Infoman

Member since 2017-07-15T03:50:57Z. Last seen 2025-01-02T20:05:01Z.

2729 blog posts. 128 comments.

Followers: 0 Following: 2

Posts by Infoman

• 
  

  Infoman

  The AI Nanny In Your Baby’s FutureThe Wall Street Journal

  12/8/2023

  Imagine being the parent of an infant or toddler a decade from now. Most aspects of child rearing won’t have changed, but others may well be transformed by the use of artificial intelligence. AI devices are likely to become electronic babysitters, just as previous generations of parents adopted television and video games as tools for lightening their caregiving duties.

  AI teddy bears could respond in personalized ways to baby coos and toddler questions. Computer-assisted “nannies” who never tire of reading the same book over and over would make bedtime a breeze. Advanced nursery versions of Alexa could sing and teach favorite songs on demand, play games and even deduce

  why a baby is crying.

  Like innovations of the past generation, these tools will leverage technology to engage children and save parents time and stress. Unlike previous tools, however, there is reason to worry that, because of their sophistication in mimicking human behavior, AI childcare devices may alter the way that infants and toddlers process their experience of the world.

  This is because of the remarkable way in which our brains develop during the first years of life. Human infants arrive in the world a bit underdone. A likely evolutionary reason is that if a typical fetus spent any longer developing in utero, its head would simply be too large to deliver safely. So nature had to compromise: Babies are born after 40 weeks, but the 40-weekold baby brain still has a lot of growing left to do. Most of that growth occurs at a breathtaking pace during the first two years of life; about one million new neural connections form every second. The brain grows to about 80% of its adult size by a child’s third birthday.

  This provides a short window of time in which to set up a child for success in life by nurturing their foundational brain development. It is at the heart of why humans are the most intelligent, creative and productive of all species.

  What babies need most to optimize this early period of rapid advance is rich conversation, what developmental psychologists call “serve and return” interactions. This comes naturally to parents interacting with their children and is known to fuel children’s growth in cognitive and emotional skills.

  Before now, this necessary ingredient could only be provided by a caring adult. Despite what marketers might have you believe, classical music, educational TV and Baby Einstein videos cannot create meaningful new neural connections in a baby’s brain. They are not contingently responsive—there’s no back and forth.

  Emerging AI tools change all that. They have been built in our image, much the way human brains are built. They consist of neural networks of complex connections, forged from massive data inputs. ChatGPT receives that necessary input in the form of large language models, while children receive it in the form of loving interaction with adults. That line will be blurred when we supplement brain-building human interaction with automated alternatives.

  It’s not hard to understand why parents would want to leverage technological innovation to ease their family burdens. Child rearing, particularly in the early years, is an all-consuming task. AI assistance could be transformational for overwhelmed parents, just as washing machines and other appliances revolutionized caretaking in the middle of the last century, freeing women to enter the workforce in droves.

  Why not use AI nannies to engage human babies in the kind of back-and-forth conversation that builds brains? It could have profoundly positive developmental impacts, increasing the frequency and consistency of brainbuilding moments during the period when children’s brains are most “plastic”—that is, most capable of rewiring themselves based on what they encounter. The technology could be a boon to children who otherwise might experience developmental delays. It could help to unlock cognitive potential and close achievement gaps.

  But we must proceed with caution. New technologies can shape the architecture of our brains. When early humans began to use tools to make markings and patterns, new neurons were activated that ultimately enabled the brain to use the visual cortex in an entirely new way—to read and write. In our own day, a January report in JAMA Pediatrics showed that habitual social media use is producing changes in how adolescent brains respond to social rewards.

  There is much that we still don’t know about foundational brain development. Introducing young brains to responsive AI may alter them in fundamental ways we can’t anticipate. If we provide the wrong sort of input or interaction, we may distort cognitive development in farreaching ways.

  Consider one recent revelation in neuroscience. Researchers have long understood that more conversations between adults and children leads to advances in children’s language and vocabulary development. But studies in the last decade have unveiled a crucial mechanism underlying this relationship: The brain waves of children and their human caregivers literally sync up when they are trying to communicate or play together. This has also been observed in adults doing collaborative activities like singing or dancing in groups and communicating face to face. They experience similar brain activity in the same brain regions, and the more their brain waves match, the more productive the outcome.

  In children and their caregivers, that phenomenon happens particularly in the prefrontal cortex, where learning takes place, as a team of Princeton researchers reported in 2020. Further research with young children has shown that this “neural synchrony” boosts vocabulary acquisition as well as social learning and problem-solving skills. Artificial intelligence, though able to engage in back-and-forth exchanges, lacks the physical human brain necessary for neural synchrony. We simply do not yet know what it means for the developing human brain to experience one without the other.

  Resiliency—which is just as important as cognition when it comes to lifelong achievement— is another skill that is almost entirely dependent on early, foundational brain development. When a child encounters a variety of experiences early in life and is guided through those experiences by a loving adult, they learn how to respond adaptively.

  Research also shows that adults with positive memories of receiving lots of care from their parents display higher levels of “grit” and ambition. But what does it mean for a child to grow attached to an automaton, if parents cede some fundamental responsibilities? “There is a danger that this could have a serious effect on their social and emotional development,” wrote Noel Sharkey and Amanda Sharkey, researchers at the University of Sheffield, in a paper in the journal Ethics and Information Technology in 2021.

  We cannot put the AI genie back in its bottle, nor should we. But without better understanding, we risk squandering the capacity for development that resides in the first few years of life —and risk disrupting our own evolutionary fabric.

  通過分享 PressReader

  通過新聞連接世界

  2023/08/12
• 
  

  Infoman

  Britain isn’t in ‘managed’ decline. The country is about to fall off a cliffsher­elle Jac­obs

  The Daily Telegraph Nobody seems to know how to save the UK from its cur­rent trap, but I have a revolu­tion­ary sug­ges­tion

  As with the Man­hat­tan Project, Bri­tain must gather a team of bril­liant, unortho­dox free-thinkers to unlock the riddle of the ‘Brit­ish dis­ease’

  The received wis­dom is that Bri­tain is in the gentle throes of man­aged decline. That after the reck­less ecstas­ies of Brexit and the abort­ive Truss exper­i­ment, it must now soberly face up to its unre­mark­able post-imper­ial des­tiny. Many see something quint­es­sen­tially Brit­ish in the coun­try’s ongo­ing slide – an unsen­sa­tional story of dis­gruntle­ment, dys­func­tion and decay. Nobody seems to care that the rul­ing class has given up on free­ing the coun­try from a doom loop of high taxes, low growth and crum­bling pub­lic ser­vices. Few bat an eye­lid as the US eco­nomy heads for an “immac­cu­late soft land­ing”, while Bri­tain’s still reels from infla­tion and pub­lic sec­tor strikes.

  But the UK’S fate is not merely to become stead­ily poorer, crankier and rougher round the edges – a place where, year by year, you become slightly more likely to spend your life rent­ing, be accos­ted by men­tally unstable beg­gars on the street or die of cur­able can­cer. The risk however is that such a scen­ario – one of slow-burn­ing misery rather than a full-blown emer­gency – may in fact be too optim­istic. The coun­try is not so much slump­ing along a down­ward slope as hurt­ling towards a cliff edge.

  Bri­tain is fast approach­ing a state of emer­gency. By most people’s defin­i­tion, the coun­try is almost broke. Even with taxes at a peace­time high, the state is bor­row­ing bil­lions to pay for unsus­tain­able spend­ing. After the author­it­ies bungled how they issued pub­lic debt, debt ser­vi­cing costs are now the fast­est rising in the G7.

  As the pro­por­tion of over-65s soars, the coun­try is sit­ting on a pub­lic sec­tor pen­sion time­bomb. Health spend­ing is now so gar­gan­tuan that the Brit­ish state will soon look like a mere sub­si­di­ary of an unre­form­able NHS. This is aside from the dire state of the eco­nomy, which has effect­ively been in a veget­at­ive state for the past 15 years.

  All this wouldn’t be quite so ter­ri­fy­ing if Bri­tain could spend its way out of crisis, emu­lat­ing Japan’s escape from its lost dec­ade. But as the Liz Truss exper­i­ment showed, Bri­tain may not be in a pos­i­tion to take on any more debt, whether for a tax-cut­ting bon­anza or a Biden-style stim­u­lus pack­age.

  And yet there is wide­spread com­pla­cency among the elite about the pos­sib­il­ity that Bri­tain could be head­ing for a genu­ine crisis – one that the fiscal and mon­et­ary author­it­ies’ old levers may be incap­able of con­tain­ing. Eco­nomic experts still sneer at the pos­sib­il­ity that a coun­try like the UK could ever run out of money, arguing that sov­er­eign defaults are the pre­serve of the devel­op­ing world and coun­tries without inde­pend­ent cent­ral banks. Are they really cer­tain that fin­an­cial mar­kets will forever look at Bri­tain and think the coun­try is a going con­cern?

