Chapter 21 introduces 13 phrase models that illustrate the essential harmonic, contrapuntal,and structural properties of the different eight-bar phrases commonly found in standardtunes.

The terms “turnaround” and “tag ending” are generic labels that do not indicate a partic-ular chord sequence; rather, they suggest the specific formal function of these progressions.In jazz, there is a certain subset of harmonic progressions whose names suggest specificchord successions. When jazz musicians use the term “Lady Bird” progression,for instance, it connotes a particular chromatic turnaround from Tadd Dameron’s tuneof the same title recorded in 1947. Figure 13.9 illustrates the chord structure of thatprogression using Model VI of harmonic realization

The ability to realize harmonic progressions on the keyboard is an essential skill for thecontemporary jazz musician, regardless of her/his primary instrument. The forthcomingmodels of keyboard playing will help to accomplish this objective

Training language models to follow instructionswith human feedback

It was only by building an additional AI-powered safety mechanism that OpenAI would be able to rein in that harm, producing a chatbot suitable for everyday use.

Dass das ägyptische Wort p.t (sprich: pet) "Himmel" bedeutet, lernt jeder Ägyptologiestudent im ersten Semester. Die Belegsammlung im Archiv des Wörterbuches umfaßt ca. 6.000 Belegzettel. In der Ordnung dieses Materials erfährt man nun, dass der ägyptische Himmel Tore und Wege hat, Gewässer und Ufer, Seiten, Stützen und Kapellen. Damit wird greifbar, dass der Ägypter bei dem Wort "Himmel" an etwas vollkommen anderes dachte als der moderne westliche Mensch, an einen mythischen Raum nämlich, in dem Götter und Totengeister weilen. In der lexikographischen Auswertung eines so umfassenden Materials geht es also um weit mehr als darum, die Grundbedeutung eines banalen Wortes zu ermitteln. Hier entfaltet sich ein Ausschnitt des ägyptischen Weltbildes in seinem Reichtum und in seiner Fremdheit; und naturgemäß sind es gerade die häufigen Wörter, die Schlüsselbegriffe der pharaonischen Kultur bezeichnen. Das verbreitete Mißverständnis, das Häufige sei uninteressant, stellt die Dinge also gerade auf den Kopf.

"There is a robust debate going on in the computing industry about how to create it, and whether it can even be created at all."

L.L.M.s have a disturbing propensity to just make things up out of nowhere. (The technical term for this, among deep-learning experts, is ‘‘hallucinating.’’)

‘‘I think it lets us be more thoughtful and more deliberate about safety issues,’’ Altman says. ‘‘Part of our strategy is: Gradual change in the world is better than sudden change.’’

OpenAI had a novel structure, which the organization called a ‘‘capped profit’’ model.

One of the most well-documented shortcomings of large language models is that they can hallucinate. Because these models have no direct knowledge of the physical world, they're prone to conjuring up facts out of thin air. They often completely invent details about a subject, even when provided a great deal of context.

language models are incredible "yes, and" machines, allowing writers to quickly explore seemingly unlimited variations on their ideas.

The application is powered by LaMDA, one of the latest generation of large language models. At its core, LaMDA is a simple machine — it's trained to predict the most likely next word given a textual prompt. But because the model is so large and has been trained on a massive amount of text, it's able to learn higher-level concepts.

What signals are available to participants,and how are they compiled into estimates of rank? Their modelassumes that knowledge of rank is noisy, but not (statistically) biased.While we can build more-sophisticated models of the biases in ourjudgments, however, Kawakatsu et al.’s (1) success highlights thevirtues of simplicity. It is possible, for example, that, even if the sig-nals are not accurate at first, we might act to make them so.

creative fields like music and writing live and die based on creativity, not financial statements and branding deals.

If my interpretation of the Retrieval quadrant is correct, it will become much more difficult to be an average, or even above average, writer. Only the best will flourish. Perhaps we will see a rise in neo-generalists.

Develop Credential Quality Guidelines and Processes

Standardization of these concepts would allow for validators to sift through credential wallets anddistinguish which credentials are most relevant in a specific use case. Critical to linking up such trustinformation is a more prominent role for dedicated trust providers in the credential ecosystem.These organizations include accreditation boards and regulators of professions, as well as otherssuch as ranking boards and private quality assurance agencies who publish quality standards foreducational organizations and maintain lists of which organizations match the criteria

Multiple initiatives have tried to make various kinds of social recommendations by issuingcredentials. However, up to this point they have worked better in closed social networks rather thanas open credentials due to the ability of social networks to tie a recommendation with the profile(and identity) of the recommender. There are also several nascent initiatives to create open linkeddata around which skills, credentials and issuers are valued by employers.

