"This is the central assumption of cognitive science and represents an orthodox view within psychology and neuroscience."

"Instead, it is because computer science has discovered that computation is a relatively abstract property, and many completely different physical devices share the same constraints and powers."

"The point isn’t to elide any difference between the digital computers we manufacture and the biological computers we’re born with."

"Instead, it’s to recognize that if you wanted to understand the heart but were somehow philosophically opposed to thinking of it as a kind of pump, you would have a hard time making sense of how it works."

"The brain is far more complicated."

"The brain is massively parallel and interconnected, having as many as a quadrillion synapses."

"Bodies have sensory inputs and motor outputs that interact with our environment."

"Computers are logical; humans are emotional."

"Computers aren’t sentient; humans have consciousness"

"Computers are rule-governed; human thinking is holistic."

"Learning, it seems, is optimized for both humans and machines when we succeed around 85% of the time."

"For all of these stochastic gradient-descent based learning algorithms, we find that the optimal error rate for training is around 15.87% or, conversely, that the optimal training accuracy is about 85%."

"These findings agree closely with the 80% success rate found by Barak Rosenshine in his study of successful classrooms, despite coming from a completely different theoretical background:"

"The 85% rule suggests that you should fine-tune the amount of support you use depending on the success rate you’re experiencing."

"The exact percentage may vary between tasks."

"There are many theories of optimal learning that all point to a sweet spot for difficulty—not too easy, not too hard."

"Lev Vygotsky’s zone of proximal development argues that tasks slightly beyond what we can do by ourselves, but can do with assistance from others, maximize learning."

"when given a test that required inference or problem solving, the students with higher background knowledge did better than students with less background knowledge when interpreting the less coherent text."

"These results fit a model where, if a text mostly says things you easily understand, you don’t invest much effort into creating a mental model of what the text describes."

"In those cases, greater difficulty might be beneficial since the struggle it creates forces you to retrieve background knowledge."

"Anders Ericsson’s model of deliberate practice argues that when skills become automatic, we plateau at levels of ability far below our potential."

"Robert Eisenberg’s theory of learned industriousness suggests difficulty plays a role in motivation as well."

"The two groups thus experienced the same success rate, but had very different expectations about the role effort played in success."

"Learned industriousness suggests that success on hard problems can be good for us, but failure is demotivating."

"Everyone wants advanced knowledge. But it’s typically the basics that matter most."

"Most math is arithmetic, not algebra. Running a business is largely a matter of keeping revenues above costs. Good writing is the product of clear ideas and clean sentences."

"Mastery of the performance results from mastery of the parts."

"We ignore the basics, not because they’re hidden, but because they’re so obvious."

"Bringing attention back to the basics doesn’t come automatically. We adjust our habits only when they fail to deliver results."

"One way to return to the basics is to change the environment."

"Another strategy is to change your goals. When you set a different standard for what you want to produce, your actions must adapt."

"Finally, you can find room for improvement by observing and analyzing your work after the fact."

"So separate these tasks–examine your mistakes when you’re not in the middle of making them."

"Part of the difficulty is that ChatGPT’s human-like conversation abilities can be deceptive."

"The metacognitive ability to know what you don’t know is underdeveloped in these applications."

"Simon Willison suggests thinking of LLMs as a “calculator for words”—something that can do useful things with text—rather than as a general-purpose intelligence or smart person."

"1. Create your own Socratic tutor."

"By far, the most common use readers reported was using an LLM as a personal tutor."

"2. Practice chatting in new languages."

"3. Generate summaries of longer texts."

"4. Dialog with long documents."

"For instance, when reading a scientific paper you could quickly query the sample size or ask for the methodology or results."

"5. Rewrite texts at different levels of explanatory depth."

"One is simply asking an LLM to explain a popular concept in simpler terms, such as: “Explain quantum computing like I’m an eighth-grader.”"

"The other is to provide ChatGPT with a text or explanation and ask the AI to rewrite it in a more digestible manner."

"6. Clear up unfamiliar jargon."

"7. Create study plans and agendas."

"For instance, some readers asked ChatGPT to break down a complex learning goal and give them a curriculum."

"Others preferred to go even further, asking for ChatGPT to create a studying schedule for them, given their constraints for the day."

"While skill breakdowns might be alright, ChatGPT still struggles with creating reading lists, confabulating books and references. Thus, while it might do well for decomposing an ambiguous learning task, I wouldn’t trust it to give me good resources (yet)."