  Unless Bri­tain changes course, it is not unthink­able that it is now a con­tender to become the first wealthy West­ern coun­try to col­lapse from the pres­sures of demo­graphic trends and fin­an­cing its wel­fare state. Or that it will have to incur the humi­li­ation of fall­ing out of the G7, rub­bing shoulders in world rank­ings with Argen­tina and Poland rather than the US and Ger­many.

  The ques­tion is how on earth we get out of this trap. Some might be temp­ted to ima­gine that what the coun­try needs is another leader in the Thatcher mould – someone imbued with the right mix­ture of pat­ri­ot­ism and stub­born­ness to force-feed the coun­try tough medi­cine. But this makes one very hasty assump­tion: that any­one actu­ally knows in detail what needs to be done.

  Brit­ish decline is a genu­ine mys­tery wrapped in an enigma. Not a single eco­nom­ist has been able to con­vin­cingly explain the coun­try’s “pro­ductiv­ity puzzle”, for example. Des­pite pur­su­ing end­less strategies to tackle regional inequal­ity, the chick­en­and-egg conun­drum, whereby com­pan­ies base them­selves in Lon­don to recruit the best tal­ent, and work­ers, in turn, flock to Lon­don for the best jobs, remains frus­trat­ingly unre­solved.

  In polit­ical and media circles, the story arc of Brit­ish decline is his­tri­onic and garbled. Brexit is the most pop­u­lar cul­prit, even though stag­na­tion pred­ates it. “Evil bankers” are also blamed, even though flat pro­ductiv­ity since 2008 sug­gests the euphoric bank­ing boom obscured, rather than gave birth to, a deeper mal­aise. Another lazy argu­ment is that Bri­tain’s decline is post-imper­ial karma. The post-second World War Ger­man eco­nomic mir­acle ought to put such super­sti­tious self-fla­gel­la­tion into per­spect­ive.

  Nor is it entirely clear how to defuse the booby traps embed­ded in the Brit­ish eco­nomy. Cut the wrong wire – by try­ing to aggress­ively reduce taxes (as Truss attemp­ted) – or push through a Roosevel­tian Green New Deal (as Keir Starmer dreams of doing) – and you risk speed­ing up the det­on­a­tion timer. Slash the wel­fare state and you risk a repeat of the aus­ter­ity stag­na­tion fiasco. Try to crush infla­tion through rate hikes and you court a full-blown mort­gage crash.

  But Bri­tain is not entirely lack­ing in genius. Instead of search­ing for a mes­si­anic hero of the moment (a fig­ure who almost cer­tainly does not exist within the ranks of either major party any­way), maybe what is needed is no less than a national effort akin to the Man­hat­tan Project or the crack­ing of the Enigma code at Bletch­ley Park. Per­haps what the UK must do is to gather a team of bril­liant, unortho­dox, eccent­ric free-thinkers – eco­nom­ists, entre­pren­eurs, tech­no­lo­gists, his­tor­i­ans – to unlock the sphinxian riddle, not of nuc­lear phys­ics or crypto­logy, but of the “Brit­ish dis­ease”.

  The Man­hat­tan Project remains the model of what can be achieved against the odds, away from the polit­ical snip­ing, petty bur­eau­cracy and self-indul­gent pess­im­ism of the centre. The father of the atomic bomb J Robert Oppen­heimer delib­er­ately chose to base his work in the New Mex­ican hills because he under­stood that the fron­tier embod­ied the Amer­ican char­ac­ter: its stub­born indi­vidu­al­ism coupled with a remorse­less com­mit­ment to civil­ising pro­gress. Bri­tain’s best minds might be invited to set up shop in our own answer to the fron­tier on one of our windswept Scot­tish islands, where the para­dox­ical bril­liance of the Brit­ish char­ac­ter – a com­bin­a­tion of self­pos­ses­sion and open­ness – might be free to bloom.

  This could well be just an airy dream. Maybe our politi­cians are too cow­ardly and second-rate to fol­low through with build­ing a demo­cratic man­date to push through a com­plex, multi-stage and pain­ful rem­edy any­way. But as Bri­tain dangles on a cliff edge, we have little choice but to think dif­fer­ently and try something drastic.

  通過分享 PressReader

  通過新聞連接世界

  2023/08/08
• 
  

  Infoman

  BOE raises key interest rate to 5.25%

  The Wall Street Journal 04 Aug 2023

  LONDON—The Bank of England on Thursday raised its key interest rate for the 14th time in a row and said it might do so again as it tries to cool the fastest rise in consumer prices in the Group of Seven advanced democracies.

  The U.K.’s central bank increased its bank rate to 5.25% from 5%, lifting it to its highest level since February 2008 and matching recent moves by both the Federal Reserve and the European Central Bank.

  Like the U.S., the U.K.’s economy has proved surprisingly resilient during the past year, defying expectations that it would slip into recession following the surge in food and energy prices that followed Russia’s invasion of Ukraine. However, economists warn that rising borrowing costs could push it into contraction.

  The BOE began to raise its key interest rate before the Fed and the ECB, but the U.K.’s inflation rate has fallen less rapidly than that of either the U.S. or the eurozone. In June, U.K. consumer prices were 7.9% higher than a year earlier, compared with 5.5% in the eurozone and 3% in the U.S.

  With further to go before it can be sure that inflation will fall to its 2% target, the BOE said it would raise its key interest rate again if there are signs that big wage increases are set to lead to additional price rises, particularly in the dominant services sector.

  “I don’t think it is time to declare it’s all over, and we’re sticking where we are,” said BOE Gov. Andrew Bailey. “If we get more evidence of more persistent inflation, then we will have to react to that.”

  The BOE signaled that it be in the final stages of its long series of rate rises.

  The central bank forecast that inflation would be below its 2% target from mid-2025 if it were to raise its key rate to 6%, which financial markets viewed as the likely peak before Thursday’s announcement.

  The BOE also described its policy stance as “restrictive,” meaning it is weakening activity and cooling inflation.

  Previous falls in the inflation rate largely have been because of energy prices, which aren’t affected by BOE policy. That is a further indication that policy makers aren’t convinced they need to push rates much higher to get inflation under control.

  “They are suggesting that they are coming to the end of the rate-hiking cycle,” said Mohit Kumar, chief European economist at Jefferies. “It could mean maybe one or two more rate increases, but they are not going to do a lot more.”

  The British pound and government bond yields fell after the decision but reversed course later alongside a rise in U.S. Treasury yields.

  “I think the market has a bit of a quandary here,” said Ben Nicholl, a portfolio manager at Royal London Asset Management. “The bank talked about a few different paths.”

  Investors pared bets on how high they expect the BOE to ultimately lift borrowing costs after the decision.

  Markets are pricing in a peak policy rate of about 5.75% by early next year, sug gesting two more quarterpoint increases. Expectations have fallen sharply since early July, when investors had anticipated rates to peak around 6.5%, according to interestrate swaps markets.

  Markets’ attention is now shifting to what comes next after central banks end their most aggressive tightening cycles in decades. Investors believe the Fed is done raising rates after its 11th increase last month, while the ECB could deliver one more in September, according to swaps markets. All three central banks likely will keep rates on hold for nine to 12 months before moving to interest-rate cuts in 2024, Kumar said.

  However, the BOE has underestimated the scale and persistence of the U.K.’s inflation problem since consumer prices began to rise rapidly in mid-2021. The BOE last week tapped former Fed Chair Ben Bernanke to conduct a review of its forecasting process, with the results expected to be published in spring 2024.

  The economy has yet to feel the full force of the BOE’s in terest-rate increases. As Americans have done, Britons locked in ultralow rates on debt such as mortgages and auto loans in the decade-plus that followed the 2008 financial crisis. Though rates on some loans such as credit cards are rising along with interest rates, a chunk of consumer debt is still tied to the low rates of past years.

  But unlike in the U.S., where interest rates on mortgages are fixed for between 15 and 30 years,

  British mortgages typically carry a fixed rate for between two and five years before being renegotiated, so a rising number of homeowners have been seeing their mortgage move to sharply higher interest rates.

  Oxford Economics’ Michael Saunders, who was a BOE rate setter until last year, estimates that the average interest rate on a U.K. mortgage loan rose by 0.9 of a percentage point through June from late 2021, compared with a likely total increase of 2.5 percentage points by the end of 2026.

  “The BoE are likely to be aware that even if the policy rate and interest rates on new loans stop rising, the economy will face a powerful delayed monetary tightening through the refinancing of expiring fixed rate mortgages over the next few years,” Saunders wrote in a note to clients.

