the virus: Its closest relatives were not in pigs or other animals but instead looked a lot like viruses that were circulating in the early 1950s, a quarter-century earlier.

Some scientists speculated that the Russian flu was not a spillover from a pig or a bird. Rather, they suggested, it had emerged from a scientific mishap, perhaps a vaccine trial that had gone wrong in the Soviet Union or China.

The vaccine makers might have used a common technique that involved producing a vaccine made of weakened viruses. Viruses growing in Petri dishes in a lab accumulate mutations that would harm them if they were infecting a person. Scientists speculated that Soviet or Chinese scientists thawed out some old flu virus to make a weakened vaccine but used faulty techniques that allowed the virus to spread from person to person.

(to be continued)

Translation

科學家窺見新疫情的起源（1/2）

研究人員開發出一種新工具，用於區分由自然發生的病毒爆發疫情和由實驗室事故導致的爆發

新冠疫情是歷史上一個非同尋常的時刻。從2019年底開始，一種科學界先前未知的病毒席捲全球，造成超過2,500萬人死亡，並造成數兆美元的經濟損失。

但一項新的研究發現，就疫情爆發而言，新冠疫情其實相當普通。

科學家比較了近幾十年來發生的七次病毒爆發，包括新冠疫情、伊波拉疫情和流感疫情。研究人員發現，在大多數情況下，這些疫情爆發之前病毒並沒有發生任何異常的基因變化。除了1977年的那次例外，所有病毒都是在動物中傳播，並因不幸的巧合而獲得了在人與人之間傳播的能力。

加州大學聖地牙哥分校的病毒學家Joel Wertheim說:「我們一次又一次地看到這種情況」。他和他的同事於週五在《細胞》（Cell）雜誌上發表了這項研究。

Wertheim博士和他的同事透過分析病毒的基因，重建了這些病毒的演化史。他們追蹤了病毒在引發疫情之前如何獲得不同類型的突變，並研究了病毒進入人類群體後這種突變模式的變化。

在其中一項研究中，科學家檢視了2009年的流感大流行。那一年，一種新的流感病毒株在北美出現，並感染了全球四分之一的人口，造成23萬人死亡。

其他針對病毒的研究表明它來自豬，而流感病毒在豬體內經常發生突變。有些突變使病毒更難傳播給其他豬。有些突變則賦予了病毒進化優勢；有些突變則沒有產生任何影響。

2009年感染人類的病毒株至少在十年前就已經分化成一個獨立的演化分支。在感染人類之前，它的進化過程看起來很普通；病毒獲得的突變和缺失的模式與科學家預期在豬群中傳播的任何流感病毒的模式相似。

只有在病毒感染人類之後，情況才發生了劇烈的變化。

流感病毒在2009年感染人類後，獲得了許多新的突變。在豬群中，這些突變可能會阻礙病毒的繁殖能力，並使其在去感染其他動物的病毒競爭中敗下陣來。

但是，一旦病毒在新宿主中定居下來，這些原有的限制就消失了。它開始適應環境，從而更容易在人類中傳播。

Wertheim博士及其同事對其他疫情進行了同樣的分析，包括2013年席捲西非的伊波拉疫情（相信起源於蝙蝠）以及2022年的麻疹病毒疫情（一種會導致疼痛水皰的病毒，相信由非洲松鼠攜帶）。研究人員一次又一次地觀察到相同的模式：最終傳播到人類的病毒在傳播前並沒有異常的演化，但在傳播後卻發生了顯著的變化。

Wertheim博士說: 「一旦病毒進入人類體內，那就是嶄新的一天」。

但新的研究發現，有一種病毒是這個規則的主要例外。其獨特的突變表明，它可能是由科學事故釋放出來的。

1977年，一場被稱為「俄羅斯流感」的全球性流行病席捲全球，因為首批病例是由蘇聯報告的。科學家對這種病毒感到困惑：它的近親並非來自豬或其他動物，而是與25年前，也就是20世紀50年代初期流行的病毒非常相似。

