PDF Summary:Nine Algorithms That Changed the Future, by John MacCormick

Algorithms are the driving force behind computing—from the search engines that organize the vast World Wide Web, to the cryptographic protocols that secure online transactions. In Nine Algorithms That Changed the Future, John MacCormick unravels the ingenious techniques that have revolutionized how we process, safeguard, and derive insights from data.

You'll explore groundbreaking algorithms, including indexing for rapid information retrieval, public key cryptography enabling secure communication, and pattern recognition through neural networks. MacCormick also unveils the theoretical underpinnings of computation, revealing undecidable problems that even the most advanced algorithms cannot solve.

(continued)...

Practical Tips

• Protect your personal data transfers by incorporating checksums into your backup routine. When backing up your data to an external hard drive or cloud service, use software that includes a checksum verification feature, such as TeraCopy for Windows or rsync for Linux, to ensure that your data is accurately copied without any errors.
• You can enhance your data integrity by using CRC-32 when creating digital archives. When archiving your personal files, such as family photos or important documents, use software that offers CRC-32 as an option for checksums. This will ensure that your archives are less likely to suffer from unnoticed data corruption over time.
• Use a checksum method to verify the accuracy of your household budget. When you write down your expenses, sum the digits of each amount spent and keep the final digit. At the end of the month, compare the checksum of your recorded expenses with the actual checksum of your bank statement to quickly check for discrepancies.
• Create a simple habit of verifying the integrity of emails containing attachments by asking senders to include a checksum value in their message. Before opening attachments, use a free checksum calculator to generate a checksum for the received file and compare it to the one provided by the sender. This practice can help you avoid malware or corrupted files.
• Create a checksum-inspired verification system for your home library. Assign a number to each bookshelf and a letter to each book. For instance, the first book on the first shelf could be 1A, the second 1B, and so on. Keep a master list with a simple checksum that adds up all the numbers. Whenever you lend out a book, update the list and checksum. When all books are returned, check the checksum to ensure all books are back and in their proper place.
• Develop a habit of manually cross-verifying important numerical data using different methods. If you're working with financial spreadsheets, for example, don't rely solely on the software's sum function. Instead, periodically perform spot checks by adding up figures using a calculator or even mentally. This practice can help you catch errors that might slip through automated checks, which could be especially useful if you're dealing with critical data where accuracy is paramount.
• Improve your online transactions by selecting platforms that implement two-factor authentication and end-to-end encryption. This adds an extra layer of security beyond what a simple checksum can provide, safeguarding your financial and personal data from potential breaches.
• Enhance your understanding of error detection by creating a visual representation of how different checksum algorithms work. Use a spreadsheet to simulate a staircase checksum or other algorithms by inputting data and visually mapping out how the algorithm checks for errors. This will help you grasp the concept without needing advanced technical skills.

Checksums Ensure Data Integrity

MacCormick emphasizes the practical significance of checksums for contemporary computing, highlighting their use in network protocols like Ethernet and TCP/IP. He also points out that Checksums play a significant role in verifying the integrity of online software releases, with MD5 and SHA-1 previously being widely employed. The author emphasizes the necessity of "hash functions" in these scenarios, which offer greater protection against deliberate tampering with the software, as opposed to random communication errors. He provides examples of popular algorithms that generate checksums, like MD5, SHA-1, SHA-256, and SHA-2, each offering different levels of resistance to mistakes and deliberate manipulation. Finally, MacCormick describes a particular version of the redundancy method called the "pinpoint trick," exemplified by two-dimensional parity used in some computer systems. He explains how this technique employs a two-dimensional array and checksum calculations for horizontal and vertical lines to precisely locate and correct any single-digit mistake in a message, illustrating the concept visually.