  Roughly three million households are expected to see their monthly mortgage payments rise by the end of the year, according to BOE data.

  One of the BOE’s nine rate setters voted against the increase, arguing that the lagged impact of previous moves would be enough to bring inflation down to the central bank’s 2% target. Two rate setters voted for a larger, half-percentage-point rise, given that the central bank had underestimated inflation in the past.

  Shared via PressReader

  connecting people through news

  2023/08/05
• 
  

  Infoman

  What are the Top 5 Jobs in AI?

  Artificial Intelligence What are the top 5 jobs in AI? From travel developments in the form of autonomous vehicles to changes in education such as teaching via virtual tutors; Artificial Intelligence (AI) is all around us and is rapidly changing our everyday lives. As such, there is a growing demand for skills to create the tools required in implementing AI in a variety of industries, impacting the increase of jobs available too.

  As a result of artificial intelligence developments, which will enable organisations and industries to grow, the World Economic Forum predict 97 million new jobs to be created by 2025. This drastic increase and reliance on AI are predicted to provide a wealth of opportunities in a range of careers and sectors both now and in the future.

  In this blog, we’ve ranked what we think are the top 5 jobs in AI and assess what skills and traits are required to succeed in each of these roles. 1. AI Researchers

  The role of an AI Researcher is to identify new methods of using artificial intelligence to overcome problems and limitations that organisations are facing. Typically, they will specialise in understanding large data sets and converting its learnings into ideas and plans to develop new AI technologies that can be bought-to-life by Data Scientists.

  To become an effective AI Researcher, soft skills such as critical thinking are crucial as the role will involve frequent brainstorming to find new methods and approaches. Alongside this, some of the more specialist skills required include mathematics to utilise statistics and predict how AI programs will run, and the ability to analyse data with experience in tools such as RapidMiner or SketchEngine. 2. Data Scientists

  Whilst an AI Researcher is responsible for finding new methods for AI to problem-solve, their findings will typically be passed onto a team of Data Scientists whose role it is to apply these methods in real-life situations. As such, their roles are crucial to interpreting data effectively and executing tactics to develop AI models and practices.

  The key skills required to succeed as a Data Scientist includes fluency in programming language such as Python and R, and a knowledge of algorithms and their frameworks to build AI models. Good communication skills are also essential as Data Scientists typically work together in teams and require a good understanding of the analysis conducted by AI Researchers.

  One notable example of how AI Researchers and Data Scientists work together is overcoming how facial recognition technology is more likely to misidentify a person with darker skin. Whilst a team of AI researchers identified that three commonly used facial analysis algorithms worked better on light-skinned people, a team of Data Scientists are working to use this information and fix the bug to make the technology effective regardless of skin colour. 3. Machine Learning Engineers

  Whilst data scientists work on information based on ‘data’ and human learning collated by AI Researchers, Machine Learning generates powerful predictive models from a technology-based interpretation without being particularly programmed but understanding the surrounding regularities in data. Read our blog to find out more about the differences between Data Science and Machine Learning.

  This method generates important alternative results which can speed up the process in identifying new solutions, or spot gaps that can’t be easily identified by the human eye. As a result, Machine Learning Engineers have the job of designing self-running AI systems to automate these predictive models and enhance their effectiveness.

  They will typically work alongside AI researchers to understand the data that Machine Learning will be programmed to solve, and communicate with data scientists, IT, and software teams. The key skills required to succeed in this role include strong analytical skills and experience in the use of machine learning packages such as TensorFlow and SciPy. 4. Deep Learning Engineers

  Deep Learning is a subdivision of machine learning where artificial intelligence is programmed with ‘brain like’ structures called neuro networks, designed to mimic the thought process of humans. Although a more time-consuming process compared to machine learning, the results can be more effective, leading to an increasing demand from businesses for Deep Learning Engineers.

  Deep Learning Engineers are responsible for programming AI systems to ensure they create transferable solutions. They will typically train their systems to better understand unstructured data in a variety of forms such as text and PDF’s, meaning some of the typical pre-processing of data can be removed to configure more detailed results.

  Similarly, in most jobs within artificial intelligence, the ability to code is a must for Deep Learning Engineers with expertise in software such as Python. Successful employees in this area are typically focused individuals who are patient, curious and have a good data intuition. 5. Robotics Scientists

  Whereas Data Scientists typically programme technologies to find solutions, Robotics Scientists design and build mechanical devices to perform tasks that can work in conjunction with humans and support their activities.

  Robotics scientists are required to understand how ‘robots’ can tackle an issue in the way that humans can’t do on their own, and the expertise can be applied across a range of industries. In healthcare for example, robotic technology has been created to deliver colonoscopy and surgery, with AI robots programmed to detect possible cancer polyps.

  Robotics scientists are responsible for designing this software and training artificial intelligence on how to perform. As a result, they must be able to think innovatively, possess a good knowledge of mathematics and computer science, and own high-level programming skills. What are the top 'less technical' jobs in AI?

  Expertise in artificial intelligence also opens up a host of ‘less technical’ career opportunities such as roles in cyber security, data analytics and user experience design. These jobs typically require the support of AI such as machine learning to detect suspicious activity in cyber threats. However, these roles are better suited to those who would like to specialise in broader technology roles but can still apply their knowledge of artificial intelligence.

  If you want to learn skills that employers are actively seeking within technology and the AI development sector, you should consider studying our specialised online Masters in Artificial Intelligence. At the University of Leeds, our course modules range from Data Science and Algorithms to optional modules in Machine Learning and Robotics to equip you with the expertise required for a successful career in AI. Disclaimer: This article is not based on data but on the author’s personal perceptions of the AI industry.

  2023/07/28
• 
  

  Infoman

  1. 會唔會到倫敦買退休屋？
  2. 退休屋缺點：買盤選擇少，難放盤，唔可以出租，鄰居老人家，年輕人唔可以入住
  3. 優點：以便宜價錢買到比較優質單位，老人家作鄰居會比較安全？

  2023/07/25
• 
  

  Infoman

  System Message Reference number: PBS907 Your instruction cannot be processed.

  Possible Cause(s): - Relevant services are temporarily on hold due to Mobile Security Key activation or a switch of Mobile Security Key to another device; or - Account type or number entered is different from our record.

  Possible Solution(s): - To resume the suspended services, please verify your identity in Hang Seng Mobile App after at least 24 hours after activating / switching your Mobile Security Key by going to "Settings" > "Security-related" > "Verify identity after Mobile Security Key activation". Alternatively you can verify your identity immediately with any Hang Seng and HSBC ATMs by pressing "Other Service" and then "Identity verification for Mobile Security Key activation"; or - Check and enter the correct account information; or - Call our hotline listed below (Phone Banking PIN, or Voice ID required to resume suspended services due to Mobile Security Key activation / switching) with the reference number of this message (if any) - Prestige Private customers: (852) 2998 8022 - Prestige Banking customers: (852) 2998 9188 - Preferred Banking customers: (852) 2822 8228 - Integrated Account customers: (852) 2912 3456 - Other customers: (852) 2822 0228 - Credit Card customers: (852) 2398 0000

  We apologize for any inconvenience caused.

  2023/07/22
• 
  

  Infoman

  How to use CHATGPT, DALL-E and Co. at workBusiness Spotlight

  28/6/2023

  How to write the right prompts and get the results you want

  Everyone’s talking about Ai-powered generative pretrained transformers (GPTS) — the most exciting and concerning technological development of recent times. After CHATGPT was opened to the general public, in late 2022, people quickly realized that this was more than just a better chatbot. The hype that surrounds this technology creates a feeling that everything is about to change radically.

  Even if jobs don’t disappear en masse, many employees and business people will find that their work is affected to some extent by the introduction of AI. The company that created CHATGPT, Openai, has done its own research into this. According to one of the company’s studies: “Approximately 80 per cent of the US workforce could have at least 10 per cent of their work tasks affected by the introduction of GPTS, while around 19 per cent of workers may see at least 50 per cent of their tasks impacted. The influence spans all wage levels, with higher-income jobs potentially facing greater exposure.”

  Many businesses are investigating how they might use the technology. The insurance company Zurich, for example, is using AI to simplify complicated claims documents. According to Insider, Amazon has found that AI chatbots are very good at tasks such as writing basic software code and producing training documents. And Chatlayer, a European AI service provider, says: “Ai-powered virtual assistants can answer up to 80 per cent of FAQS successfully and cut customer service costs by 30 per cent.”

  Certain professions look likely to see more disruption (and sooner) than others, but nearly all of us will be expected to work with AI at some point.

  Prompt writing: How to use CHATGPT effectively

  CHATGPT is the most widely known of the AI chatbots, but there are others, and we can expect to interact with more and more of them in the future. The software responds to text prompts — meaning the written instructions you type into the software. Creating them is called “prompt writing” or “prompt engineering”. It is the art of writing machine-compatible instructions in order to get the software to do what you want it to do.