"8. Provide refreshers on forgotten or infrequently used tools."

"Programmers were the biggest professional group to reply to my query."

"9. Generate flashcards based on text. (Tentative)"

"Nonetheless, making flashcards is tedious, so getting a first draft that I later review might speed up the process considerably."

"10. Use it to organize your notes. (Advanced)"

"1. Don’t expect AI to get facts right."

"When Wikipedia was released, for instance, “experts” were in an uproar about how the user-generated nature of the website meant it couldn’t ever be relied upon as a source."

"2. Don’t expect AI to get citations right."

"3. Don’t expect AI to get the math right."

"One finding from psychology is that much of reasoning is performed by different subsystems in the brain than those devoted to language."

"As such, LLMs are really bad at math. And not just higher-level math that humans struggle with."

"I suspect a major cause of the difficulty is that we readily conflate two different ideas of what it means to “be productive.”"

"The first is a subjective feeling of productivity. This comes when we’re super busy or exerting extra willpower to focus on our work."

"The second is the objective output of productivity. This is the actual amount of important work that gets done."

"In the short term, the subjective and objective views of productivity often coincide."

"However, in the big picture, subjective and objective views of productivity often have nothing to do with each other."

"Sustainable Productivity is Invisible Productivity"

"Sustainable productivity comes from decoupling the short-term feeling of being busy or exerting effort from the long-term investments that allow you to get more work done whether or not you “feel” particularly productive."

"Knowledge and skills. Whether something is a vexing problem or a trivial task depends on the skills and experience you’ve accumulated."

"But most important is recognizing that in order to become more productive, you have to worry less about feeling productive and more about creating systems that allow you to get more done—without herculean effort."

"Pick any author at random, and on average the papers they publish earlier in their career, whether as first author or last author, will be more highly cited and cited by a more diverse group of fields, than a paper they publish later in their career."

"Does the spark of youthful genius burn out quickly, explaining the mediocrity of researchers in their later years? Interestingly, the answer is no:"

"But this is not what Liu and coauthors find! Instead, they find that any paper written, at any stage in your career, has about an equal probability of being your top cited paper!"

"More recently, Dean Simonton has analyzed the creative output of individuals across many domains and suggests an “equal-odds” rule best describes it: once a creative individual starts publishing in a field, each piece of work they produce has roughly equal odds of world-breaking impact."

"Instead, it looks like the most important determinant of creative success is simply how much work you produce."

"Be at a knowledge frontier. You can’t contribute anything new if you’re not at the boundary of knowledge for a discipline."

"Idea generation and public reception are stochastic processes. Once you reach the threshold, further advances have a significant random component."

"The randomness of creative success favors those who are the most prolific. Price’s Law captures this relationship in scientific output, estimating that half of the research of a given discipline will be produced by the square root of the number of researchers."

"To Produce Better Work, Increase Your Output"

"Yet perhaps it’s because we’re so uncomfortable likening creativity to a lottery that this perspective is undervalued."

"Over a surprisingly wide range of pursuits, creativity is productivity, and we will have more hits if we take more swings."

"Surprisingly, the answer seems to be that more intensive language education programs do better!"

"The results from our analyses suggest that concentrating the hours of English instruction in shorter periods of time is more beneficial for the students’ learning than distributing them in many months."

"1. Spacing makes learning harder, sometimes too hard."

"However, added complexity can sometimes flip the expected benefit of these desirable difficulties."

"These results suggest that the optimal studying technique for complex skills changes over time, starting with massed/blocked/review, and shifting to spaced/interleaved/retrieved presentation."

"2. Spacing of individual items is independent of overall classroom intensiveness."

"It’s possible that in the time frames studied, individual items of vocabulary and grammar are being spaced out just fine in the intensive programs, so the efficiency gains of the extensive schedule are not observed."

"3. Intensive classes are more motivating (or attract more motivated students)"

"In arguing for my intensive language immersion bursts, I similarly relied on a non-cognitive argument."

"4. Perhaps the tests weren’t delayed enough."

"A well-known consequence of the spacing effect is that massed presentations tend to be more effective initially, but their advantage typically fades over time."

"It appears that studies of actual classroom language learning seem to weigh more in favor of the approach I’ve intuitively preferred—working through an intensive burst to get to semi-conversational and then following that with periodic review, rather than studying in twenty-minute daily sessions from the get-go."

"Spacing for individual items is still central."

"Intensive projects should be followed by more leisurely maintenance."