一些科學家推測，俄羅斯流感並非源自豬或禽類。他們認為，它可能是科學事故的產物，例如蘇聯或中國疫苗試驗中出現的失誤。

疫苗生產商可能採用了一種常見的技術，即用減弱病毒去製造疫苗。在實驗室培養皿中培養的病毒會累積突變，如果它感染人類，這些變異會對病毒造成傷害。科學家推測，蘇聯或中國的科學家解凍了某種舊的流感病毒來製造減弱病毒疫苗，但由於使用了錯誤的技術，導致病毒在人與人之間傳播。

（待續）

US February Employment Statistics: Employment Declines by 92,000 from Previous Month - A Significant Drop

Recently NHK News on-line reported the following:

米 2月雇用統計 就業者は前月比9万2000人減 大幅な減少に

2026年3月6日午後10時49分

(2026年3月7日午前1時10分更新)

雇用統計

円相場や株価に影響を及ぼすアメリカの雇用統計が発表され、2月の農業分野以外の就業者は前の月から9万2000人の大幅な減少となりました。増加を見込んでいた市場予想を大きく下回り、金融市場では雇用情勢の減速への警戒感が再び強まることも予想されます。

アメリカ労働省が6日に発表した2月の雇用統計によりますと、景気の動向を敏感に示す農業分野以外の就業者は、前の月から9万2000人の大幅な減少となりました。

市場の予想では5万5000人程度の増加が見込まれていましたが、これを大きく下回り、アメリカメディアも予想外の減少と相次いで報じています。

また、失業率は前の月から0.1ポイント悪化し、4.4％でした。

専門家の間ではアメリカを襲った寒波や医療大手のストライキなどが影響した可能性があるとの指摘も出ていますが、雇用情勢が減速傾向にある可能性を示唆する内容となっていて、金融市場では、警戒感が再び強まりそうです。

今回の統計の内容などを踏まえてFRB＝連邦準備制度理事会は景気を下支えするための利下げのペースを判断していくことになりますが、アメリカなどとイランとの間の攻撃の応酬を受けて原油価格が大きく上昇していて、インフレが再加速することへの懸念も広がるなか、難しいかじ取りを迫られることになりそうです。

Translation

US February Employment Statistics: Employment Declines by 92,000 from Previous Month - A Significant Drop

Employment Statistics

The US employment statistics, which influenced the yen exchange rate and stock prices, had been released, showing a significant decrease of 92,000 non-agricultural workers in February compared to the previous month. This fell far short of market expectations, which had anticipated an increase, and was expected to rekindle concerns about a slowdown in the employment situation in financial markets.

According to the February employment statistics released by the US Department of Labor on the 6th, non-agricultural workers, a sensitive indicator of economic trends, fell sharply by 92,000 from the previous month.

Market expectations had projected an increase of around 55,000, but this figure fell significantly short, and US media were also reporting one after another that the decrease was unexpected.

Furthermore, the unemployment rate worsened by 0.1 percentage point from the previous month to 4.4%.

While some experts suggested that the cold wave that hit the US and strikes at major healthcare companies might have had an impact, the report suggested that the employment situation might be slowing down, which was likely to heighten concerns in financial markets once again.

Based on the contents of this data, the Federal Reserve would decide on the pace of interest rate cuts to support the economy, but with oil prices rising sharply due to the exchange of attacks between the US and Iran, and concerns about renewed inflation accelerating, the Federal Reserve would likely face a difficult decision.

              So, US employment statistics is showing a significant decrease of 92,000 non-agricultural workers in February compared to the previous month. This number falls far short of market expectations and may generate concerns about a slowdown in the employment situation.

中國家長將家庭作業外判給人工智能 (2/2)

Recently the New York Times reported the following:

China’s Parents Are Outsourcing the Homework Grind to A.I. (2/2)

Parents in China are turning to A.I. chatbots and other tools to help their children gain an edge and ease the fighting over homework.

The NYT - By Vivian Wang and Jiawei Wang - Vivian Wang reported from Beijing. Jiawei Wang reported from Seoul.

March 2, 2026

(continue)

Creating Learning Games

Yin Xingyu, 37, uses A.I. chatbots like DeepSeek in her job as a marketer in Shenzhen. She started to wonder whether the tool could also help her 6-year-old daughter.