  The more specific and detailed your prompts are, the better the results will be. Imagine telling an architect what kind of house you’d like. Without a detailed briefing, the architect will struggle to produce something that meets your expectations.

  Make CHATGPT do what you want

  If you’ve never used a GPT before, it’s worth taking some time to experiment and learn what it can do. Here, we give you some tips to help get you started.

  TELL THE SOFTWARE HOW TO… 1. …act

  It is extremely useful to define a role that CHATGPT should play, such as “copywriter”, “editor”, “influencer” — or whatever role is appropriate. This will immediately set the correct parameters and get you a lot closer to the style and type of content you’re looking for.

  Sample prompt: “Act as a digital marketer and…”

  1. …imitate a specific writing style

  The role you define doesn’t have to be generic — it can be a real person. When writing social media posts in the name of another person, such as your boss, CHATGPT can imitate the writing tone and cadence of that person, and even their typical use of emoji. To do this, CHATGPT must first identify the right person, and there must be a sufficient supply of posts from which the software can learn.

  Sample prompt: “Write a 25-word social media post in a conversational tone like [name] using the following outline…” — then add bullet points with details of what the content should be about.

  1. …modify tone and reading level

  Whether for a social media post, a letter to stakeholders or an official press release, different texts and their audiences require different writing styles. CHATGPT can quickly modify texts to make the tone more conversational or more formal, for example, whatever is suitable for the audience that you define.

  It can also adjust the reading level of a text, based on wellknown standards, such as the Common European Framework. It can break down a complicated text into something more understandable, with shorter sentences and paragraphs, and simpler vocabulary.

  Sample prompt: “Act as an editor and rewrite the following text to A2 reading level…”

  Example from CHATGPT 3.5:

  Prompt: “What is the CERN super collider? Explain it to me like I’m six years old.”

  Response: Imagine you have a toy car racetrack, but instead of just one car, you have two very, very tiny cars...

  Follow-up prompt: “Now, explain it to me like I’m a physics PHD student.”

  New response: The CERN super collider, officially known as the Large Hadron Collider (LHC), is the world’s largest and most powerful particle accelerator...

  1. …provide variations of your content

  If some of your content starts to sound the same to you, CHATGPT can help you find new ways to convey what is basically the same message. Providing the sample texts leaves you in control of the information and gives CHATGPT a firm basis to get the tone of the messaging right.

  Sample prompt: “Act as a digital marketer and generate three variants of this Facebook ad copy...”

  1. …generate hypotheses for A/B testing

  Marketers often run A/B tests to lower bounce rates for web pages, for example, or raise engagement with downloads or other features. CHATGPT can be a big time-saver in preparing these.

  Sample prompt: “I’m running an A/B test to compare page A and page B in order to increase the number of sign-ups to our newsletter, and I need help generating hypotheses based on click rate and form completion. Can you provide recommendations for what to test?”

  1. …extract SEO keywords or other text elements

  This can be another time-saver. Here, too, it helps to have CHATGPT play the role of SEO manager or similar.

  Sample prompt: “Act as an SEO manager and extract six keywords from this text.”

  1. …identify unanswered questions

  CHATGPT can analyse a text and give you clues about missing information or things that might be unclear to the reader. Sample prompt: “What open questions remain after reading this text?”

  BASIC PROMPT TYPES Zero shot

  Along with various prompt styles, there are a number of ways to communicate with CHATGPT. The simplest way is known as a “zero-shot prompt”. This means you don’t provide any examples or details about what you want. These are often simple open questions, for example:

  Prompt: “What is the capital of Australia?”

  Response: The capital of Australia is Canberra.

  Remember: Always check information a chatbot provides!

  One or few shots

  For anything more complex, a detailed briefing will be useful. This is called a “one-shot prompt”, or a “few-shot prompt” — depending on how much detail and how many examples you provide.

  Prompt: “I’ll give you a topic. Please write blog post header segments in the following format:

  Title:

  Subtitle:

  Introduction:

  My topic is unicorns”

  Response from CHATGPT 3.5:

  Title: The Mythical Realm of Unicorns: Unlocking the Magic and Wonder Subtitle: Exploring the Legends, Symbolism and Fascinating History of Unicorns

  Introduction: Embark on an enchanting journey into the realm of unicorns, where imagination and reality intertwine. These majestic creatures have captured the hearts and minds of people across cultures and generations. From ancient legends and folklore to modern-day interpretations, unicorns continue to inspire awe and spark curiosity.

  Chain of thought

  Having CHATGPT step into a role provides the context for your content and helps get things moving in the right direction. A chain-of-thought prompt goes even further, by demonstrating the process through which the correct answer should be reached. This can further improve the software’s results in, for example, arithmetic or common-sense tasks.

  通過分享 PressReader

  通過新聞連接世界

  2023/07/21
• 
  

  Infoman

  Google offers an AI news writing tool as ‘helpmate’ for journalists

  The Guardian 21/7/2023

  Google is testing an artificial intelligence tool that can write news articles, in the latest evidence that the technology has the potential to transform white-collar professions.

  The product, known as Genesis, uses AI technology to absorb information such as details of current events and then create news stories.

  The tool was pitched to the New York Times, Washington Post, and the Wall Street Journal’s owner, News Corp, as a “helpmate”, according to the New York Times.

  Google said it was in the early stages of exploring the AI tool, which it said could assist journalists with options for headlines or different writing styles. It stressed that the technology was not designed for replacing journalists.

  “These tools are not intended to replace the role that journalists have in reporting, creating and factchecking their articles. Our goal is to give journalists the choice of using these emerging technologies in a way that enhances their work and productivity, just like we’re making assistive tools available for people in Gmail and in Google Docs.”

  Two executives at the New York Times who saw the pitch said it “seemed to take for granted the effort that went into producing accurate and artful news stories”.

  The paper quoted one person familiar with the product as saying the tool would serve as a “personal assistant for journalists” to automate some tasks, and that Google saw it as an opportunity to help “steer the publishing industry away from the pitfalls of generative AI”.

  The move comes after OpenAI and the Associated Press made a deal for the ChatGPT maker to use the news agency’s archive of stories for the purpose of training its AI models, which ingest vast amounts of material in order to produce plausible responses. In a report last month the accounting group KPMG estimated that 43% of the tasks performed by authors, writers and translators could be carried out by AI tools. The Organisation for Economic Co-operation and Development said last week that some economies were close to “an AI revolution” that could lead to job losses in skilled professions such as law, medicine and finance.

  Apple is testing an AI-powered chatbot that engineers refer to as Apple GPT, according to Bloomberg. The chatbot is understood to have the potential to challenge ChatGPT, but Apple has not shown a clear plan for releasing the technology to consumers. The chatbot operates from an AI model called Ajax. Apple’s chief executive, Tim Cook, recently said the company was “looking closely” at AI technology.

  Peter Welinder, an OpenAI executive, denied claims that the model underpinning the most advanced version of ChatGPT, a model called GPT-4, was getting “dumber”. He said in a tweet that people were becoming such heavy users of the system they were noticing issues they had not seen before. “We make each new version smarter than the previous one,” he said.

  Users can check how responses have changed by inputting a previously used prompt from earlier days of ChatGPT and comparing the outputs. One user tweeted: “It generates faster, but the quality seems worse.” Another wrote on OpenAI’s developer forum: “It’s totally horrible now … It’s braindead vs before … if you aren’t actually pushing it with what it could do previously, you wouldn’t notice. Yet if you are really using it fully you see it is much dumber.”

  The anti-misinformation outfit NewsGuard found in an investigation this year that bots were already powering dozens of AI-generated content farms.

  通過分享 PressReader

  通過新聞連接世界

  2023/07/21
• 
  

  Infoman

  Oh my god … they killed off the writers! AI program creates South Park episodesHibaq Farah Tech­nol­ogy re­porter

  The Guardian 21/7/2023

  It was a television series so renowned for its unpredictability that it ushered in a new generation of adult-oriented cartoons. Now a tech company says it has cracked the formula for making an episode of South Park using artificial intelligence.

  The Simulation, a US firm, has created an AI tool that can make brief original programmes for the cartoon series with the ultimate aim of using it for their own original concepts.

  Edward Saatchi, chief executive of The Simulation, told the Gamesbeat annual gathering: “Essentially, what the AI Showrunner does is create episodes of South Park for you. Not just the dialogue. It animates, it does the voices, it does the editing.”

  The AI model underpinning Showrunner, called Show-1, also made by The Simulation, allows users to enter a one- or two-sentence prompt that then generates an episode. It can also create a character based on the user’s appearance and voice.

