<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Ajay Shrestha on Medium]]></title>
        <description><![CDATA[Stories by Ajay Shrestha on Medium]]></description>
        <link>https://medium.com/@ajay-shrestha?source=rss-658421888120------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*3CatWyBMDccfFkI5M6RHwQ.jpeg</url>
            <title>Stories by Ajay Shrestha on Medium</title>
            <link>https://medium.com/@ajay-shrestha?source=rss-658421888120------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 07:07:57 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ajay-shrestha/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[AI Didn’t Come From Computer Science Alone]]></title>
            <link>https://ajay-shrestha.medium.com/ai-didnt-come-from-computer-science-alone-96fefeb5a99e?source=rss-658421888120------2</link>
            <guid isPermaLink="false">https://medium.com/p/96fefeb5a99e</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[philosophy]]></category>
            <category><![CDATA[neuroscience]]></category>
            <dc:creator><![CDATA[Ajay Shrestha]]></dc:creator>
            <pubDate>Sun, 08 Mar 2026 21:07:45 GMT</pubDate>
            <atom:updated>2026-03-09T01:51:30.278Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong>How Psychology, Neuroscience, Philosophy and Other Fields Shaped Modern AI</strong></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*YtQaPIhq5OGmpkNK" /><figcaption>Photo by <a href="https://unsplash.com/@tjsocoz?utm_source=medium&amp;utm_medium=referral">Tim Bish</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a> (Yale Beinecke Library)</figcaption></figure><p>AI is often portrayed as a product of computer science and maths alone. A closer look reveals many of its central ideas came from psychology, neuroscience, philosophy, and even biology. Several early AI pioneers were even trained outside traditional computer science. The goal of this post is to highlight how some of these other disciplines contributed to what we now know as modern AI.</p><h3>Psychology</h3><p>The early prototype of the artificial neural network architecture, the perceptron, was built by Frank Rosenblatt, a psychologist. His work attempted to model how biological neurons could enable machines to learn from experience instead of being pre-programmed. Early AI, including the much-touted IBM Deep Blue in the 1990s, relied largely on handcrafted rules and brute-force search. This approach wasn’t really novel and definitely didn’t scale.</p><p>The shift toward artificial neural network–based learning from data and feedback marked a departure from rigid logic toward adaptive systems that improve with experience (data). That shift did not come from computer science alone. It came from borrowing models of learning, perception, and adaptation from other disciplines, including psychology.</p><p>The works of Noam Chomsky and B. F. Skinner, major figures in two branches of psychology, i.e., cognitive psychology and behavioral psychology respectively, also played important roles in AI’s development, but in opposing ways. Chomsky heavily criticized Skinner’s ideas as inadequate. Chomsky argued that intelligence requires internal mental structures, while Skinner argued that behavior can be shaped through reinforcement from the environment.</p><p>Decades later, elements of both perspectives appeared in AI. Chomsky emphasized internal mental structures, a concept that parallels the idea of internal representations learned by deep neural networks. Skinner’s work on reinforcement and behavior influenced ideas that later shaped reinforcement learning (RL), another major branch of modern AI. Deep learning (DL) focuses on learning internal representations from data and underpins much of modern AI, while reinforcement learning trains agents through reward and punishment mechanisms and is goal or outcome based.</p><p>Google DeepMind’s AlphaGo used both deep learning (DL) and reinforcement learning (RL) to defeat the world’s number 2 champion in the Chinese board game Go in 2016. Both Chomsky’s and Skinner’s psychology theories ultimately inspired complementary AI approaches.</p><h3><strong>Neuroscience</strong></h3><p>Artificial neural networks, the internal structure of modern AI, are directly inspired by the brain. Artificial neurons, layers, and weighted connections abstract biological neural structures and synapses. The main idea is that intelligence emerges from interconnected units that learn through adjustment.</p><p>In the 1960s, Marvin Minsky at MIT asked his students to solve a computer vision problem over the summer. The assumption was that perception could be reduced to a computer science, i.e., logical, rules-based, if-then-else coding problem. Biological vision reflects millions of years of evolution and immense neural complexity. Minsky underestimated the complexity of the problem.</p><p>A reasonable solution to his problem emerged fifty years later at the ImageNet competition, when a deep neural network significantly outperformed its competitors. The breakthrough was not due to better handcrafted rules. It came from new learning algorithm, hardware and layered neural architectures inspired by biology.