Practical Tips

• Apply the concept of redundancy from network protocols to your daily tasks to prevent errors. For instance, when you make a to-do list, create a secondary list or a digital copy that acts as a 'checksum'. If you ever find discrepancies between the two, you'll know something was missed or added incorrectly, prompting you to review and correct the information. This strategy uses the underlying principle of checksums to maintain accuracy and reliability in your personal organization system.
• Create a habit of checking the authenticity of software updates by setting reminders. Whenever you receive a notification for a software update, before proceeding, set a calendar reminder to first verify the checksum. This ensures you maintain a routine check and don't overlook the security step in the rush to update.
• You can verify the integrity of software you download by using hash values provided by the software developers. When you download software, look for a hash value (often an MD5 or SHA-256) on the developer's website. Use a simple online hash calculator to compute the hash of the downloaded file and compare it to the one provided. If they match, the software hasn't been tampered with.
• Educate yourself on the basics of cryptographic hash functions through free online courses or tutorials. This knowledge will help you make informed decisions about data security in your everyday technology use. For example, when you're setting up a home network or choosing a service for storing sensitive information, you'll be able to ask the right questions regarding their security protocols and ensure they're not relying on outdated algorithms like MD5 or SHA-1.
• Enhance your fitness routine by pinpointing fundamental exercises that work multiple muscle groups and incorporating them into every workout for redundancy. For example, include squats, push-ups, and planks in your routine, as they are compound movements that target various muscles and can be modified for increased difficulty over time.
• You can create a simple game that uses the concept of two-dimensional parity to teach error detection and correction. Design a board game where players must navigate a grid, and errors are intentionally introduced into their paths. Players must detect and correct these errors to advance, mirroring the way two-dimensional parity works in data transmission.
• Develop a habit of double-checking important handwritten information using a self-made parity system. Whenever you jot down critical numbers like phone numbers or account numbers, add your own parity digit at the end. When you need to use the information, check the parity digit to ensure no single-digit errors have occurred. This practice can help prevent mistakes in everyday life where a single digit can make a big difference.

Algorithms for Learning Machines and Recognizing Patterns

This section delves into algorithms for recognizing patterns, a branch of AI focused on empowering computers to identify, classify, and learn from variable data. Unlike the deterministic tasks previously discussed, identifying patterns demands flexible algorithms that adapt to the inherent uncertainties and variations present in real-world data. The author examines three core techniques: classifiers that use nearest neighbors, trees that aid in decision-making, and network-based neural models.

Nearest-Neighbor Classifiers Accurately Classify Variable Data Like Digits

MacCormick introduces nearest-neighbor algorithms, a straightforward but powerful technique that lets computers classify data based on similarities to known, labeled examples. The author describes the fundamental concept of nearest-neighbor classification, noting that it involves assigning the class of the most similar training example to an unclassified data point.

Algorithm Classifies Inputs Based on Resemblance to Examples

MacCormick illustrates the nearest-neighbor technique with a straightforward scenario of predicting the political party donation preference of a person based on their residence. He shows how a map with labeled data points indicating the donation history of nearby houses could help provide an educated guess about the political affiliation of someone at an unknown address. He further elaborates on the concept of "K-nearest-neighbors," where instead of only considering the single closest neighbor, the algorithm examines the K nearest neighbors and selects the most common class among them. He explains that the choice of an appropriate value for K depends on the specifics of the problem and often requires combining theoretical understanding with empirical testing.

Context

• Beyond predicting political preferences, nearest-neighbor algorithms are used in various fields such as image recognition, recommendation systems, and anomaly detection.
• The algorithm relies on a distance metric to determine how close data points are to each other. Common metrics include Euclidean distance, Manhattan distance, or more complex measures depending on the data's nature.
• Feature scaling or normalization is often necessary to ensure that each feature contributes equally to the distance computation, especially when features have different units or scales.
• Cross-validation can be used to empirically determine the best K by testing different values and evaluating their performance on a validation set.
• The presence of noise in the data can affect the choice of K. A larger K can help average out noise, while a smaller K might be more sensitive to it.

Nearest-Neighbor Classifiers Achieve High Accuracy With Minimal Training

MacCormick delves into the challenge of specifying "nearest" neighbors for data types that lack clear-cut measures of distance, such as images of handwritten digits. He demonstrates how a computer can calculate how far apart two digit images are by quantifying the percentage of pixels that differ between them, showcasing an example with images of the digit '2'. MacCormick emphasizes that this approach, using a thoughtfully crafted distance measure, can achieve surprisingly high accuracy, reaching over 99.5% for handwritten digit recognition. He emphasizes that this remarkable result rivals the achievements of more intricate pattern recognition systems, attesting to the simplicity and effectiveness of nearest-neighbor methods. Ultimately, the author acknowledges that this classification method stands out from others by not having a clearly defined learning period. The information is directly used during classification without undergoing any preprocessing or extraction of class-specific features.

