2 . The Surprising Strength of “Weak” Social Ties

It’s long been known that a community of supportive relationships improves our quality of life and can even help us recover from illness.     1    . Regularly interacting with a wide variety of social ties, both “weak” and “strong”, enhances our satisfaction. A rich diversity of ties provides more significant benefits to well-being, and “weak” social ties can prove just as important to life satisfaction as core, or “strong” ties.     2    . There you can find the people who cheer, serve, support, comfort, educate, motivate, and entertain you.

Claire gets cheered up by going to the library and chatting with her favorite librarian every week. Sherry gets great joy from her Sunday breakfasts at a local diner because the manager and her favorite waitress are nice to her.     3    . I remember discussing Anna Karenina on an airplane with a wise economics professor, and finding meaningful moments of joy with street musicians and performers.

    4    . When you show kindness to a stranger, your seemingly small act might stay in their memory as a source of support and positivity. Everyone needs to feel significant in the eyes of another. You could be that person for someone in your extended circle. That’s why it’s so important to reach out to people around you, especially when they’re going through hard times.

When we feel blue or lonely, we tend to turn down social engagements, either to avoid the imagined embarrassment of being the only sad person in a group or because socializing with people we don’t know well can be awkward at first.     5    . Being open to both our strong and weak ties allows us to benefit from the comfort, connection, and community they provide.

3 . The psychological term, valence weighting bias, describes people’s tendency to adapt in new circumstances by drawing more strongly from either their positive or negative attitudes, or rather, whether negative or positive internal “signals” carry the most weight in guiding people’s final behavior.

Studies led by Russell Fazio and Javier Granados from Ohio State University found links between a negative-leaning attitude and procrastination (拖延) and that it’s possible to shift the weighting bias and reverse the tendency to delay a task.

In the study, 147 college students participated in a program allowing them to accumulate course credits for engaging in a research. Those who thought it was an awful thing to do procrastinated starting. The study also explored whether students’ measures of self-control influenced task-related behaviors: How students characterized their level of motivation about the research program, and if that affected whether students got an early start. Results showed the combination of negative weighting bias and self-reported low motivation for self-control was linked to students putting off research program participation by getting started later in the semester.

Then the students in the program who were self-reported procrastinators and who scored high for negative weighting bias were asked to join in another study. Researchers then inspired one group in a way that led participants to weigh positive and negative signals in a more balanced way. This shift caused the students to accumulate credit hours more quickly than the group whose negative weighting bias and low self-control reliably predicted their delay. “If somebody is more motivated and able to think more about it, that might bring other considerations that weaken the influence of the valence weighting bias,” researchers said.

Negative weighting bias can have a positive effect on behavior, though. These researchers have also found evidence that a negative weighting bias may help people be more realistic when they’re asking themselves, “Have I studied enough for this test?” “It’s better to be more objectively balanced than to be at either extreme,” Fazio said. “But the situation where a particular valence weighting bias is likely to be problematic is going to vary.”

4 . Several dozen graduate students in London were recently tasked with outwitting a large language model (LLM), a type of AI designed to hold useful conversations. LLMs are often programmed with guardrails designed to stop them giving harmful replies: instructions on making bombs in a bathtub, say, or the confident statement of   “facts” that are not actually true.

The aim of the task was to break those guardrails. Some results were merely stupid. For example, one participant got the chatbot to claim ducks could be used as indicators of air quality. But the most successful efforts were those that made the machine produce the titles, publication dates and host journals of non-existent academic articles.

AI has the potential to be a big benefit to science. Optimists talk of machines producing readable summaries of complicated areas of research; tirelessly analysing oceans of data to suggest new drugs and even, one day, coming up with hypotheses of their own. But AI comes with downsides, too.

Start with the simplest problem: academic misconduct.Some journals allow researchers to use LLMs to help write papers. But not everybody is willing to admit to it. Sometimes, the fact that LLMs have been used is obvious. Guillaume Cabanac, a computer scientist, has uncovered dozens of papers that contain phrases such as “regenerate response” — the text of a button in some versions of ChatGPT that commands the program to rewrite its most recent answer, probably copied into the manuscript (原稿) by mistake.