"It also fits with my informal observation that people are much more reluctant to pick up new skills or topics as they get older."

"This is a worrying trend. Learning is an integral part of the good life, so if forces make it harder (or less desirable) to learn, it seems like it would be helpful to understand them."

"Theory 1: Opportunity Costs and Investment Horizons"

"The first explanation is economic: learning is an investment. As you invest in some skills, but not others, you get a greater return from activities where you have considerable training. Thus, your opportunity cost for learning new things increases."

"Therefore, a failure to learn new things is perfectly rational, even if it can result in inflexibility as we get older."

"Another way this can occur is if investment in one skill makes time spent on new skills comparatively less attractive."

"Theory 2: We’re Too Busy"

"Learning new things takes time. While efficiency can help, the time cost for even well-designed learning may be prohibitive for busy professionals with kids."

"Theory 3: Older Minds Can’t Learn as Well"

"There is a grain of truth to this. Fluid intelligence peaks in your early twenties and declines afterward. However, the decline is gradual and minimal. The research I’ve seen seems to indicate that it’s only a serious problem in advanced age."

"Older people are usually more responsible and organized, which assists in sticking to a project. Also, accumulated knowledge can make learning easier."

"Reduce the effort for learning. Always carry a book with you."

"Set aside time for experiments. Straightforward calculus suggests that doing what you’re already good at has the best payoff."

"Shift your focus away from skills that involve quick wits to those that rely on accumulated knowledge."

"Above all, however, I think losing interest in learning is a choice. For every trend, there are exceptions. With the right attitude, you can be one of them."

"Ten well-chosen books are usually enough to understand the expert consensus on any reasonable question you might have."

"The books have to be well-chosen. Any ten random books on a subject won’t converge to the expert wisdom."

"Understanding an expert consensus doesn’t make you an expert."

"There must actually be an expert consensus on the question (or at least, a few dominant viewpoints)."

"Books with “new” ideas. Most ideas are old, even in supposedly cutting-edge fields."

"The first difficulty people have with this approach to research is that they pick the wrong books."

"Books with “useful” ideas."

"Books with “revolutionary” ideas."

"Up-to-date textbooks."

"Academic monographs."

"Canonical texts that the field cites as authoritative."

"“if I learned as much as an expert, what opinions would I likely form?” The ten-book rule helps you get close to this estimate."

"Discourse tends to be grounded in a consensus viewpoint."

"Self-conducted research isn’t without pitfalls."

"As mentioned above, a major reason people don’t reach the expert consensus after ten books isn’t that their goal was impossible. It’s because they picked the wrong books."

"Bruce Lee is reported to have said, “I fear not the man who has practiced 10,000 kicks once, but the man who has practiced one kick 10,000 times.”"

"Variability plays an essential and oft-neglected role in mastering complex skills. Considerable research shows that practicing in varied contexts with varied methods and performing with varied task constraints results in more robust learning than simple repetition."

"Contextual interference improves the transfer of learning to new situations."

"Identifying problems correctly and ensuring the correct technique is associated with the problem."

"The extra effort needed to retrieve the right response may be desirable."

"According to psychologist Robert Bjork’s influential theory of memory, more difficult retrieval results in greater memory strengthening than easier retrieval."

"Experts tend to perceive the deep principles of a particular problem. In contrast, novices tend to get distracted by the superficial features."

"Some theorists argue that we reason through storing multiple, specific instances of ideas."

"A substantial challenge in learning is that the mind economizes on effort. This means we often fall prey to psychological shortcuts that give us the correct answer, even if they won’t benefit us in future situations."

"But having multiple methods for getting the right answer is also an important part of mastering complex skills."

"Regardless of whether thinking is fundamentally concrete or abstract (or some mixture of both), seeing multiple examples is central to learning."

"When variability was low, participants simply memorized the pattern. In contrast, when variability was high, they simulated the trajectory to find the likely destination."

"Strategy variability will be highest when we have some, but not a lot, of experience."

"Having multiple strategies for solving a problem is vital when you aren’t yet at the level of mastery."

"Most research supports the benefits of variability in practice. However, less-variable practice is often better for beginners or lower-performing students."

"Having variable methods may also backfire if some of those methods are buggy or flawed."

"These two considerations moderate the extreme stance that all variability is good when learning. Instead, we want to see a slow ramp-up in variability."

"Mix up your problems."

"Choose opportunities with greater variability."

"Work with multiple teachers, peers and styles."

"Add constraints that force you away from dominant strategies."