Other Perspectives

• The statement may not fully acknowledge the existence of alternative non-Euclidean distance measures, such as cosine similarity or Jaccard index, which have been successfully applied to complex data types and could mitigate the challenges mentioned.
• The method may not generalize well to more complex image recognition tasks where the subject of interest is not centered or varies in position within the image frame.
• The high accuracy reported could be specific to the dataset used for testing, and performance might vary significantly with different datasets, especially those with more variability or noise.
• In practice, nearest-neighbor methods can be computationally expensive for large datasets because they require comparing a new instance to every instance in the training set.
• Nearest-neighbor classifiers do require a form of learning period, as the model must store the training data for future comparison, which is an implicit learning process.
• In many real-world applications, raw data can contain outliers or errors that preprocessing steps are designed to mitigate; skipping these steps could result in a classifier that is less robust to such anomalies.

Tree Structures Adapt Tests to Sort Inputs

MacCormick shifts the focus to decision trees, an intuitive and effective way to recognize patterns that closely resembles the popular guessing game. Just as players ask targeted questions with binary answers to identify an object, decision trees utilize a series of simple tests to categorize inputs.

Decision Trees Extract Features From Complex Data, Indicating Spam

MacCormick presents a straightforward example using a decision tree to decide if you should bring an umbrella. The tree involves simple questions that require a binary answer about current weather conditions like rain, humidity, and cloud cover to guide the user to the final decision. He then connects this concept to the job of identifying "web spam" -- irrelevant web pages designed to manipulate search engine rankings. The author explains how a classification model can be trained to categorize web pages based on their content characteristics, highlighting a specific example where researchers at Microsoft Research employed this model to identify spam with up to 90% accuracy.

Other Perspectives

• The interpretability of decision trees can decrease as the complexity of the data increases, making it harder to understand which features are truly important for the decision-making process.
• Decision trees for weather prediction may require more than just binary questions to capture the nuances of meteorological data.
• Decision trees can guide users to a final decision, but they may not account for uncertainty or probabilistic outcomes, which can be better handled by other models like Bayesian networks or random forests.
• Decision trees can become overly complex and prone to overfitting when trying to handle the vast and diverse features of web pages, which can reduce their effectiveness in a real-world setting.
• The term "irrelevant" is subjective and can vary depending on the user's perspective or the context in which a web page is found; what is irrelevant to one user or search query may be relevant to another.
• Spammers can adapt their strategies to evade detection by these models, necessitating continuous model updates and retraining.
• The use of such a model could raise privacy concerns if it involves analyzing the content of web pages in a way that infringes on the rights of website owners or users.

Efficient Classification With Minimal Tests in Tree Structure

MacCormick emphasizes that the tree used for decisions in the spam detection example was automatically generated by software after analyzing a dataset of labeled web pages. This automatic learning, he explains, involves building the tree by selecting questions that provide the most useful information at each step, progressively refining the classification until "pure" sets of examples representing only junk mail or non-junk mail are reached. MacCormick contrasts decision trees with the nearest-neighbor method, noting that decision trees invest upfront effort in the learning stage, resulting in an efficient classification phase that requires only a few simple tests to arrive at a prediction. Unlike the nearest-neighbor technique, which involves no learning but requires comparison with every training example during classification.

Practical Tips

• Develop a habit of regularly updating your email preferences on websites you subscribe to. This proactive approach involves going through your subscriptions and notifications settings to opt-out of irrelevant or excessive communications. By doing this, you reduce the chances of receiving unwanted emails that could be potential spam, keeping your inbox cleaner and more manageable.
• Create a visual sorting chart to categorize your emails manually, using color-coded labels for "junk" and "non-junk." Start by printing out a simple tree diagram with two branches labeled accordingly. As you go through your inbox, use colored stickers or markers to label each email on the chart. This hands-on approach will help you recognize patterns in what you consider junk or non-junk, improving your manual sorting skills over time.
• Use a free online flowchart tool to map out a complex decision you're facing, such as choosing a new career path or planning a large purchase. Input all the variables and possible outcomes to visualize the decision tree. Share the interactive tree with friends or family for feedback, which can provide new perspectives and help you make a more informed choice.
• Apply the decision tree concept to your budgeting by categorizing expenses with a set of binary choices. Start with a broad category like "necessary" or "optional" and then break it down further into more specific choices, such as "fixed" or "variable" under the necessary expenses. This method can help you quickly identify areas where you can cut costs or reallocate funds.
• Engage in a social observation exercise to apply nearest-neighbor principles to understanding social dynamics. When you're in a social setting, like a party or a meeting, observe how people group together based on apparent similarities. Try to predict new arrivals' groupings by comparing their characteristics (such as attire, behavior, or whom they know) to those already present, reflecting on how this comparison-based method can be both time-consuming and insightful.