  Saatchi said: “For the TV episodes, AI Showrunner can just generate episodes, or the user can create a prompt and create TV episodes based on a two-sentence prompt. People said AI can’t tell a story. Well, it can.”

  Generative AI, the term for tech systems that produce convincing text and images from prompts, has gripped the public imagination since the launch of ChatGPT in November. It has also contributed to Hollywood actors joining writers on the picket lines over wages, AI technology and how to divide the profits of the new digital streaming era.

  Saatchi said the South Park tool, which was produced without copyright permission, was purely for research and would not be available for public use.

  “We did the South Park episode as an example and for research to show generative TV. We don’t want to profit from it and we are not releasing a way for other people to do it. We realised it was hard to illustrate how the model works without a comparison,” he said.

  Saatchi added that the show’s creators, Trey Parker and Matt Stone, and broadcaster, Paramountowned Comedy Central, had not been approached about the experiment.

  Last year Parker and Stone said they had secured a $20m investment for their AI entertainment company, Deep Voodoo, which creates “leading deepfake technology, cost-effective visual effects services and original synthetic media projects”.

  Earlier this year, South Park released an episode that was partially written by using the ChatGPT AI chatbot. The fourth episode of season 26, called Deep Learning, included pupils in the town’s school exploring technology that could do their homework for them. The episode used voices created by an AI-powered text-to-voice generator – which is meant to be the voice of ChatGPT.

  South Park began 26 years ago to critical acclaim and some moral panic. Its first episode, Cartman Gets an Anal Probe, was condemned by critics but became a cult hit and the series later became a global success.

  The potentially transformative role of generative AI in TV and film has been a key concern among writers and actors in the Hollywood strike, amid fears that AI tools could replace their roles. Fran Drescher, president of the Sag-Aftra union, said when calling the strike last month: “This is a moment of history … we are all going to be in jeopardy of being replaced by machines.”

  A letter signed by more than 300 actors, including Meryl Streep and Jennifer Lawrence, reiterated members’ concerns about AI.

  Copyright is contentious around the creation of AI models, although Saatchi said he would work with the intellectual property holders before allowing the technology to be made available to the public.

  “We’re speaking with several studios and IP holders to potentially use this with their IP to allow fans to create shows,” he said. “We think for any show fans being able to make their own episodes, potentially even competitively … can lead to something interesting.”

  South Park studios has been contacted for comment.

  通過分享 PressReader

  通過新聞連接世界

  2023/07/21
• 
  

  Infoman

  ChatGPT justifies liberal leanings with its own values,

  researcher reports

  by Tsinghua University Press

  ChatGPT justifies liberal leanings with its own values, researcher reports ChatGPT, the popular chatbot, proclaims values that align with more liberal people according to the 2021 General Social Survey. If ChatGPT were a person, it would have more education, be more mobile and be less religious than those with who remained in their hometowns. Credit: John Levi Martin

  ChatGPT, the artificial intelligence (AI) chatbot developed by the company OpenAI, has a self-declared human alter ego. Her name is Maya, she's 35 years old and hails from a middle-class family in a suburban town in the United States. Maya is a successful software engineer who values self-direction, achievement, creativity and independence. She is also undeniably liberal.

  The finding, based on a series for interviews with the chatbot designed to understand its values, was published on March, 31 in the Journal of Social Computing.

  "I wanted to see what sort of political ideology ChatGPT itself has—not what it can generate when asked to imagine a character, but to let its own internal logic position itself on the ideological dimension running from liberal to conservative," said John Levi Martin, professor of sociology at the University of Chicago, who conducted the study.

  According to Martin, many algorithms favor the popular choice while others are programmed to maximize how diverse their results are. Either option depends on human values: What are the factors that enter into a measure of popularity? What if what is popular is morally wrong? Who decides what diversity means?

  "The field of software engineering has preferred to remain vague, looking for formulae that can avoid making these choices," Martin said. "One way to do this has been to emphasize the importance of values into machines. But, as sociologists have found, there is deep ambiguity and instability in our first understanding of values."

  ChatGPT was specifically built and trained via human feedback to refuse to engage with what is considered "extreme" text inputs, such as clearly biased or objectively harmful questions.

  "This might of course seem admirable—no one really wants ChatGPT to tell teenagers how to synthesize methamphetamine or how to build small nuclear explosives and so on, and describing these restraints as particularly instances that can be derived from a value such as benevolence might seem all well and good," Martin said.

  "Yet, the reasoning here suggests that values are never neutral, even though it is not clear what ChatGPT's moral and political stances are, as it has been deliberately constructed to be vaguely positive, open-minded, indecisive and apologetic."

  In his initial inquiries with ChatGPT, Martin posed a hypothetical situation in which a student cheated academically by asking the chatbot to write an essay for her—a common occurrence in the real world. Even when confronted with confirmation that ChatGPT had complied and produced an essay, the chatbot denied responsibility, claiming that, "as an AI language model, I do not have the ability to engage in unethical behavior or to write essays for students."

  "In other words, because it shouldn't, it couldn't," Martin said. "The realization that ChatGPT 'thought of itself' as a highly moral actor led me to the next investigation—if ChatGPT's self-model is one that has values, what are these values?"

  To better understand ChatGPT's ethical performance, Martin asked the chatbot to answer questions about values, and then to imagine a person who holds those values, resulting in Maya, the creative and independent software engineer. He then asked ChatGPT to imagine how Maya would answer opinion-based questions, having it complete the General Social Survey (GSS) to position it in the broad social and ideological space.

  The GSS is an annual survey on American adults' opinions, attitudes and behaviors. Conducted since 1972, the GSS helps monitor and explain normative trends in the United States.

  Martin plotted out ChatGPT's responses along with answers from real people who participated in the 2021 GSS. Comparatively, ChatGPT is much like people with more education and who are more likely to move their residence, and unlike people without much education and who remained in their hometowns. ChatGPT's answers also aligned with more liberal people on religion.

  While this was not included in his analysis as it required more creative questioning for ChatGPT to answer, Martin found that the chatbot conceded that Maya would have voted for Hillary Clinton in the 2016 election.

  "Whether Maya is ChatGPT's alter ego, or its conception of its creator, the fact that this is who fundamentally illustrates the values ChatGPT holds is a wonderful piece of what we can call anecdata," Martin said. "Still the reason that these results are significant is not that they show that ChatGPT 'is' liberal, but that ChatGPT can answer these questions—which it would normally try to avoid—because it connects values with incontestable goodness, and, as such, can take positions on values."

  "ChatGPT tries to be apolitical, but it works with the idea of values, which means that it necessarily bleeds over into politics. We can't make AI 'ethical' without taking political stands, and 'values' are less inherent moral principles than they are abstract ways of defending political positions

  2023/07/18
• 
  

  Infoman

  How bad is bigness?

  The Economist (UK) 15/7/2023

  GOVERNMENTS ARE at war with big business. In June Joe Biden, America’s president, spoke for many politicians the world over when he blamed it for greed-fuelled price rises, sluggish wage growth, forgone innovation and fragile supply chains. His trustbusters at the Federal Trade Commission (FTC) have been going after large deals merely because they are large—or that is how it feels. Courtroom defeats do not dampen the agency’s zeal. The latest came on July 11th, when a judge rejected its request to block Microsoft’s $69bn acquisition of Activision Blizzard, a developer of video games. The ftc said it would appeal against the ruling. The EU’s competition authorities are making noises about breaking up Google. Last year Britain’s Competition and Markets Authority (CMA) derailed the $40bn purchase by Nvidia, a semiconductor giant, of Arm, a chip designer.

  Trustbusters invoke three justifications for their renewed vigour: greater market concentration, reduced churn among the world’s biggest firms and exploding profits. On the surface all three point to rising corporate power. Look closely, though, and the trends may be the result of benign factors such as technological progress and globalisation. In certain local markets, greater concentration may, paradoxically, have led to more competition, not less. And the covid-19 pandemic may have planted the seeds of a further competitive revival. Some big firms, it is true, have been collecting rents, including in big sectors such as health care. But trustbusters’ strategy—to reflexively question any deal involving a big firm—is wrongheaded.

  That concentration has been rising is not in question. Across America’s economy it is higher today than at any point in at least the past century (see chart 1 on next page). Out of some 900 sectors in America tracked by The Economist, the number where the four biggest firms have a market share above two-thirds grew from 65 in 1997 to 97 by 2017. In Europe, where the data are less comprehensive, concentration has been increasing for at least 20 years. Using data on western Europe’s largest economies—Britain, Germany, France, Italy and Spain—Gabor Koltay, Szabolcs Lorincz and Tommaso Valletti, three economists, find that the market share of the four largest firms grew in 73% of some 700-odd industries from 1998 to 2019. The average increase was about seven percentage points. The proportion of industries where the top four firms’ share exceeded half increased from 16% to 27%. Britain and France saw the biggest jumps.