One of the clear signs that the bottleneck to low-income adults working moreresults from their lack of opportunities is provided by looking at their hours of workover the business cycle. When the economy is strong and jobs are plentiful, low-incomeworkers are more likely to find work, find work with higher pay, and be able to securemore hours of work than when the economy is weak. In 2000, when the economy wasclose to genuine full employment, the unemployment rate averaged 4.0 percent and thepoverty rate was 11.3 percent; but in 2010, in the aftermath of the Great Recession, theunemployment rate averaged 9.6 percent and the poverty rate was almost 15.1 percent.What changed in those years was not poor families’ attitudes toward work but simplythe availability of jobs. Among the bottom one-fifth of nonelderly households, hoursworked per household were about 40 percent higher in the tight labor market of 2000than in recession- plagued 2010.Given the opportunity for work or additional work hours, low-income Americanswork more. A full-employment agenda that increases opportunities in the labor market,alongside stronger labor standards such as a higher minimum wage, reduces poverty.

natural-language processing is going to force engineers and humanists together. They are going to need each other despite everything. Computer scientists will require basic, systematic education in general humanism: The philosophy of language, sociology, history, and ethics are not amusing questions of theoretical speculation anymore. They will be essential in determining the ethical and creative use of chatbots, to take only an obvious example.

Houston, we have a Capability Overhang problem: Because language models have a large capability surface, these cases of emergent capabilities are an indicator that we have a ‘capabilities overhang’ – today’s models are far more capable than we think, and our techniques available for exploring the models are very juvenile. We only know about these cases of emergence because people built benchmark datasets and tested models on them. What about all the capabilities we don’t know about because we haven’t thought to test for them? There are rich questions here about the science of evaluating the capabilities (and safety issues) of contemporary models. 

11/30 Youth Collaborative

Misleading Templates There is no consistent re-lation between the performance of models trainedwith templates that are moderately misleading (e.g.{premise} Can that be paraphrasedas "{hypothesis}"?) vs. templates that areextremely misleading (e.g., {premise} Isthis a sports news? {hypothesis}).T0 (both 3B and 11B) perform better givenmisleading-moderate (Figure 3), ALBERT andT5 3B perform better given misleading-extreme(Appendices E and G.4), whereas T5 11B andGPT-3 perform comparably on both sets (Figure 2;also see Table 2 for a summary of statisticalsignificances.) Despite a lack of pattern between

Insum, notwithstanding prompt-based models’impressive improvement, we find evidence ofserious limitations that question the degree towhich such improvement is derived from mod-els understanding task instructions in waysanalogous to humans’ use of task instructions.

Suppose a human is given two sentences: “Noweapons of mass destruction found in Iraq yet.”and “Weapons of mass destruction found in Iraq.”They are then asked to respond 0 or 1 and receive areward if they are correct. In this setup, they wouldlikely need a large number of trials and errors be-fore figuring out what they are really being re-warded to do. This setup is akin to the pretrain-and-fine-tune setup which has dominated NLP in recentyears, in which models are asked to classify a sen-tence representation (e.g., a CLS token) into some

Antibiotic resistance has become a growingworldwide concern as new resistance mech-anisms are emerging and spreading globally,and thus detecting and collecting the cause– Antibiotic Resistance Genes (ARGs), havebeen more critical than ever. In this work,we aim to automate the curation of ARGs byextracting ARG-related assertive statementsfrom scientific papers. To support the researchtowards this direction, we build SCIARG, anew benchmark dataset containing 2,000 man-ually annotated statements as the evaluationset and 12,516 silver-standard training state-ments that are automatically created from sci-entific papers by a set of rules. To set upthe baseline performance on SCIARG, weexploit three state-of-the-art neural architec-tures based on pre-trained language modelsand prompt tuning, and further ensemble themto attain the highest 77.0% F-score. To the bestof our knowledge, we are the first to leveragenatural language processing techniques to cu-rate all validated ARGs from scientific papers.Both the code and data are publicly availableat https://github.com/VT-NLP/SciARG.

“The metaphor is that the machine understands what I’m saying and so I’m going to interpret the machine’s responses in that context.”

understanding mental models of learning will make it easier to think about learning problems.

Business ModelWill I get charged at some point? How do you make money to run this product?TBD

In the end, universities live off of being exclusive.

many of their parents are still paying back their student loans.