Neural Networks Model Brain Structures to Learn Patterns

MacCormick introduces artificial neural systems, drawing inspiration from the remarkable abilities of the human brain to recognize patterns. He describes how the inherent complexity of simulating human cognition has challenged computer scientists for decades. He briefly touches on Alan Turing's early attempts to model the brain using computers, along with the subsequent emergence of neural nets.

Neural Networks Recognize Patterns Like Faces or Written Characters Without Explicit Programming

MacCormick then shifts to discussing the biological basis of neural networks, explaining how real brains are composed of interconnected neurons that transmit signals, and how individual neurons aggregate input signals from other neurons to determine whether to "fire' and transmit their own signals. The author transitions from biological neural systems to digital ones, introducing a basic artificial neuron model. This model uses a threshold value: if the input signals' sum reaches or exceeds it, the simulated neuron activates; otherwise, it stays inactive. He illustrates this with a basic neural system for the familiar issue of deciding whether to use an umbrella, showcasing how three input neurons representing weather conditions stimulate a two-neuron network to produce a decision on using an umbrella. As an example of a real-world application, MacCormick introduces the "sunglasses problem," where a neural network model must be trained to determine if sunglasses are on a face. He presents a schematic representation of a network designed for this task, highlighting the complexity involved in connecting every pixel of the input image to each central neuron, resulting in thousands of connections to be "tuned" or adjusted.

Practical Tips

• Use smartphone apps that translate written characters in real-time. These apps often use neural networks to recognize and translate text from one language to another. By trying out these apps on different types of text, you can observe the capabilities and limitations of current pattern recognition technologies.
• Use a daily journal to track decision-making patterns, akin to observing neural pathways. Record decisions you make each day and the outcomes they lead to. Over time, you'll start to see patterns emerge, similar to how neural pathways strengthen with repeated use. This can help you understand your own 'mental wiring' and make more informed choices.
• Create a 'neural' feedback loop for habit formation by setting up a system of rewards and consequences. For each new habit you're trying to establish, decide on a positive reinforcement for when you follow through and a consequence for when you don't. This mimics the way neurons strengthen or weaken connections based on feedback, and it can help solidify your habits more effectively.
• Apply the activation threshold principle to improve decision-making by creating a decision matrix. List the factors that are important to you for a particular decision and assign a threshold value to each. When evaluating options, score them against your matrix. Only consider options that meet or exceed your threshold values, ensuring you make choices aligned with your priorities.
• Develop a personal decision-making applet using IFTTT (If This Then That) or similar services that triggers reminders based on weather conditions. Set up an applet that sends you a notification to bring an umbrella when the chance of rain exceeds a certain percentage. This mimics the neural decision-making process by using technology to automate the decision for you.
• Create a collage to visualize the concept of pixel connectivity. Cut out several images of faces with sunglasses from magazines or printouts and arrange them on a poster board. Then, using colored strings or markers, draw lines connecting points across the different faces that correspond to the sunglasses' position. This visual representation can help you grasp the complexity of connections a neural network must establish to recognize sunglasses in an image.
• Explore the basics of neural networks by using simple online simulators. Online platforms like TensorFlow Playground allow you to tweak neural network parameters and see the effects in real-time without any coding knowledge. Adjusting the settings can give you a hands-on understanding of how changes in a network can affect its ability to recognize patterns or images.

Training AI Networks: Adjusting Connections to Minimize Errors

MacCormick explains how a network's parameters function as "dials," allowing the network's behavior to be tuned. Instead of manual tuning, which he notes is practically impossible for real-world networks with numerous connections, the author introduces the concept of "learning," a process in which these parameters are automatically adjusted according to training data, a set of labeled examples. MacCormick then outlines how this training happens: the network processes each training sample, compares its computed output to the "target" output it should achieve for that sample, and makes small adjustments to the connections to improve its future predictions. Applying this concept to the issue of identifying sunglasses, the author describes how after training, the weights of the network connections encode information that aids in recognizing sunglasses. Visualizing the connection strengths as an image reveals patterns resembling sunglasses, demonstrating how intelligent-like behavior can arise from a network initially trained using only labeled examples of faces with and without sunglasses. Finally, MacCormick presents the outcome of an experiment where a network trained on a collection of faces achieved 85% classification accuracy on identifying sunglasses in unseen images. He encourages the reader to analyze the images where the network fails, highlighting those that are truly hard to classify, even for people, and acknowledging the unexplained failures that are common in recognizing patterns.