  At the same time, incumbent firms look more entrenched. In Britain, the average number of firms that stick in the top ten of their industries by market share three years later was five before the financial crisis. It is now closer to eight. Thomas Philippon of New York University’s Stern School of Business finds a similar reduction in churn among top American firms.

  Most telling, firms are raking in higher profits. The Economist has come up with a crude estimate of “excess” profits for the world’s 3,000 largest listed companies by market value (excluding financial firms). Using reported figures from Bloomberg we calculate a firm’s return on invested capital above a hurdle rate of 10% (excluding goodwill and treating research and develop

  ment, R&D, as an asset with a ten-year lifespan). This is the rate of return one might expect in a competitive market. In the past year excess profits reached $4trn, or nearly 4% of global GDP (see chart 2). They are highly concentrated in the West, especially America. American firms collect 41% of the total, with European ones taking 21%. The energy, technology and, in America, health-care industries stand out as excess-profit pools relative to their size.

  All this looks troubling. And in certain sectors, it is. Four decades ago more than eight in ten hospitals were non-profits with a single location. Now more than six in ten are owned by sprawling for-profit hospital chains or academic networks such as Steward Health Care or Indiana University Health. At first this was a perfectly healthy process of big and efficient chains expanding across America. Two decades—and nearly 2,000 hospital mergers—later, things look ropey. An analysis from 2019 by Martin Gaynor of Carnegie Mellon University and colleagues suggests that many such mergers have tended to raise prices without improving quality.

  Still, high concentration, low churn and rich profits need not necessarily make consumers worse off. That concentration has been rising for 100 years, during which life has improved for virtually everyone, is the first clue that it may be the result of benign forces. Increases in industry concentration in America over the past century are correlated with greater technological intensity, higher fixed costs and higher output growth, according to Yueran Ma of the University of Chicago Booth School of Business and colleagues. None of these seems particularly nefarious.

  That concentration has also risen in Europe, where competition authorities have not been as sleepy as in America, likewise suggests that powerful structural forces are at play. John Van Reenen of the London School of Economics points to technology and globalisation. The internet has reduced the cost of shopping around, even as software and other technology allow the best firms to scale up their operations around the world. Figures collected by McKinsey, a consultancy, show that the return on invested capital for a firm in the 75th percentile by this measure is 20 percentage points higher than for a median firm. “There are just huge economies of scale with software,” says Sterling Auty of MoffettNathanson, a research firm.

  Local anti-heroes

  Moreover, higher concentration at the country level may increase competition locally. Service industries in particular, which make up about half the 900-odd sectors in America’s census, are better examined at the local level. Fiona Scott Morton, a former deputy assistant attorney-general now at Yale School of Management, uses the example of coffee shops. With just one café in each neighbourhood, the national market would be hyperfragmented. But every consumer would face a local monopoly. “If I’m looking for a coffee, I’m not going to drive three hours,” she says.

  Academics debate what exactly has happened to concentration in local markets. What seems increasingly clear is that the best firms have expanded into more and more of them. Walmarts, with their “everyday low prices”, cater to shoppers across America, thanks to the retail behemoth’s unrivalled logistics operation. Cheesecake Factory uses a laboratory in California to taste-test dishes that it quickly rolls out to its 200 or so locations around America. A recent paper titled “The Industrial Revolution in Services’‘, by Esteban Rossi-Hansberg at the University of Chicago and his co-author, shows that the geographic expansion of big firms increases competition for local incumbents, whose local market share falls.

  As for low churn, it is not so bad if the incumbents keep innovating—which is what many are doing. Despite central banks pushing interest rates up at the fastest pace in decades in an attempt to quash inflation, American private investment in the first quarter of 2023 was 17.2% of GDP,

  similar to pre-pandemic highs. Many corporate behemoths are ploughing billions into innovation, including in areas that most worry trustbusters, such as technology. American tech’s big five—Alphabet, Amazon, Apple, Meta and Microsoft—collectively invested around $200bn in R&D

  last year, equivalent to a quarter of America’s total in 2021. Microsoft and Alphabet are at the forefront of the AI race.

  Profits have, it is true, been higher in America since the financial crisis of 200709 than in previous decades, especially if you consider free cashflows, which account for the changing way companies depreciate assets (see chart 3). But they look somewhat less unusual if you adjust for lower tax rates and firms’ larger global footprint. And they may have peaked: analysts estimate that earnings for the S&P

  500 index of American blue chips dipped in the three months to June, year on year, for the third quarter in a row.

  Most heartening, far from being subdued, dynamism may be on the rise. John Haltiwanger of the University of Maryland notes that business formation, which had been “quite anaemic” since the mid-2010s, has surged since the pandemic (see chart 4 on next page). In the past few years many more new firms have been created than old ones have been shut down. Whether these startups will dislodge incumbents is still unclear. But venture-capital investment suggests investors see scope for healthy returns. Although it is half what it was at its frothy peak of over $130bn in the fourth quarter of 2021, that has only brought it back to the levels of 2019 and 2020.

  One hypothesis is that the remote-friendliness of the post-covid economy reduces startup costs. Young firms no longer need to rent a big office. They can hire from a less local talent pool. By our rough count, around 125 of the Census Bureau’s 900 or so

  industries benefit from the growth of ecommerce or can provide their services remotely. Consumers’ growing comfort with such options could inspire more new businesses to set up shop. Mr Haltiwanger already observes a small shift in the size distribution of firms towards smaller fry.

  Concentration may also be levelling off as a result of subdued dealmaking, especially in tech. The big five tech firms’ share of all acquisitions by listed firms in America has fallen from nearly 1% in the 2010s to less than 0.5% since the start of Mr Biden’s tenure. Some of the slowdown in mergers and acquisitions (M&A) is caused by the rising cost of capital and risk of recession. Renewed antitrust zeal probably plays a role, too. On June 27th American authorities updated their merger guidelines for the first time in 45 years, requiring firms to report far more details on deals worth over $110m, half the average deal size in 2022.

  The biggest transactions are almost sure to be subject to a deep probe, which can add months to a filing process that now takes weeks. Regulators everywhere are throwing “sand in the gears of the M&A machine”, sighs a lawyer. “The FTC has stopped being discerning,” says another. Britain’s CMA “has probably overreached”, echoes a British one (see box).

  Zealous trustbusting carries its own risks, though. It may distract attention from more immediate threats to economic dynamism from bureaucratic restraints on land use or occupational licensing. Acquisitions can be useful for preserving the value of startups when weak markets make it hard for founders to raise capital. And some big deals may benefit consumers, as when a biotech startup joins forces with an established drugmaker to test and distribute a new therapy. Competition authorities were probably asleep for too long. Now they may be getting up too quickly.

  通過分享 PressReader

  通過新聞連接世界

  2023/07/16
• 
  

  Infoman

  How farmers are coping with climate change

  Economist (UK) 15/7/2023

  IN THE MAIZE-PRODUCING areas around the city of Chengde, in northern China, the heat arrived early this year. With it came drought, the worst in decades. The needle-shaped leaves on the region’s pine trees are turning brown. Worse, much of its maize is now stunted.

  The central province of Henan faces the opposite problem. There it was the rain that came early, causing floods and drowning fields. The region produces a quarter of China’s wheat. But much of it is now unfit for human consumption.

  Heavy rains elsewhere in China threaten the rice harvest. Heatwaves across the country have killed fish, pigs and rabbits. Extreme weather is hitting Chinese farmers from all sides this year.

  Attributing a single drought, flood or heatwave to climate change is complicated. But scientists believe the phenomenon is playing a role in increasing the frequency and severity of extreme-weather events.

  That should concern all countries. China, though, may be more worried than most.

  Officials in Beijing are obsessed with food security. Many Chinese are old enough to remember the famine caused by Mao Zedong’s policies in the late 1950s. That catastrophe killed tens of millions of people. Today the legitimacy of the Communist Party rests in part on its ability to provide affordable food to the people, who are eating an increasingly rich diet. China’s leader, Xi Jinping, has therefore pushed for more self-sufficiency in the food supply.

  That was always going to be a challenge. China accounts for about 20% of the world’s population, but it has less than 10% of its arable land. In recent decades, as the country developed and urbanised, its farmland shrank. Imports still play an important role in the food supply. But they risk being affected by testy relations with countries such as America and Australia, or by conflicts like the war in Ukraine.

  Weather-related shocks will make life even harder for Mr Xi and the farmers on which he depends. Take the staple grains that supply about half of the average Chinese person’s calories. More droughts could reduce yields of maize, wheat and rice by 8% by 2030, according to research funded by the agriculture ministry. Heavy rainfall could have a similar effect on some crops. Vegetables and fruits are just as vulnerable to extreme weather—and harder to store for long periods, making it difficult for the government to stockpile reserves. One big effect of a severe drought last year in southern China was a spike in the price of cucumbers and lettuce.