Harris said this model is often better for the textbook authors OpenStax works with, whom Harris called "the long tail" behind the minority of financially successful academic authors -- those who wouldn't necessarily sell enough units to make a lot in royalties, but who are committed to their work nonetheless.

"We are fully committed to providing affordable, high-quality learning solutions for students," Joyner said. "We are excited to think openly and collaboratively with key partners like OpenStax to ensure that we, and our authors, are able to reach as many students as possible in new and highly accessible ways."

Ernest Hemingway was one of the most recognized and influentialnovelists of the twentieth century. He wrote in an economical,understated style that profoundly influenced a generation of writersand led to his winning the Nobel Prize in Literature in 1954.

It was as if Silicon Valley had made a secret pact to subsidize the lifestyles of urban Millennials. As I pointed out three years ago, if you woke up on a Casper mattress, worked out with a Peloton, Ubered to a WeWork, ordered on DoorDash for lunch, took a Lyft home, and ordered dinner through Postmates only to realize your partner had already started on a Blue Apron meal, your household had, in one day, interacted with eight unprofitable companies that collectively lost about $15 billion in one year.

Free public projects private projects starting at $9/month per project

In her 2021 book "Bet on Yourself," which features a foreword by Schmidt, Hiatt lays out the two key ways she "up-leveled" her career."First I have prioritized finding a manager who is modeling the career path I want to take and embodies the leadership qualities I want to possess," she wrote. "Second, I have chosen roles that surround me with top quality people and a depth of opportunities to grow with them."

Our brains work not that differently in terms of interconnectedness.Psychologists used to think of the brain as a limited storage spacethat slowly fills up and makes it more difficult to learn late in life. Butwe know today that the more connected information we alreadyhave, the easier it is to learn, because new information can dock tothat information. Yes, our ability to learn isolated facts is indeedlimited and probably decreases with age. But if facts are not kept

Most writing is chasing clout, rather than insight

Founded in partnership with a team of entrepreneurial journalists who believe in a better model to create excellent content while narrowing the synapse between elite creators and their audiences.

Aligning editorial mission and business model is critical.

Axios Pro is bundling newsletters together in a high-priced subscription product ($2,500 for the bundle; $599 each) aimed squarely at deep-pocketed investors.

Legal Cases: Open Education

also into business models that can better serve the interests of both students and educators.

“Speed kills.” If you are able to be nimble, assess the ever-changing environment, and adapt quickly, you’ll always carry the advantage over any opponents. Start applying the OODA Loop to your day-to-day decisions and watch what happens. You’ll start to notice things that you would have been oblivious to before. Before jumping to your first conclusion, you’ll pause to consider your biases, take in additional information, and be more thoughtful of consequences.

When you act fast enough, other people view you as unpredictable. They can’t figure out the logic behind your decisions.

Boyd made use of the Second Law of Thermodynamics. In a closed system, entropy always increases and everything moves towards chaos. Energy spreads out and becomes disorganized. Although Boyd’s notes do not specify the exact applications, his inference appears to be that a fighter pilot must be an open system or they will fail. They must draw “energy” (information) from outside themselves or the situation will become chaotic. They should also aim to cut their opponent off, forcing them to become a closed system.

The second concept Boyd referred to is Heisenberg’s Uncertainty Principle. In its simplest form, this principle describes the limit of the precision with which pairs of physical properties can be understood. We cannot know the position and the velocity of a body at the same time. We can know either its location or its speed, but not both.

Boyd referred to three key principles to support his ideas: Gödel’s theorems, Heisenberg’s Uncertainty Principle, and the Second Law of Thermodynamics. Of course, we’re using these principles in a different way from their initial purpose and in a simplified, non-literal form.

Gödel’s theorems indicate any mental model we have of reality will omit certain information and that Bayesian updating must be used to bring it in line with reality. For fighter pilots, their understanding of what is going on during a battle will always have gaps. Identifying this fundamental uncertainty gives it less power over us.

If the opponent uses an unexpected strategy, is equipped with a new type of weapon or airplane, or behaves in an irrational way, the pilot must accept the accompanying uncertainty. However, Boyd belabored the point that uncertainty is irrelevant if we have the right filters in place.

If we can’t cope with uncertainty, we end up stuck in the observation stage. This sometimes happens when we know we need to make a decision, but we’re scared of getting it wrong. So we keep on reading books and articles, asking people for advice, listening to podcasts, and so on.