Ms. Yin didn’t know how to code, so she turned to what’s known as “vibecoding” instead: using A.I. models to build software by describing what you want in plain language. She worked with DeepSeek to build an interactive English word game for her daughter. The chatbot wrote the code for her.

She is now sharing the prompts on social media so other parents can input them into their own chatbots and replicate her games.

Ms. Yin has also experimented with other models. She used Google’s A.I. image generator, Nano Banana Pro, to create comic strips that used her daughter’s Chinese vocabulary words, starring her favorite characters from movies like “Zootopia” and “Frozen.”

Ms. Yin said she didn’t think that her daughter would become dependent on A.I., because she designed the games to prioritize active thinking over passive stimulation. She plans to encourage her daughter to use the tools even more as she gets older; for example, by using chatbots to brainstorm ideas for essays.

“Most likely that’s how the future will be, and I want her to get used to it from a young age,” she said.

‘A.I. Self-Study Rooms’

Not every effort to use A.I. for education has been successful. Some companies have rolled out products that critics say are more hype than substance.

So-called “A.I. self-study rooms,” for example, are advertised as physical spaces where students can learn from A.I.-powered tablets that tailor learning plans to individual needs. Fees range from a few dollars an hour to hundreds of dollars a month.

A state media report from 2024 about the trend showed a classroom in Zhejiang Province lined with cubicles, where students sat quietly in front of tablets that assessed their completion of assignments for accuracy and speed.

But some parents and former employees have complained that the “A.I.” is merely a marketing facade for prerecorded lessons or other less advanced technology, and that the tablets are just basic, off-the-shelf devices.

State media outlets have also accused the operators of some of the centers of trying to circumvent a 2021 ban on for-profit tutoring that was meant to shield children from having too much homework and families from spending too heavily. Many tutoring services have continued to operate underground. (The study centers have said that A.I. is doing the teaching, not tutors, so the ban does not apply.)

Many of the study rooms have already shut down, according to media reports. The New York Times twice tried to visit one in Beijing, only to find it locked and empty, with posters purporting to show reviews from satisfied parents still on the walls.

May Zeng, 24, worked at a study room in Jiangxi Province for two months last year, where she was in charge of making sure that students didn’t slack off. She thought that parents didn’t care as much about A.I. as they did about having somewhere to put their children.

Still, A.I. was getting used — by Ms. Zeng herself. As part of her duties, she had to write feedback on each student’s progress.

“When I found I really had nothing to say, I’d just throw it to the A.I.,” she said. “In this A.I. self-study room, I was the one using the most A.I.”

Translation

中國家長將家庭作業外判給人工智能 (2/2)

中國的家長們正在轉向人工智能聊天機械人和其他工具，以幫助孩子在學習上取得優勢，並減少因家庭作業而起的爭吵

（繼續）

創建學習遊戲

37歲的Yin Xingyu在深圳擔任行銷人員，她在工作中經常使用像DeepSeek這樣的人工智能聊天機械人。她開始思考，這個工具是否也能幫助她6歲的女兒。

Yin女士不懂程式設計，所以她轉而使用所謂的「感知編碼」：透過用簡單的語言描述需求，利用人工智能模型來建立軟件。她用DeepSeek配合，為女兒開發了一款互動式英語單字遊戲。聊天機械人為她編寫出程式碼。

她現在在社交媒體上分享與電腦溝通的提示，以便其他家長可以用它們輸入到自己的聊天機器械中，從而複製她的遊戲。

Yin女士也嘗試了其他模型。她使用Google的人工智能影像產生器 Nano Banana Pro，創作了一些簡短漫畫，漫畫中使用了她女兒正在學習的中文詞彙，扮演是她最喜歡電影裡的角色，例如《瘋狂動物城》(Zootopia) 和《冰雪奇緣》(Frozen) 。

Yin女士表示，她不認為女兒將會變得依賴人工智能，因為她設計的這些遊戲旨在培養孩子的主動思考能力，而不是被動而非被動的刺激。她計劃鼓勵女兒在長大後更多地使用這些工具；例如，使用聊天機械人來集思廣益，為作文尋找靈感。