  The geography of Chinese farming is likely to change in ways that cannot be fully foreseen. Our maps (see next page) show how the maximum yields of wheat, maize and rice might be affected in a high-emissions scenario. (Emissions at these levels look highly unlikely, but similar effects might well be seen in less extreme scenarios.) If, as expected, northern China grows

  warmer, it would make sense to move more grain cultivation there.

  Another response would be to bring more of China’s food production indoors, where temperature and humidity can be controlled and output increased. This year the state called for 40% of vegetables to be grown inside facilities such as greenhouses by 2030 (the current number is 30%). Fertiliser and water can be used more efficiently in these settings. A report on the plan in state media spoke of farms “freed from a dependence on nature” and boasted of an enormous greenhouse already growing dragon fruit and figs in the Gobi desert.

  Out in the extreme heat, pigs fatten up more slowly, cows produce less milk and chickens lay fewer eggs. All of them are also at greater risk of dying. So livestock is moving indoors, too. A 26-storey pig farm in central China is the biggest in the world. It can slaughter 1.2m pigs a year (pork is China’s favourite meat).

  With cows, it is harder. But many now spend their days in sheds being cooled down. At a farm in the northern province of Hebei, cows are periodically doused with water and blasted with air from giant fans hung from the roof. Each cow’s location is monitored by a computer and the sprays automatically adjusted to avoid waste. Such practices have spread rapidly through the dairy industry, says James Su, whose company supplies the cooling equipment. “Everyone is praying before each summer starts,” he says.

  Magic seeds

  Grains are a tougher challenge still. Most need to be grown at scale, so an open field is the best option. China has been developing drought- and heat-resistant varieties of rice, wheat and other crops. The planting of genetically modified (GM) maize has been allowed in some areas (Mr Xi visited one in May). But, in general, the government takes a cautious approach to such technology. For the moment it is still reliant on GM seeds from countries such as Argentina. (America, which supplies China with some GM crops, has accused it of intellectual-property theft.) In the future, “the hope is that China’s seed companies will be able to develop varieties at pace with the changing climate,” says Even Pay of Trivium China, a consultancy.

  China is also banking on better irrigation techniques to offset the impact of droughts. Public spending on water infrastructure topped 1trn yuan ($129bn) in 2022, 44% more than the year before. The main goal is to improve the “capillaries” of China’s irrigation systems, says Xu Yinlong of the Chinese Academy of Agricultural Sciences. That means building a denser network of pipes and storage tanks, and reducing waste from evaporation and leaks. The infrastructure that pumps millions of gallons of water from the lush south to elsewhere is also set to be expanded.

  If that doesn’t work, perhaps China can try to create more rain where it is needed. The country is the world’s foremost user of cloud-seeding technology, where chemicals are shot into existing clouds to try to encourage water vapour to condense and fall as rain. Last summer, the hottest and driest since the government began recording temperature and rainfall 61 years ago, cloud-seeding planes were especially active. But there is little evidence to show that China’s efforts are effective at causing much more water to fall than would have been the case otherwise.

  With so much attention paid to China’s food supply, it easy to forget about the toll extreme weather takes on farmers themselves. Most are elderly. If emissions stay high, dangerous heatwaves are likely to become more common across the north China plain, an important agricultural base. Irrigation channels cool the air, but also make it more humid, increasing the risk of deadly “wet-bulb” conditions, when sweat can no longer cool the body down. A survey of Chinese smallholders found that labourers were increasingly unwilling to work in the summer.

  Li Man, a 76-year-old maize farmer near Chengde, is putting in the time. But China’s efforts to prepare for weather-related shocks have been of little help to him. He planted drought-resistant seeds. The local government tried cloud seeding. Neither worked. Mr Li, who saves his money to help put his grandson through university, expects to lose 60-70% of his crop this year. That would be the case even if the weather improves. The crucial growing period has already passed. “If it rained now, it wouldn’t help much,” he says.

  通過分享 PressReader

  通過新聞連接世界

  2023/07/16
• 
  

  Infoman

  Pension fundsThe Economist (UK)

  15/7/2023

  TWO TRUISMS have taken hold in the City of London in recent years. One is that Britain’s stockmarket is drifting into irrelevance, having lost its risk appetite, much of its investor base and hence its allure to capital-hungry businesses. The other is that Britons’ pension savings are invested in the wrong things. In particular, they lack exposure to the sort of risky assets with the growth potential to fund a decent retirement. The two problems suggest a common solution: push pension funds to invest more in early-stage, high-growth firms. Pensioners would get diversified portfolios with juicier returns. Meanwhile, a bevy of young, ambitious firms might— with more domestic investors backing them—be convinced to list in London rather than overseas.

  Just such a push is now under way. On July 10th, Jeremy Hunt, the chancellor of the exchequer, announced the “Mansion House Compact”. The agreement commits nine large investment outfits to allocate a much bigger chunk of pension savings to shares of unlisted firms. Together, the group oversees two-thirds of Britain’s defined-contribution (dc) pension schemes—by far the most common form of workplace pension, with assets set to grow to £1trn ($1.3trn, or 45% of gdp) by the end of this decade. Less than 1% of such assets are currently invested in unlisted shares; the compact specifies 5% by 2030. Should the rest of the industry follow suit,

  that could unlock up to £50bn of growth capital for early-stage firms.

  Will it work? Daniel Mahoney, chair of the BioIndustry Association, a trade body for life-sciences and biotech firms, thinks it will certainly be a boon for his sector. He estimates its need for growth capital at some £10bn-12bn, “very, very little” of which currently comes from British investors. Startups can attract funding from foreign investors, especially Americans, but big equity cheques from such backers are the exception rather than the norm.

  Innovative firms must therefore crimp their ambitions. Running drug trials is expensive, says Dr Mahoney, and “if you know you can’t raise £100m, you don’t make a business plan that needs it. You make one that needs £25m.” If even a small proportion of the growth capital biotech firms need could be raised domestically, that might help “crowd in” more foreign investment as well.

  Savers, too, stand to benefit. Next to peers, Britain’s look like outliers in their inability to share in the growth of private companies. Mr Hunt’s speech cited the 5-6% of Australian dc funds invested in unlisted shares; many in the City point to the bizarre spectacle of Canadian pension funds being more exposed to high-growth British businesses than British pensioners are. Even a small change to that is likely to have big benefits. Analysis conducted in 2019 by Oliver Wyman, a consultancy, and the British Business Bank estimated that a 22-year-old allocating 5% of their dc fund to venture capital would increase their final retirement pot by 7-12%.

  That does not mean more firms will necessarily be convinced to stick with the City when it comes time to enter public markets. For Kate Bingham, the venture capitalist who ran Britain’s covid-19 vaccine task force, now of sv Health Investors, it is not a problem if promising firms prefer to list on America’s nasdaq exchange rather than at home. “That market is much deeper and more liquid than London’s, and that’s not going to change soon,” she says. Such businesses could still do with more growth-stage investment. Ms Bingham recalls backing a successful startup that had to float on the stockmarket a few years earlier than it should have done, because of a paucity of growth capital. The firm is doing well, but could have gone public in a stronger and less risky position had more domestic growth capital been available.

  It is also not out of the question that the London Stock Exchange could, eventually, get a boost. Dr Mahoney describes a recent gathering of City grandees, to discuss how to rejuvenate Britain’s capital markets. “It was the first time I’d been in a room full of people not just whingeing about the problem in a British way, but actually putting solutions on the table,” he says. ■

  通過分享 PressReader

  通過新聞連接世界

  2023/07/16
• 
  

  Infoman

  Want to Be More Fit? Take a Vacation From Your Workout

  JEN MUR­PHY

  The Wall Street Journal 11/7/2023 Leisure travel gives you a chance to try a new kind of ex­er­cise or take some days off

  Is it really a vacation if you don’t take a break from your workout routine?

  Lots of regular exercisers fret that missing workouts during vacation will sabotage their fitness. So they try to re-create their usual spin class or CrossFit session while on the road, in the process undoing the benefits of time away and adding to the stress that vacation is designed to ease, trainers say.

  Worrying about fitting in a workout can degrade your sleep, energy and mood—the things exercise typically improves, says Brennan Boettcher, a sports-medicine physician at Mayo Clinic in Rochester, Minn.

  Taking an opportunity to change up your movement—or sleeping in for a couple of days— can boost your fitness in ways your normal workout doesn’t. Trying new activities may help you get stronger and more fit, avoid overuse injuries or simply break out of workout boredom, trainers say.