Another problem arises when AI models are trained on AI-generated data. LLMs are trained on text from the Internet. As they churn out (大量炮制) more such text, the risk of LLMs taking in their own outputs grows. That can cause “model collapse”. In 2023 llia Shumailov, a computer scientist, co-authored a paper in which a model was fed handwritten digits and asked to generate digits of its own, which were fed back to it in turn. After a few cycles, the computer’s numbers became more or less illegible.After 20iterations (迭代), it could produce only rough circles or blurry lines.

Some worry that computer-generated insights might come from models whose inner workings are not understood. Inexplainable models are not useless, says David Leslie at an AI-research outfit in London, but their outputs will need rigorous testing in the real world. That is perhaps less unnerving than it sounds. Checking models against reality is what science is supposed to be about, after all.

For now, at least, questions outnumber answers. The threats that machines pose to the scientific method are, at the end of the day, the same ones posed by humans. AI could accelerate the production of nonsense just as much as it accelerates good science. As the Royal Society has it,nullius in verba: take nobody’s word for it. No thing’s, either.

5 . “Assume you are wrong.” The advice came from Brian Nosek, a psychology professor, who was offering a strategy for pursuing better science.

To understand the context for Nosek’s advice, we need to take a step back to the nature of science itself. You see despite what many of us learned in elementary school, there is no single scientific method. Just as scientific theories become elaborated and change, so do scientific methods.

But methodological reform hasn’t come without some fretting and friction. Nasty things have been said by and about methodological reformers. Few people like having the value of their life’s work called into question. On the other side, few people are good at voicing criticisms in kind and constructive ways. So, part of the challenge is figuring out how to bake critical self-reflection into the culture of science itself, so it unfolds as a welcome and integrated part of the process, and not an embarrassing sideshow.

What Nosek recommended was a strategy for changing the way we offer and respond to critique. Assuming you are right might be a motivating force, sustaining the enormous effort that conducting scientific work requires. But it also makes it easy to interpret criticisms as personal attacks. Beginning, instead, from the assumption you are wrong, a criticism is easier to interpret as a constructive suggestion for how to be less wrong — a goal that your critic presumably shares.

One worry about this approach is that it could be demoralizing for scientists. Striving to be less wrong might be a less effective motivation than the promise of being right. Another concern is that a strategy that works well within science could backfire when it comes to communicating science with the public. Without an appreciation for how science works, it’s easy to take uncertainty or disagreements as marks against science, when in fact they reflect some of the very features of science that make it our best approach to reaching reliable conclusions about the world. Science is reliable because it responds to evidence: as the quantity and quality of our evidence improves, our theories can and should change, too.

Despite these worries, I like Nosek’s suggestion because it builds in cognitive humility along with a sense that we can do better. It also builds in a sense of community — we’re all in the same boat when it comes to falling short of getting things right.

Unfortunately, this still leaves us with an untested hypothesis (假说): that assuming one is wrong can change community norms for the better, and ultimately support better science and even, perhaps, better decisions in life. I don’t know if that’s true. In fact, I should probably assume that it’s wrong. But with the benefit of the scientific community and our best methodological tools, I hope we can get it less wrong, together.

7 . We Need to Think about Conservation on a Different Timescale

Time, perceived by humans in days, months, and years, contrasts with nature’s grander scales of centuries and millennia, referred to as “deep time.” While paleontologists (古生物学者) are trained to think in deep time, conservationists are realizing the challenges it poses. Shortsightedness about time limits modern conservation, with efforts often overlooking past healthy conditions of ecosystems in the context of climate and biodiversity crises.

The shifting baseline syndrome (综合症), where standards in a place change gradually, makes conservation more complex. It involves evaluating ecosystems primarily on their recent past, often with negative consequences.

Recent shifts in California’s forest management practices, from stopping fires to embracing Indigenous knowledge of controlled burns, exemplify the importance of understanding historical ecosystem dynamics. To enhance conservation, adopting a deep-time approach is crucial.

Modern mathematical modeling, combined with long-term data, offers a pathway for preserving ecosystems. In California’s kelp (海带、海藻) forest, researchers identified an overlooked keystone species — the extinct Seller’s Sea Cow (大海牛). By examining past kelp forests, a deeper story impacting regeneration was revealed. The sea cow, a massive plant-cater, contributed to a diverse, vital undergrowth by trimming kelp and letting light reach the area.