她說:「未來很可能就是這樣，我希望她從小就習慣它」。

“人工智能自學室”

並非所有將人工智能應用於教育的嘗試都取得了成功。一些公司推出的產品被批評者認為華而不實。

例如，所謂的「人工智能自學室」被宣傳為實體空間，學生可以在這裡使用人工智能平板電腦學習，這些平板電腦會根據個人需求量身定制學習計劃。費用從每小時幾美元到每月幾百美元不等。

2024年的一篇官方媒體報導了這一趨勢，報導中展示了浙江省的一間教室，教室裡排列著一排排隔間，學生們安靜地坐在平板電腦前，平板電腦會評估他們完成作業的準確性和速度。

但一些家長和前員工抱怨說，「人工智能」只不過是預先錄製的課程或其他不太先進技術的行銷噱頭，而這些平板電腦只是基本的現成設備。

官方媒體也指責一些自學中心的經營者試圖規避2021年禁止營利性課外輔導的禁令，該禁令旨在避免孩子作業過多，也避免家庭支出過高。許多輔導機構仍在暗中運作。 （這些輔導中心聲稱，授課的是人工智能，而非輔導老師，因此禁令對他們無效。）

據媒體報道，許多輔導室已經關閉。 《紐約時報》曾兩度試圖造訪北京的一家輔導室，卻發現它大門緊閉，空無一人，牆上還貼著一些據稱是家長好評的海報。

24歲的May Zeng去年在江西省的一家輔導室工作了兩個月，負責督促學生認真學習。她認為，家長更關心的是，是否有地方可以安置孩子，而不是人工智能。

然而，人工智能確實被Zeng本人使用了 。她作為工作的一部分，需要為每個學生的學習進度撰寫出回應。

她說:「當我發現自己沒什麼好說的時候，我就把它交給人工智能寫」; 「在這個人工智能自學室裡，我是使用人工智能最多的人」。

So, parents in China are turning to A.I. to gain a competitive edge. Some are making interactive learning games or using chatbots to grade their children’s homework, while others are using it to get past a language barrier. Their eagerness to embrace A.I. for learning is a sharp contrast to many in the United States who worry that it may feed students misinformation or erode critical thinking. Apparently, there is a huge different between these two countries due to cultural difference etc.

Note:

1. In the computer world, a “prompt” (提示) is essentially an instruction or input given to a computer system by a user to get a response or action from the computer. The exact meaning depends on context, but here are the main uses: 1. Command Line Prompt (classic meaning) - A prompt is the symbol or text that shows a computer is ready for input. Example: C:\Users\>; 2. Prompt as a Request for Input - Programs often “prompt” users to enter information. Example: “Enter your password:”; “Are you sure you want to continue? (Y/N)”; 3. AI Prompt (modern meaning) - With tools like ChatGPT, a prompt is the text you give an AI to guide its response. Example prompts: “Write a story about space travel”; “Explain quantum physics simply”; “Summarize this article”. (ChatGPT)

Iran Fired Ballistic Missiles at Joint US-UK Base in Indian Ocean, a US Bomber Launch Base: Target Missed

Recently CNN.co.jp reported the following:

イラン、インド洋の米英共同基地へ弾道ミサイル発射も命中せず　米爆撃機の出撃拠点

2026.03.21 Sat posted at 16:40 JST

  （CNN） イランが20日午前、インド洋のディエゴガルシア島にある米英の共同軍事基地に向けて中距離弾道ミサイルを発射したことが分かった。米当局者が明らかにした。

当局者によると、イラン沿岸から約3810キロ離れた基地に向けてミサイル2発が発射されたが、いずれも命中しなかったという。

ディエゴガルシア島は米国の重爆撃機にとって重要な航空拠点で、長年にわたり、米本土から遠く離れた地域への作戦に不可欠な米軍の海外出撃拠点の役割を果たしてきた。

CNNはホワイトハウスや米国防総省、英国防省、ディエゴガルシアの基地を管轄地域に収める米インド太平洋軍にコメントを求めている。

ディエゴガルシア島はモーリシャスから北東1609キロあまりに位置する群島、チャゴス諸島の一部。英国は1814年、フランスの皇帝ナポレオンを撃退後、モーリシャスと合わせてチャゴス諸島を取得した。