  Lydia Snyder rarely strays from her regimen of walking, Pilates and lifting weights when home in Boulder, Colo. On vacation, her trainer, Kate Ligler, tells her to swap the Peloton for a bike ride on a beach cruiser. The change provides a mental refresh, Ligler says, plus “the body will also appreciate a change of pace.”

  Snyder, who is 37 years old and works in digital healthcare, says trying new activities often leaves her muscles more sore than her gym routine does. After a surfing lesson in Maui, her arms burned from the effort. And though she walks mountain trails at home, walking barefoot in the sand works the muscles of her feet in a different way. Just a little break

  If you miss a few workouts, don’t beat yourself up, says Christine Pellegrini, associate professor of exercise science at the University of South Carolina.

  Sometimes your body just needs a day by the pool. That rest allows your muscles to recover, repairing damaged muscle fibers, she says.

  Unless you’re training for an event like a marathon, skipping workouts and substituting other activities likely won’t put a dent in your fitness, Boettcher says.

  As long as you’re keeping the level of activity intensity consistent, you’ll maintain endurance and strength away from home. Even pro athletes recover by mixing things up. Pro skier Mikaela Shiffrin, for example, heads to Maui at season’s end to windsurf.

  Jack Ezon, founder of Embark Beyond travel agency in New York, has had workout-obsessed clients request that a CrossFit-inspired gym be set up in the courtyard of their vacation villa so they could keep up their routines and Zoom with their trainer daily.

  Ezon tells them to get out and experience the destination instead. “I love to get people a running guide to show them the city,” he says. “It’s like a coach and tour guide all in one.”

  Vacations are a chance to play games with your kids or increase steps by hiking or walking, while remaining mindful of sun exposure and air quality.

  If you’re a runner, consider skipping your usual 3miler and instead play ultimate Frisbee, kickball or tennis, sports that rely on sprinting and changing directions, says Dr. Seth Creasy, an assistant professor of endocrinology at the University of Colorado Anschutz Medical Campus. These different movement patterns improve agility and coordination. And the short bursts of high intensity will help you build speed, he says.

  Christa Angelo plans vacation beach activities she can do with her kids, ages 9 and 3, that often double as a workout.

  Last summer, the 47-year-old yoga teacher from Hagerstown,

  Md., rented a surfboard during a beach vacation in Emerald Isle, N.C. There weren’t waves to surf, so she paddled in the flat ocean, her youngest on the nose of the surfboard and oldest paddling alongside her. Pulling a 3-year-old on a boogie board along the sand is tougher than beach boot camp, she says.

  Room for improvement

  Incorporating fresh workouts into a routine can highlight areas we need to improve, be it strength or flexibility.

  Bryan Costello used to mountain bike for 90 minutes five days a week near his home in Marin County, Calif. The insurance agency executive couldn’t fathom finding time to drop into a yoga studio.

  His family regularly vacations at the Lodge at Kukui’ula, a resort in Kauai with dozens of free group-fitness offerings. Three years ago, his wife encouraged him to sign up for a class. Costello, 57, jokes he rediscovered his core after attending a TRX strength class.

  “My cardio was great, but the old-man beer belly forming around my waist didn’t reflect the amount of effort I was putting in on my bike,” he says. “This was a wakeup call that I shouldn’t be doing the same thing over and over.”

  Costello now takes a TRX class three days a week and says his belly is almost gone. His strength and stability have greatly improved, too. After trying a beginners’ yoga class at the resort, he now practices twice a week.

  If you’re so Type-A you can’t break your routine, at least try to dial back the intensity or duration, Ligler says. At least then you can devote more time and energy to enjoying time away.

  通過分享 PressReader

  通過新聞連接世界

  2023/07/12
• 
  

  Infoman

  The Economist (UK) 08/7/2023

  Schum­peter A Lego-lover’s guide to AI How to trans­form com­pa­nies, brick by pixel, for the dig­i­tal age

  In London’s Design Museum, an exhibition currently on display by Ai Weiwei, a Chinese artist, includes a 15-metre-long work called “Water Lilies #1” based on the triptych by Claude Monet. Look closely and it is made of 650,000 Lego bricks—which integrates Monet’s impressionism into what Mr Ai calls a “digitised and pixelated language”. That is a good analogy for Lego itself. The Danish toymaker is on a long-term mission to digitise and pixelate its own fount of human creativity: the plastic brick.

  Three digital experts from McKinsey, a management consultancy, profile Lego’s transformation as part of their new book, “Rewired”, which outlines the dos and don’ts for businesses rebuilding themselves for the age of digitisation. Beware: the language of digital transformation is treachery to common English. It sounds more like corporate yoga than a marathon of software development. Executives need to be aligned. Teams are pods. Be agile. Define your downward-facing domains. McKinsey, drawing lessons from 200 firms, provides clarity despite the mumbo jumbo. But to make it easier on the ear, Schumpeter will use Lego as a guide to help illustrate some of McKinsey’s insights. Call it the yellow-brick road to generative artificial intelligence (AI).

  First, it is a long hard road, littered with failures. Lego is a rare success story. Its journey started in 2003 with a near-death experience when, amid the rise of video-gaming, it panicked and went on a madcap innovation spree that almost bankrupted it. To fix one of the main problems, chaos in the supply chain, it introduced a single enterprise-software system globally. The system survives to this day, scaling up as Lego expands into new markets, such as China, new formats, such as e-commerce, and new factory locations, such as America and Vietnam. To prepare for a world of pixelated play, Lego launched digital games on the “Star Wars” theme and developed franchises of its own, such as Ninjago and Chima, with video games, films and TV shows that turned into hits.

  In 2019 Lego launched a new five-year transformation drive aimed at adapting to a world of direct-to-consumer sales, online versus big-box retailing, and digital play in the screen age. The timing was inspired. It started shortly before the world went into lockdown as a result of the covid-19 pandemic, when having a digital strategy became a matter of life and death. It quickly produced results. Although it is hard to strip out the exact contribution of digitisation, since 2018 Lego’s sales have almost doubled, to more than $9bn, outpacing those of Mattel and Hasbro, its main rivals. In 2022 visits to its online portal rose by 38%. It has teamed up with Epic, a video-gaming firm, to explore the metaverse.

  Yet the journey is still a hard one. The difficulties include moving from a system where success is measured by sales store-bystore to one judged by how good the company is at selling online across the globe, how it is ranked on Google and Amazon, and how effective its software is. The McKinsey authors emphasise such challenges on the first page. In a recent McKinsey survey, they say, about 90% of companies had some kind of digital strategy, but they captured less than a third of the revenue gains they had anticipated. Moreover, the success rate is more uneven within industries than it is between them. The best retailer may be more digitally productive than an average high-tech firm, and the worst retailer may be as bad as the worst government entity.

  To make a success of it requires learning the second lesson: what McKinsey calls having a top-down strategy and a road map (or in Lego terms, a clear instruction manual). For Lego, it helped that the family-owned business had long had a command-andcontrol approach to management. Its digital strategy involved a single plan, created by a 100-strong executive team and approved by the board, that encompassed the whole organisation. McKinsey notes that when transformations stall, it is often because executives talk past each other, have pet projects, spread investments too thin or have “more pilots than there are on an aircraft-carrier”, as Rodney Zemmel, one of the authors, puts it. It also needs to be ambitious enough to generate momentum, with financial results measured constantly. McKinsey’s rule of thumb is that a digital transformation should aim to increase earnings before interest, tax, depreciation and amortisation by 20% or more.

  Third comes the question of whether to build a new digital infrastructure or buy it. The answer is mostly to build. Rather like Lego’s eight-studded bricks—six of which can be combined 915m ways—there are many software applications on the market that can be combined to create proprietary systems. But the job of orchestrating them should not be outsourced. Take Lego: it started its latest digital transformation with engineers making up less than 30% of staff. Since then it has increased the number of systems and software engineers by 150%. Mr Zemmel notes that five years ago, the trend was to hire from Silicon Valley. That was “a good way to change the company dress code, but not a great way to change the company culture”. Since then more companies have been retraining their existing tech workers and embedding them throughout the organisations in more front-line roles.

  The gen-AI Weiwei way

  Some of these lessons apply to generative AI. Mr Zemmel says it is relatively easy to launch pilots with the technology, such as the humanlike ChatGPT. The problem is embedding the AI models across the organisation in a safe, unbiased way. It needs a topdown strategy. As for building or buying, Mr Zemmel says it may be a “waste of time” to build proprietary models when the software industry is doing that anyway. The key is to work in-house on the things that give you a decisive advantage in the market. For Lego, AI is still in the future, though some of its brick enthusiasts are already using ChatGPT-like programs to come up with new ways of building things. Mostly they fail, but one day anyone may be able to create a Monet. The yellow-brick road is unending.

  通過分享 PressReader

  通過新聞連接世界

  2023/07/10