Practical Tips

• Experiment with different routines to optimize your daily productivity by treating each part of your routine as a "dial." For example, adjust your wake-up time, exercise duration, or work-break intervals slightly every week and monitor how these changes affect your overall productivity and well-being. Keep a journal to track the adjustments and their impacts, helping you find the optimal settings for your personal productivity.
• Opt for smart home devices with self-optimizing features to reduce the need for manual network management. Devices like smart thermostats or lighting systems often come with algorithms that learn your habits and adjust settings accordingly, minimizing the need for you to manually change settings as your environment or preferences evolve.
• Start a learning journal to reflect on your experiences and tweak your approach accordingly. After each learning session, jot down what worked well, what didn't, and any ideas you have for improving the process. This could be as simple as realizing that studying in shorter, more frequent intervals keeps you more engaged than long sessions, leading you to break up your study time differently.
• Engage in citizen science projects that require data labeling, such as classifying galaxies or identifying wildlife in photos. These projects often provide a tutorial on how to label the data, giving you practical experience with the process and its significance in contributing to scientific research. This activity will help you grasp the collective effort needed to build reliable datasets for scientific inquiry.
• Implement a budgeting app to manage your finances, comparing your spending against your budget, and tweaking your habits to save money. Input your income and expenses into the app, and it will help you identify areas where you're overspending. If you notice you're consistently overspending on dining out, you could adjust by setting a lower budget for restaurants and cooking more meals at home.
• Use spaced repetition software for learning new concepts or vocabulary. These programs are designed to present information at optimal intervals, similar to how neural networks are trained over multiple iterations. By reviewing material just as you're about to forget it, you can strengthen your memory connections in a way that's analogous to how a network's connection weights are optimized for recognition.
• Use color-coded sticky notes to represent different types of connections in your daily life, such as family, friends, work colleagues, and hobbies. Place these notes on a wall in a pattern that resembles sunglasses, with each lens dedicated to a different category. Over time, you can rearrange the notes to reflect changes in the strength or importance of these connections, giving you a dynamic, visual tool to monitor and adjust your social and professional priorities.
• Experiment with photo tagging on social media to understand how labeled examples work. Upload a batch of photos and tag the people or objects in them. Over time, notice how the platform's algorithms start to recognize and suggest tags for these entities. This hands-on activity will show you the practical application of networks learning from labeled examples, as the system uses your tags as data points to learn and make future predictions.
• Enhance your photography skills by taking pictures of sunglasses in various settings and lighting conditions. Analyze your photos to see how different angles, lighting, and backgrounds affect the visibility and recognition of the sunglasses. This hands-on approach will give you a better understanding of visual cues and how they can be captured effectively.
• You can start a personal log to track and categorize any tech issues you encounter daily. By keeping a detailed record, you'll begin to see patterns and classifications of network problems in your own life, which can help you troubleshoot more effectively. For example, if your home Wi-Fi frequently drops, noting the times and circumstances can lead to a better understanding of whether it's a device issue, a signal interference, or a service provider problem.
• Start a 'failure analysis' group with friends or colleagues where you collectively discuss instances of unexplained failures in pattern recognition. Each member brings a personal or professional example of an unexpected failure, and the group analyzes the situation to identify potential blind spots or biases that may have affected the pattern recognition process. This collaborative approach can help uncover common pitfalls and improve individual and collective understanding of pattern recognition failures.

Computation Theory and the Limitations of Algorithms

This part of the book takes a turn towards the limitations of algorithms, challenging the notion that computing devices can solve any problem. MacCormick explains the concept of "undecidability," focusing on the issue of halting first presented by Alan Turing. The chapter then delves into a specific example proving that no program can detect every potential malfunction in other programs, highlighting the limitations of computation that extend beyond practical constraints.