モーリシャスは1968年に独立したが、チャゴス諸島は今も英国の支配下にある。

米国と英国は71年、この地域におけるソ連の軍事的影響力を抑える目的で、ディエゴガルシア島に大規模な基地を建設した。

ディエゴガルシア島はイラク侵攻に寄与したほか、アジア各地への任務に飛ぶ爆撃機の重要な着陸拠点として機能してきた。米国による身柄引き渡しとの関連も指摘される。

トランプ米大統領は今年、チャゴス諸島の主権をモーリシャスに引き渡す計画を巡り、英国を「愚かだ」と批判した。

英国は条約の条件に基づき、チャゴス諸島の主権をモーリシャスに移譲する方針。ディエゴガルシアにある基地については、英国が毎年1億3600万ドル（約216億円）をモーリシャスに支払って99年間租借する予定で、米英は引き続き施設を使用できる。

Translation

Iran Fired Ballistic Missiles at Joint US-UK Base in Indian Ocean, a US Bomber Launch Base: Target Missed

March 21, 2026 (Sat) Posted at 4:40 PM JST

(CNN) A US official confirmed that Iran had fired medium-range ballistic missiles at a joint US-UK military base on Diego Garcia Island in the Indian Ocean on the morning of March 20.

According to the official, two missiles were fired at the base, located approximately 3,810 kilometers off the Iranian coast, but neither hit its target.

Diego Garcia Island is a crucial air base for US heavy bombers and has long served as an essential overseas launch base for US military operations far from the US mainland.

CNN had reached out for comment to the White House, the US Department of Defense, the UK Ministry of Defense, and also the US Indo-Pacific Command which oversaw the Diego Garcia base.

Diego Garcia was part of the Chagos Islands, an archipelago located approximately 1,609 kilometers northeast of Mauritius. Britain acquired the Chagos Islands, along with Mauritius, in 1814 after repelling Napoleon the French emperor.

Mauritius gained independence in 1968, but the Chagos Islands remained under British control.

In 1971, the United States and Britain constructed a large base on Diego Garcia with the aim of countering Soviet military influence in the region.

Diego Garcia played a role in the invasion of Iraq and had served as an important landing base for bombers flying to various parts of Asia. Its connection to the US extradition had also been suggested.

This year, US President Trump criticized Britain as "stupid" over its plan to transfer sovereignty over the Chagos Islands to Mauritius.

Britain intended to transfer sovereignty over the Chagos Islands to Mauritius based on the terms of the treaty. The base in Diego Garcia was to be leased by the UK for 99 years, with the UK paying Mauritius $136 million (approximately 21.6 billion yen) annually, allowing the US and UK to continue using the facilities.

So, Iran has fired two medium-range ballistic missiles at a joint US-UK military base on Diego Garcia Island in the Indian Ocean located approximately 3,810 kilometers from the Iranian coast, but neither hit its target. Apparently, up to this moment, Iran is unable to fire the missiles accurately over such a long distance, and for sure the accuracy will be improved eventually.

中國家長正將家庭作業外判給人工智能（1/2）

Recently the New York Times reported the following:

China’s Parents Are Outsourcing the Homework Grind to A.I. (1/2)

Parents in China are turning to A.I. chatbots and other tools to help their children gain an edge and ease the fighting over homework.

The NYT - By Vivian Wang and Jiawei Wang - Vivian Wang reported from Beijing. Jiawei Wang reported from Seoul.

March 2, 2026

Across China, where education is famously cutthroat, parents are turning to artificial intelligence to gain a competitive edge. Some are making interactive learning games or using chatbots to grade their children’s homework. Others are using A.I.-powered gadgets to get past a language barrier.

Their eagerness to experiment is one example of how Chinese users are embracing A.I. for learning, even as many in the United States worry that it feeds students misinformation or erodes critical thinking. This cultural rift is backed by data: A 2025 global survey led by the services firm KPMG found that more than 90 percent of Chinese said they felt optimistic about the technology, compared to just over 50 percent in the United States.