</p><h3><strong>Philosophy</strong></h3><p>Alan Turing set a lofty goal in 1950 when he defined a test of machine intelligence. If we cannot tell whether the entity on the other side of a conversation is human or machine, the machine could be considered intelligent.</p><p>Since many foundational large language models (LLMs) now appear to pass this test, Mustafa Suleyman has proposed a new benchmark. His version of the Turing test asks whether an AI system could turn a $100,000 investment into $1 million on a retail platform within a few months.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UeHwQZsnqZDdnmAIKcj_ZA.png" /><figcaption>Figure 1: Bloom’s taxonomy organized as a pyramid of intelligence levels | Source links <a href="https://en.wikipedia.org/wiki/Bloom%27s_taxonomy#/media/File:Bloom%27s_revised_taxonomy.svg">1</a>/<a href="https://en.wikipedia.org/wiki/Bloom%27s_taxonomy">2</a> (CC BY 4.0)</figcaption></figure><p>Frameworks such as Bloom’s Taxonomy describe a hierarchy of cognitive abilities, ranging from remembering and understanding to analyzing, evaluating, and creating. With the advent of generative AI, AI systems can now perform tasks across all levels of this hierarchy.</p><p>To some, the ultimate Turing test would be AI gaining sentience. Yann LeCun has suggested that consciousness may be an emergent property of computational systems. Others remain skeptical and argue that intelligence and consciousness are fundamentally different phenomena.</p><h3><strong>Statistics</strong></h3><p>One could argue that machine learning is applied statistics at scale. During the pre-deep learning phase, AI was dominated by probabilistic models, Bayesian inference, and statistical learning theory. The ability to generalize from data, rather than simply memorize it, comes directly from statistical. Traditional statistical methods often struggled with very large datasets or complex, nonlinear relationships between variables. Deep learning addressed these challenges. Statistics principles nonetheless provided the foundation for modern AI.</p><h3><strong>Evolutionary Biology and Nature</strong></h3><p>Biology and nature have solved extremely complex problems over millions of years. It is therefore no surprise that they have influenced many meta-heuristic algorithms, such as genetic algorithms, swarm intelligence, and ant colony optimization. These algorithms are commonly used to solve hard problems in computer science and AI. Instead of explicitly pre-programming solutions, these algorithms perform the right amount of explorational and exploitation to arrive as near optimal solutions for hard problems. My PhD dissertation explored this idea by using a genetic algorithm to optimize neural network architectures and hyperparameters. Here is the link to it if you are interested.</p><h4>My PhD Dissertation (AI) — 2019</h4><p><a href="https://www.proquest.com/docview/2322784986/631422EFED204CFDPQ/">Enhanced Deep Network Designs Using Mitochondrial DNA Based Genetic Algorithm and Importance Sampling</a></p><h3><strong>Summary</strong></h3><p>The term “AI” was coined in the 1950s, and the field has gone through multiple boom and bust cycles since then. A number of researchers drew from different spheres of human knowledge and persisted through several AI winters, often with little funding and many skeptics.</p><p>These efforts culminated in the breakthrough moment of the 2012 ImageNet competition, when a deep learning model developed by Geoffrey Hinton’s team dramatically outperformed competitors. Hinton, who won the Nobel Prize in 2024 and often called the “godfather of AI,” earned his bachelor’s degree in psychology before later completing a PhD in AI. He has frequently credited his psychology background as an important influence on his work in machine learning.</p><p>The timeline of AI also reminds us that major breakthroughs require time and the right conditions. Without advances in computing hardware such as GPUs and the availability of large datasets like ImageNet, Hinton’s breakthrough might have taken much longer.</p><p>The generalized take away from this post is that breakthroughs often emerge from cross domain knowledge, cross functional collaboration, patience, and the alignment of the right conditions.</p><h3>Related Post</h3><p><a href="https://blog.ajayshrestha.com/machine-learning-vs-software-development-4b23cc182742?source=list---------0-------769ecf09d3ea----------------------------">Machine Learning Vs Software Development</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=96fefeb5a99e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Money Protocol in the age of AI]]></title>
            <link>https://ajay-shrestha.medium.com/money-protocol-in-the-age-of-ai-455f81b516ef?source=rss-658421888120------2</link>
            <guid isPermaLink="false">https://medium.com/p/455f81b516ef</guid>
            <category><![CDATA[defi]]></category>
            <category><![CDATA[finance]]></category>
            <category><![CDATA[bitcoin]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Ajay Shrestha]]></dc:creator>
            <pubDate>Wed, 31 Dec 2025 21:58:22 GMT</pubDate>
            <atom:updated>2026-01-14T03:34:12.952Z</atom:updated>