No Program Can Be Created That Detects All Crashes

MacCormick notes that despite computers' immense processing power, there remain tasks they cannot perform well. He then questions if some problems might be impossibly hard, problems for which no algorithm could ever be devised. He states upfront that it's a definitive 'yes' -- some problems are provably unsolvable by computers. The author then introduces "proof through contradiction," a technique commonly employed in both formal and informal reasoning. He demonstrates the technique with a simple example proving that Abraham Lincoln wasn't born in 1520, based on known facts about Lincoln's presidency and the dates of the U.S. Civil War. MacCormick then focuses on how dependable software is, noting that despite improvements in recent decades, crashes still occur. He examines the role of automated software-checking tools in improving reliability, then asks if these tools could ever find every potential issue in every program. The author then sets out to prove that flawless crash-detection programs can't exist.

The Undecidability of Halting Problems

MacCormick describes the problem of halting, a foundational concept in computer science first presented by Alan Turing. The author explains that the Halting Problem asks whether it's possible to write a program that can evaluate all other programs and determine if they will eventually halt (terminate) or run forever. Turing's groundbreaking work demonstrated that the challenge of ascertaining if a program will halt is "undecidable," meaning it is impossible to create a general algorithm that will solve it. MacCormick connects the Halting Problem to crash-detection, noting that it too is an undecidable problem, highlighting that there are multiple classes of problems that defy solutions through computer programs.

Context

• While the Halting Problem is undecidable in general, specific instances can sometimes be resolved. Programmers often use heuristics and testing to address halting-like issues in practical scenarios.
• Turing's proof involves self-referential logic, similar to the paradoxes found in set theory, such as Russell's Paradox, which deals with sets that contain themselves.
• Alan Turing introduced the Halting Problem in 1936, a time when the foundations of computer science were being established, significantly influencing the development of the field.
• While a general solution is impossible, practical methods such as static analysis, testing, and runtime monitoring are used to identify and mitigate potential crashes in specific cases, though they cannot guarantee complete coverage.

Perfect Crash-Detection Program Impossible via Contradictory Proof

MacCormick builds up to the proof by introducing the concept of "yes-no" programs, which simply output either affirmative or negative answers. He describes a series of increasingly complex binary output programs, illustrating their behavior with numerous examples. First, he introduces ProgramA.exe and ProgramB.exe, simple programs that perpetually generate the outputs "affirmative" and "negative," respectively. MacCormick then moves on to SizeChecker.exe, a program that evaluates the size of an input file and says "yes" if it exceeds 10 kilobytes, and "no" otherwise. Finally, he defines NameSize.exe, which checks an input file's name and outputs a positive response if it contains a minimum of one character and "no" otherwise.

He then transitions to a more intricate yes-no program, AlwaysYes.exe, which outputs "yes" only if its input is another yes-no program that perpetually outputs "yes." The author defines YesOnSelf.exe, which returns "yes" only if a program provided as input gives "yes" when it receives itself as an input. After analyzing various inputs to YesOnSelf.exe, MacCormick points out a paradoxical situation: when given its own code as input, YesOnSelf.exe can seemingly output either an affirmative or a negative and still fulfill its definition. Introducing AntiYesOnSelf.exe, which reverses YesOnSelf.exe's results, the author demonstrates how this program leads to a contradiction when used on itself. Since AntiYesOnSelf.exe cannot exist, and it can be easily created from YesOnSelf.exe, MacCormick concludes using a logical progression that neither program is possible. He then builds on this insight to show that AlwaysYes.exe is impossible, as it could be easily transformed into YesOnSelf.exe.

Finally, MacCormick leverages these concepts to prove the impossibility of a program that detects crashes. He starts by assuming there is a program called CanCrash.exe and guides the reader through a series of conceptually simple transformations, mimicking the earlier proof for AntiYesOnSelf.exe. The author demonstrates that by strategically modifying CanCrash.exe, one can create a new program, AntiCrashOnSelf.exe, that will inevitably contradict itself. This contradiction, he asserts, proves the initial assumption of CanCrash.exe's existence is false.