The enthusiasm in China has fueled a sprawling, often unchecked, marketplace for educational technology that is worth more than $43 billion by some estimates, and where gimmicks and exaggerated marketing are common. Yet for some families, the tools are providing genuine relief. Three parents shared videos of their routines, showing how A.I., while imperfect, has made parenting and teaching their children a little easier.

As a mother of two with a full-time public relations job, Zheng Wenqi, 42, had little time to practice English herself, let alone teach her children. She knew her 9-year-old son needed more conversational experience but didn’t know where to turn.

“There just wasn’t an opportunity for him to start talking,” said Ms. Zheng, who lives in northern China’s Heilongjiang Province.

Then she saw a livestream promoting a gadget she could wear to make her conversant in English.

It has two parts: a mask that covers her mouth, and a speaker that hangs around her neck. Ms. Zheng speaks Chinese into the mask, which also muffles her voice. Then, a translation comes out of the speaker. She began wearing it around the house, for 30 to 60 minutes a day.

The roughly $375 device, called Native Language Star, draws on speech and language models developed by several Chinese technology firms, according to the company, based in Shenzhen, that makes the device.

Ms. Zheng said the translations were sometimes stiff. But she said that after about a month, her son was speaking more confidently and initiating conversations.

Ms. Zheng also uses the device with her 5-year-old daughter, who had never learned English before. The child can now describe daily tasks, like getting dressed and putting her shoes on.

“We say, ‘Now is English time, let’s all speak in English,’ and I’ll put that thing on,” Ms. Zheng said. “And then they just say whatever they know.”

A Chatbot With ‘Eyes’

Li Linyun, a stay-at-home mother, used to fight with her 10-year-old daughter, Weixiao, over her studies.

Now Ms. Li has delegated supervision of Weixiao’s schoolwork to an A.I. chatbot.

“It’s a 24-hour online teacher, and it’s knowledgeable and extremely patient,” said Ms. Li, who lives in Hunan Province, in central China.

Ms. Li uses Doubao, China’s most popular A.I. chatbot, which was created by ByteDance, the parent company of TikTok. It has a camera function, which parents refer to as Doubao’s “eyes.” People can use it to learn more about their surroundings, for instance by asking Doubao to identify plants or give more details about museum artifacts. (ChatGPT has a similar function for paying users; Doubao’s is free.)

After more experimentation, Ms. Li learned that Doubao could explain grammar rules better than she ever could.

Weixiao said she liked that she could ask Doubao to repeat explanations as many times as she needed, while her teachers moved on quickly. “It explains in more detail, so I can understand,” she said.

Ms. Li also began asking Doubao to grade completed homework assignments by uploading a photo to the app.

The chatbot identifies wrong answers and corrects them, Ms. Li said, though it sometimes makes mistakes.

The camera can also monitor Weixiao’s posture. But Ms. Li said she rarely used that feature, because her daughter didn’t like the feeling of being watched.

Ms. Li said she wasn’t worried about feeding so much footage of Weixiao to the chatbot. In the social media age, “we don’t have a lot of privacy anyway,” she said.

And the benefits were more than worthwhile. She no longer had to spend hundreds of dollars a month on English tutoring, and Weixiao’s grades had improved. “It makes educational resources more equitable for ordinary people,” Ms. Li said.

Her relationship with Weixiao had improved, too, she said. “To ease tensions in a parent-child relationship, you can’t spend too much time on homework,” she said. “Just encouraging her is enough.”