Speed is a crucial element of military decision-making. Using the OODA Loop in everyday life, we probably have a little more time than a fighter pilot would. But Boyd emphasized the value of being decisive, taking initiative, and staying autonomous. These are universal assets and apply to many situations.

There’s a difference between making decisions and enacting decisions. Once you make up your mind, it’s time to take action. By taking action, you test your decision out.

This part of the loop needs to be flexible and open to Bayesian updating. In some of his notes, Boyd described this step as the hypothesis stage. The implication is that we should test the decisions we make at this point in the loop, spotting their flaws and including any issues in future observation stages

He recommended a process of “deductive destruction”: paying attention to your own assumptions and biases, then finding fundamental mental models to replace them.

He identified the following four main barriers that impede our view of objective information: Our cultural traditions – we don’t realize how much of what we consider universal behavior is actually culturally prescribed Our genetic heritage – we all have certain constraints Our ability to analyze and synthesize – if we haven’t practiced and developed our thinking skills, we tend to fall back on old habits The influx of new information – it is hard to make sense of observations when the situation keeps changing

To orient yourself is to recognize any barriers that might interfere with the other parts of the OODA Loop. Orientation means connecting yourself with reality and seeing the world as it really is, as free as possible from the influence of cognitive biases and shortcuts.

If you want to make good decisions, you need to master the art of observing your environment.

Competent scientists do not believe their own models or theories, but rather treat them as convenient fictions. ...The issue to a scientist is not whether a model is true, but rather whether there is another whose predictive power is enough better to justify movement from today's fiction to a new one. Steve Vardeman, 1987. Comment. Journal of the American Statistical Association 82 : 130-131. [kw]

A series of studies conducted by Frédéric Vallée-Tourangeau, a professor of psychology at Kingston University in Britain; Gaëlle Vallée-Tourangeau, a professor of behavioral science at Kingston; and their colleagues, has explored the benefits of such interactivity. In these studies, experimenters pose a problem; one group of problem solvers is permitted to interact physically with the properties of the problem, while a second group must only think through the problem. Interactivity “inevitably benefits performance,” they report.

Moving mental contents out of our heads and onto the space of a sketch pad or whiteboard allows us to inspect it with our senses, a cognitive bonus that the psychologist Daniel Reisberg calls “the detachment gain.”

“The real economics of college have shifted so much during the last 70 years, and we have not made adjustments to all those changes. Students are in an equation that has not adapted to the circumstances.”

How a memory palace works When we’re learning something new, it requires less effort if we connect it to something we already know, such as a physical place. This is known as elaborative encoding. Once we need to remember the information, we can “walk” around the palace and “see” the various pieces. The idea is to give your memories something to hang on to. We are pretty terrible at remembering things, especially when these memories float freely in our heads. But our spatial memory is actually pretty decent, and when we give our memories some needed structure, we provide that missing order and context. For example, if you struggle to remember names, it can be helpful to link people you meet to names you already know. If you meet someone called Fred and your grandmother had a cat called Fred, you could connect the two. Creating a multisensory experience in your head is the other part of the trick. In this case, you could imagine the sound of Fred meowing loudly. To further aid in recall, the method of loci is most effective if we take advantage of the fact that it’s easiest to remember memorable things. Memory specialists typically recommend mentally placing information within a physical space in ways that are weird and unusual. The stranger the image, the better.

The problem is, algorithms were never designed to handle such tough choices. They are built to pursue a single mathematical goal, such as maximizing the number of soldiers’ lives saved or minimizing the number of civilian deaths. When you start dealing with multiple, often competing, objectives or try to account for intangibles like “freedom” and “well-being,” a satisfactory mathematical solution doesn’t always exist.

many other systems that are already here or not far off will have to make all sorts of real ethical trade-offs

In a way, the essential premise of the collab-house business model is not far from that of pornographic entertainment. (Where else do talent and crew and cadres of management congregate in furnished mansions to produce intimate content?) Interestingly, but maybe not surprisingly, many TikTok influencers, including some here at the Clubhouse, have made the crossover from social media to pornography, using apps such as OnlyFans to post nude pics for their legions of subscribers.

Later in my visit, Chase Zwernemann, the twenty-one-year-old VP of talent management, will tell me that “we really see ourselves like influencing professors.” And if this weren’t enough, under the Clubhouse aegis is a trio of TikTok houses, each of which corresponds, apparently, to a different level of academe. There’s Clubhouse BH—the grad school—which is meant for “our more seasoned influencers.” (If this phrase conjures for you images of geriatrics taking selfies in suggestive postures, please know that by “seasoned influencers” they simply mean people who have been in the business a while and have thus reached the ripe old age of twenty-two or twenty-three.) Beneath that is Clubhouse FTB, which apparently serves as the undergraduate program. And finally, there’s Not a Content House, the high school of the Clubhouse venture, one meant to appeal to an even younger demographic.