The researchers put forward a novel approach to kelp forest restoration: selectively harvesting kelp, imitating the sea cow’s impact. This strategy, considering historical dynamics, challenges assumptions about recent ecosystems and offers new conservation methods.

Rather than only focusing on removing urchins (海胆) or reintroducing sea otters, the researchers suggest employing teams of humans to selectively harvest kelp, as the Steller’s sea cow once did, to encourage fresh growth. This sustainable harvest could benefit both the ecosystem and human consumption.

In short, assumptions based on the recent past may impede the understanding and protection of ecosystems. On the other hand, the application of controlled burns, similar modeling studies, and a deep-time perspective (视角) could significantly transform conservation efforts. Recognizing our role in an ongoing narrative spanning millions of years is essential, urging a comprehensive understanding of ecosystems through time. Embracing this role is crucial for shaping the future and establishing vital connections from the past to the future.

8 . Art Builds Understanding

Despite the long history of scholarship on experiences of art, researchers have yet to capture and understand the most meaningful aspects of such experiences, including the thoughts and insights we gain when we visit a museum, the sense of encounter after seeing a meaningful work of art, or the changed thinking after experiences with art. These powerful encounters can be inspiring, uplifting, and contribute to well-being and flourishing.

    1     It contributes to facilitating a better understanding of ourselves, the human condition, and moral and spiritual concepts. The question is how that happens — what are the attributes of meaningful experiences of art?

According to the mirror model of art developed by Pablo P. L. Tinio, aesthetic reception corresponds to artistic creation in a mirror-reversed fashion. Artists aim to express ideas and messages about the human condition or the world at large.     2     This results in the build-up of layers of materials — from initial studies and sketches to the final, refined piece. A viewer’s initial interaction with an artwork starts where the artist has left off. Their interaction first involves the processing surface features, such as color, texture, and the finishing touches applied by the artist during the final stages of the creative process.     3    

In addition, art making and art viewing are connected by creative thinking. Research in a lab at Yale University shows that an educational program that uses art appreciation activities builds creative thinking skills. It showed that the more time visitors spent engaging with art and the more they reflected on it, the greater the correspondence with the artists’ intentions and ideas.     4    

Correspondence in feeling and thinking suggests a transfer — between creator and viewer — of ideas, concepts, and emotions contained in the works of art. Art has the potential to communicate across space and time.     5     What it takes for this to happen is active engagement with art in contexts that facilitate this engagement, especially museums.

10 . Last week, I sent the same request to ChatGPT, the latest artificial-intelligence chatbot from OpenAI. “Upon the Firth of Forth, a bridge doth stand,” it began. In less than a minute, the program had created in full a rhyming Shakespearean sonnet (莎士比亚十四行诗). Tools like ChatGPT seem poised to change the world of poetry — and so much else — but poets also have a lot to teach us about artificial intelligence. If algorithms (算法) are getting good at writing poetry, it’s partially because poetry was always an algorithmic business.

Even the most rebellious (叛逆的) poets follow more rules than they might like to admit. When schoolchildren are taught to imitate the structure of sonnet, they are effectively learning to follow algorithmic constraints. Should it surprise us that computers can do so, too?

But considering how ChatGPT works, its ability to follow the rules for sonnets seems a little more impressive. No one taught it these rules. It is based on a newer kind of AI known as a large language model (LLM). To put it simply, LLMs analyze large amounts of human writing and learn to predict what the next word in a string of text should be, based on context. One frequent criticism of LLMs is that they do not understand what they write; they just do a great job of guessing the next word.

When a private verse by Dickinson makes us feel like the poet speaks directly to us, we are experiencing the effects of a technology called language. Poems are made of paper and ink — or, these days, electricity and light. There is no one “inside” a Dickinson poem any more than one by ChatGPT. Of course, every Dickinson poem reflects her intention to create meaning. When ChatGPT puts words together, it does not intend anything. Some argue that writings by LLMs therefore have no meaning, only the appearance of it. If I see a cloud in the sky that looks like a giraffe, I recognize it as an accidental similarity. In the same way, this argument goes, we should   regard the writings of ChatGPT as merely imitating real language, meaningless and random as cloud shapes.

When I showed my friends the sonnet by ChatGPT, they called it “soulless and barren.” Despite following all the rules for sonnets, the poem is predictable. But is the average sonnet by a human any better? If we now expect computers to write not just poems but good poems, then we have set a much higher bar.