(to be continued)

Translation

中國家長正將家庭作業外判給人工智能（1/2）

中國的家長們正轉向人工智能聊天機器人和其他工具，幫助孩子在學習上取得優勢，並減少因家庭作業而起的爭執

在中國，教育競爭異常激烈，家長正轉向人工智能以獲得競爭優勢。有些家長會製作互動式學習遊戲，或使用聊天機器人批改孩子的作業。另一些家長則使用人工智能設備來克服語言障礙。

他們樂於嘗試，這正是中國用戶接受人工智能用於學習的一個例子，雖然許多美國人擔心人工智能會向學生灌輸錯誤訊息或削弱批判性思考。這種文化差異有數據佐證：畢馬威會計師事務所（KPMG）主導的一項2025年全球調查發現，超過90%的中國人對這項技術持樂觀態度，而美國祇有略高於50%的人持相同看法。

中國的熱情催生了一個龐大且往往缺乏監管的教育科技市場，據估計，該市場價值超過430億美元，各種噱頭和誇張的行銷手段層出不窮。然而，對一些家庭來說，這些工具確實帶來了實質的幫助。三位家長分享了他們的日常視頻，展示了人工智能雖然並不完美，但確實讓他們的育兒和教育工作變得輕鬆了一些。

42歲的Zheng Wenqi是一位有兩個孩子的母親，同時也從事全職的公關工作。她幾乎沒有時間自己練習英語，更別提教孩子練習了。她知道自己9歲的兒子需要更多的口語練習，但不知該向誰練習說英語。

住在中國北方的黑龍江省的Zheng女士說道:「他根本沒有機會開口說英語」。

後來，她看到一個直播，推廣一款可以穿戴的電子設備，幫助她學習英語。

這款裝置由兩個部分組成：一個遮住嘴部的面罩和一個掛在脖子上的揚聲器。Zheng女士對著面罩說中文，面罩也會降低她的聲音。然後，揚聲器會播放翻譯。她開始每天在家裡戴它30到60分鐘。

根據生產這款設備的公司介紹，這款名為「母語之星」（Native Language Star）的設備售價約為375美元，其語音和語言模型由多家中國科技公司共同開發，生產公司位於深圳。

Zheng女士說，翻譯有時略顯生硬。但她說，大約一個月後，她的兒子開始自信地用英語交流，並主動發起對話。

Zheng女士還用這款設備教她5歲的女兒，女兒之前從未學過英語。這女孩現在可以描述一些日常活動，例如穿衣服和穿鞋。

Zheng女士說: 「我們會說，『現在是英語時間，大家一起說英語吧！』然後我會戴上那個東西」; 「之後他們就會說他們知道的英語了」。

 

擁有「眼睛」的聊天機械人

全職媽媽Li Linyun以前常常因為10歲的女兒Weixiao的學習而和她吵架。

現在，Li女士把Weixiao的作業的輔導委託給了人工智能聊天機械人。

居住在中國中部湖南省的Li女士說:「它就像一位24小時線上老師，知識淵博，而且非常有耐心」。

Li女士使用的是豆寶 (Doubao)，這是中國最受歡迎的人工智能聊天機械人，由TikTok的母公司位元組跳動 (ByteDance)開發。它有一個攝影機功能，家長稱之為豆寶的「眼睛」。人們可以利用它來了解更多周圍環境的信息，例如讓豆寶識別植物或提供更多關於博物館藏品的細節。 （ChatGPT 為付費用戶提供類似功能；豆寶的功能是免費的。）

經過更多嘗試，Li女士發現豆寶講解語法規則比她自己講得更好。

Weixiao說她很喜歡可以不停地重複讓豆寶講解，而老師們講得很快。 她說:“它講得更詳細，所以我能理解”。

Li女士也開始讓豆寶批改完成的作業，只需將照片上傳到應用程式即可。

Li女士說，聊天機械人可以識別錯誤答案並進行糾正，儘管它有時也會出錯。

攝影機還可以監控Weixiao的舉動。但Li女士說她很少使用這個功能，因為她女兒不喜歡被監視的感覺。

Li女士說，她並不擔心把Weixiao的那麼多影像資料給聊天機械人。她說，在社群媒體時代，「我們本來就沒太多隱私可言」。

同時，這樣做的好處顯而易見。她不用再每月花幾百美元請英文家教了，Weixiao的成績也提高了。Li女士說: 「這讓一般人更容易獲得教育資源」。

她還說，她和Weixiao的關係也改善了。 她又說:“要緩解親子關係緊張，不能在作業上花太多時間”，“鼓勵她就足夠了。”

（待續）