Microsoft 365 Business Standard and Microsoft 365 Business Premium plans will include all the features above for up to 300 attendees

In viewing academia as a business, you should always give customers what they want, and this applies on two levels. First, always consider the demand for the research product. This is much easier said than done. Anyone can acknowledge that the customers are always right, but truly listening to them and extracting what they need is difficult, especially if you have your own personal desires with respect to the product (in this case, the research). Talk to the funding customer constantly. Second, most students are, in effect, employees, and the adviser is a boss who doubles as a customer. In some respects, your adviser will provide your pay cheque, or at least govern it. Thus, do what the customer requires. In addition, always consider your audience when writing and presenting. In the case of a thesis, the audience is your adviser and committee. Again, talk to the customers constantly.

Anyone who treats research as a business tends not to be well received in academia, but they likely have the funding necessary to drive advances, and they may eventually be wealthy.

Amidst the global pandemic, this might sound not dissimilar to public health. When I decide whether to wear a mask in public, that’s partially about how much the mask will protect me from airborne droplets. But it’s also—perhaps more significantly—about protecting everyone else from me. People who refuse to wear a mask because they’re willing to risk getting Covid are often only thinking about their bodies as a thing to defend, whose sanctity depends on the strength of their individual immune system. They’re not thinking about their bodies as a thing that can also attack, that can be the conduit that kills someone else. People who are careless about their own data because they think they’ve done nothing wrong are only thinking of the harms that they might experience, not the harms that they can cause.

Substack insists that advances are determined by “business decisions, not editorial ones”. Yet it offers writers mentoring and legal advice, and will soon provide editing services.

Record labels are another endangered middleman. They have historically taken care of turning a song into a hit, in return for an ongoing share of revenues. But more and more artists are going it alone. More than 60,000 new songs are uploaded to Spotify every day, most by bedroom-based rockstars who can use new online services to handle the logistics themselves. UnitedMasters, a music-distribution platform which bills itself as “a record label in your pocket”, recently raised $50m in a venture-capital round led by Apple. Tools like Splice make recording easier. Companies like Fanjoy take care of merchandise.And financing is getting simpler. One startup, HIFI, helps artists manage their royalties, paying them regularly and fronting them small sums to make up shortfalls. Another, Karat, extends credit to creators based on their follower count. Helped by such services independent artists took home 5.1% of global recorded music revenues last year, up from 1.7% in 2015, calculates MIDiA Research, a consultancy. In the same period the share of the three largest record labels fell from 71.1% to 65.5%.

Paywall success stories are a mix of specialized news, often catering to the ultrawealthy (WSJ), superstar news orgs focused on specific national and international news (NYT) or news with billionaire backstops (WP). The civic function of news is not met by any of these models. But reader-supported, open access news, like Canadaland and The Halifax Examiner are filling in the gaps. The co-op model is a most welcome adjunct to these success stories.

In Trailblazer, models are completely empty. They solely contain associations and finders. No business logic is allowed in models.

A popular strategy for bootstrapping networks is what I like to call “come for the tool, stay for the network.” The idea is to initially attract users with a single-player tool and then, over time, get them to participate in a network. The tool helps get to initial critical mass. The network creates the long term value for users, and defensibility for the company.

the idea of decentralized business models. These are business models distributed across a blockchain network not centralized in a traditional corporation. They can be fully autonomous too, meaning no humans involved. There's big potential for true peer-to-peer models like ride-sharing without Uber as a middleman taking fees. Owners of driverless cars will someday be able to put their autonomous vehicles to work. So even the big disruptors can be disrupted by this new technology.

But the inverse trajectory, from which this essay takes its name, is now equally viable: “come for the network, pay for the tool.” Just as built-in social networks are a moat for information products, customized tooling is a moat for social networks.1 This entrenchment effect provides a realistic business case for bespoke social networks. Running a bespoke social network means you’re basically in the same business as Slack, but for a focused community and with tailored features. This is a great business to be in for the same reasons Slack is: low customer acquisition costs and long lifetime value. The more tools, content, and social space are tied together, the more they take on the qualities of being infrastructure for one’s life.

Most business writing lacks a meaningful engagement with the question of whether the strategies, tactics, and trends on offer are good, in a larger and longer term sense. It is negligent not to address these questions.