Practical Tips

• Use a "yes-no" checklist to streamline your morning routine. Create a list of actions that are essential for starting your day right, such as "Did I make my bed?" or "Have I packed a healthy lunch?" By going through this checklist each morning, you ensure that you don't skip important steps and establish a productive tone for the day.
• Create a daily affirmation alarm on your phone to remind you to acknowledge something positive about your day. Set multiple alarms with personalized affirmations that resonate with your goals and aspirations. For example, if you're working on becoming more patient, your alarm could say, "I am calm and patient in all situations."
• Create a positivity playlist with songs that have uplifting lyrics and melodies. Play this playlist when you're feeling down or when you need a boost. The idea is to create an auditory environment that counteracts the perpetual "negative" output, similar to how you might use music to improve your mood or motivation.
• You can enhance your digital organization by creating a simple naming convention for your files that ensures each name is descriptive and unique. Start by identifying the main categories of files you handle regularly, such as "Photos," "Receipts," or "Projects." Then, establish a format that includes the category, date, and a brief description, like "Photos20220412Cats" or "Receipts20220412Groceries." This will help you quickly locate files and ensure that each file name is meaningful and contains at least one character.
• Engage with a friend or family member in a mutual accountability partnership where you both commit to sending each other a daily message affirming a positive action or thought. This mirrors the concept of a program that outputs "yes" perpetually, reinforcing positive behavior or thoughts consistently. If your goal is to be more active, you might send each other a message every day you complete your workout, like "Crushed my 30-minute walk today!"
• Enhance your self-awareness by writing a daily reflection that ends with the question, "How did today's experiences reflect my understanding of myself?" This practice helps you to see patterns in your behavior and thoughts by consistently relating daily events back to your self-perception. For example, if you notice you're often frustrated in traffic, you might reflect on whether this aligns with your view of yourself as a patient person and what that might mean for other areas of your life.
• Engage in a "devil's advocate" conversation with a friend where you defend a belief or decision you've made, and they take on the role of challenging it. For example, if you believe you should invest in a certain stock, your friend would argue against it. This exercise can help you see the other side of your decisions and beliefs, strengthening your critical thinking and decision-making skills.
• Try to identify real-life systems or rules in your daily routine that might lead to a contradiction if turned on themselves. For instance, if you have a personal rule to question all advice given to you, question what happens when you apply this rule to itself. Does it hold up, or does it create a paradox?
• Develop a habit of asking for feedback from friends or colleagues to create a realistic self-view. After completing a task or project, reach out to at least two people for their honest opinions on your performance. This helps you understand your strengths and areas for improvement without leaning too heavily on self-criticism or self-praise.
• Experiment with a "Yes Audit" by tracking your affirmative responses for a month and categorizing them based on their alignment with your personal goals. Create a simple spreadsheet with columns for the request, your immediate response, the reason for your response, and the outcome. At the end of the month, review the data to identify trends and determine if your yeses are leading to personal growth and satisfaction. This exercise can reveal whether you're leaning towards an AlwaysYes.exe mentality and guide you towards more selective and self-beneficial agreements.
• Practice mental simulations of "CanCrash.exe" events during downtime. When you're waiting in line or commuting, visualize a system you depend on failing and mentally walk through the steps you would take to adapt. This mental exercise can improve your ability to think on your feet and respond effectively to unexpected challenges.
• Create a 'contradiction jar' much like a swear jar, but for self-contradictory statements or actions. Every time you catch yourself saying one thing and doing another, put a small amount of money in the jar. This tangible consequence will make you more aware of inconsistencies. For instance, if you advocate for healthy eating but grab fast food for lunch, you contribute to the jar and acknowledge the contradiction.

Other Perspectives

• The threshold of 10 kilobytes is arbitrary and may not be suitable for all contexts where file size is relevant.
• The behavior of YesOnSelf.exe when given its own code as input may not necessarily result in both affirmative and negative outputs; instead, it could lead to an undefined state or an error due to self-reference, which is not accounted for in the description.
• The argument may not account for probabilistic or heuristic approaches that can effectively detect crashes in practical scenarios, even if they are not perfect.
• The creation of AntiCrashOnSelf.exe assumes that the behavior of CanCrash.exe is fully understood and predictable, which may not be the case in practice.
• The logical progression used to prove the non-existence of CanCrash.exe could contain flaws or oversights that, if corrected, would invalidate the conclusion.

Fundamental Challenges Beyond the Capability of Computers

MacCormick examines the broader implications of undecidable problems, highlighting that while computer programs are incapable of perfectly solving them, that does not prevent researchers from developing practical solutions that address these problems most of the time.

Unsolvable Problems Lack Efficient Solutions

MacCormick acknowledges that while undecidability places a theoretical constraint on computing power, it doesn't always reflect the practical challenges encountered when using computers. He notes that numerous decidable problems, like the Traveling Salesman Problem, lack efficient solutions, rendering them practically intractable for large input sizes despite being theoretically solvable. He further notes that even undecidable problems can frequently be effectively tackled in a more focused way. MacCormick explains that while a program capable of finding all errors in every piece of software is impossible, developers can still create programs that detect the majority of crashes across various kinds of software. This, he argues, highlights the valuable role of partial solutions that address typical cases of practically undecidable problems. These practical solutions have significantly improved the dependability of software in recent decades.

Context

• Understanding undecidability helps computer scientists recognize the boundaries of algorithm design, guiding them to focus on problems that are solvable or to develop partial solutions for those that are not.
• Despite its complexity, TSP has practical applications in logistics, manufacturing, and DNA sequencing, where approximate solutions are often used.
• Partial solutions are valuable because they often address the most common or critical cases, providing significant benefits in terms of efficiency and reliability. This approach is widely used in fields like optimization and artificial intelligence.
• Instead of finding all errors, many tools use heuristics and approximations to identify the most likely or critical errors, accepting that some issues may remain undetected.
• This practice involves frequent testing and updates, allowing developers to catch and fix crashes quickly as new code is integrated into the software.
• These are problems for which no algorithm can be constructed that always leads to a correct yes-or-no answer. An example is the Halting Problem, which asks whether a given program will finish running or continue indefinitely.
• The widespread use of open source software allows for community-driven improvements and peer reviews, which can lead to more reliable and secure software solutions.

Church-Turing Thesis: Human Thought and Computational Limits

MacCormick delves into the philosophical implications of undecidability by questioning if the brain has the same limits as computers. He explains that if a machine were able to perfectly replicate the brain, then the Church-Turing hypothesis would suggest that the brain's computational abilities are equivalent to those of a computer. This would imply that humans are also incapable of resolving unsolvable issues.

MacCormick then acknowledges the contentious nature of the debate surrounding simulating the brain with computers. While no obvious physical constraints exist based on our current knowledge of the brain, different philosophical theories propose that the conscious "mind" might be fundamentally different from any computational process, potentially possessing abilities that defy simulation. He reminds the reader that, during the 1930s, "computer" often meant someone who performed calculations. Turing's work "On Computable Numbers" not only made significant contributions to computer science by identifying undecidable problems, but it also presented a strong argument for the equivalence of computations performed by humans and those performed by Turing's abstract "machines," which we would call "computers" today. This equivalence, called the Church-Turing thesis, holds that any computer and potentially humans share the same computational capabilities.

MacCormick emphasizes the profound implications of the Church-Turing principle, noting that if its strongest version proves true, then the limitations of undecidability would apply not only to computers but also to human thought, implying that we too are bounded by these computational limits. He concludes by acknowledging the ongoing debate about whether the Church-Turing Thesis is true, emphasizing the profound implications it might have on our understanding of human intelligence and its relationship to the computational capabilities of computers.

Context

• This theory suggests that cognitive processes are a form of computation. If true, it implies that the brain operates similarly to a computer, processing information through algorithms and data structures.
• A Turing machine is a theoretical construct that manipulates symbols on a strip of tape according to a set of rules. It is used to model the logic of any computer algorithm, serving as a fundamental concept in the theory of computation.
• Some theorists propose that quantum computing might overcome certain classical computational limits. However, whether this extends to solving undecidable problems remains speculative and is a topic of ongoing research.
• Simulating the brain raises ethical questions about identity, consciousness, and the potential creation of sentient machines. These concerns add another layer to the debate.
• This philosophical view, notably associated with René Descartes, posits that the mind and body are distinct entities. The mind, or consciousness, is non-physical and cannot be fully explained by physical processes or computational models.
• Human decision-making often involves emotions and intuition, which are not easily reducible to logical algorithms. These aspects of cognition might involve processes that are not straightforward to simulate.
• The work of human computers influenced the development of programming languages and early computer science principles. Their methods and organizational skills informed the design of early software and hardware systems.
• While the thesis provides a theoretical framework, practical limitations such as processing power, memory, and time can affect what computers can actually compute in real-world scenarios.
• While influential, the Church-Turing thesis is not a formal theorem that can be proven; it is a hypothesis about the nature of computation and its relationship to human thought.
• The development of AI and machine learning has intensified discussions about the thesis, as these technologies increasingly mimic human cognitive functions, prompting questions about the nature and limits of machine intelligence.