            <content:encoded><![CDATA[<h4>What will AI Agents use for store of value and payments?</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*dRXLnR47RWev4doA" /><figcaption>Photo by <a href="https://unsplash.com/@kevinrajaram?utm_source=medium&amp;utm_medium=referral">Kevin Rajaram</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Yuval Noah Harari wrote in <em>Sapiens</em> that money is an <em>inter-subjective fiction</em>, i.e., a shared belief that enables humans to cooperate at scale. Money is a social construct, but a powerful one. It has allowed humans to trade, coordinate, and build complex societies and civilizations.</p><p>What constitutes money has changed over time. It evolved from collectible objects such as seashells, to metal coins, gold coins, gold-backed paper currencies and eventually fiat currencies. The defining property of money is collective belief and trust that it will be accepted in exchange for goods and services, and its supply is somewhat constrained (scarce or controlled).</p><p>That trust can come from different sources. It may be backed by centralized authorities like governments and central banks (fiat currencies), derived from natural scarcity and production costs (gold), or enforced by decentralized consensus and software (bitcoin). Gold, for example, is scarce and requires significant energy to mine, which contributed to its durability as money across cultures for thousands of years. Fiat currencies, on the other hand, reflect trust in and enforcement by government.</p><p>Historically, objects tend to become money through a gradual evolution of utility, first as a store of value, then as a medium of exchange, and eventually as a unit of account. In addition to scarcity, sound money also exhibits properties such as divisibility, portability, durability, fungibility, and verifiability.</p><p>On top of base money, layers of financial infrastructure emerge to support payments and financial services at scale. When we swipe a credit card or send an international wire, multiple intermediaries exchange information, update ledgers, and wait for final settlement. A wide range of financial products from loans, equities, derivatives, savings instruments, all reference the base money, serving users with different risk tolerances, time preferences, and return expectations.</p><p>So far, humans have been firmly in the driver’s seat of this system. But what happens when an AI-agent-to-AI-agent economy begins to take shape? These agents will increasingly operate autonomously, with the ability to make decisions, custody money, and execute transactions with minimal human involvement. They may exchange value with each other in return for services rendered, e.g., API calls, data access, compute, or other services.</p><p>This raises a natural question: <strong>what will AI agents use as money?</strong></p><p>One option is simply to plug AI agents into existing financial systems, bank accounts, credit cards, Apple Pay, Google Pay and call it a day. But current fiat-based payment rails were designed for human institutions, not autonomous machines. For example, one AI agent might charge another $0.001 for an API call, a transaction that is impractical on traditional financial rails. Beyond this, AI agents may encounter several additional frictions.</p><ol><li><strong>Slow final settlement<br>- </strong>Credit card payments typically take 1–3 business days to settle<br>- International SWIFT wires can take 1–5 business days</li><li><strong>Lack of viable micropayments<br>- </strong>Transactions below a cent are impractical or impossible</li><li><strong>Excessive fees<br>- </strong>Credit card processing fees, wire fees, overdraft fees</li><li><strong>Cross-border friction<br>- </strong>Multiple intermediaries, holidays, and limited operating hours</li><li><strong>Limited access<br>- </strong>Around 6 million people in the US and over 1 billion globally are unbanked [1]</li><li><strong>Inflation and monetary risk<br>- </strong>20–40 countries experience double-digit inflation [2]<br>- Even 2% annual inflation compounded over 100 years results in an 86% loss of purchasing power</li><li><strong>Centralized control<br>- </strong>Policy failures, mismanagement, single points of failure, and exclusionary practices</li></ol><p>For autonomous agents optimizing for speed, cost, and reliability, these frictions matter.</p><p>Alternatives to fiat include bitcoin and stablecoins. While the internet enabled peer-to-peer transfer of information without centralized intermediaries, bitcoin was the first system to enable peer-to-peer transfer of value without a trusted intermediary. It also solved the problem of digital scarcity as there will only ever be 21 million bitcoins, i.e., something no previous technology had achieved.</p><p>Stablecoins, particularly considering regulatory developments such as the GENIUS Act [3], may also address some inefficiencies of traditional finance by enabling faster settlement and lower-cost transfers while maintaining price stability. Each option comes with trade-offs: bitcoin offers censorship resistance and a fixed supply; stablecoins introduce issuer and regulatory trust; and traditional finance largely represents the status quo.</p><p>Like language, money exhibits strong network effects and tends to converge toward dominant standards. This has been true for the base protocol of the internet, i.e., TCP/IP and may also apply to AI systems. E.g., in the early days of the Internet, there were number of competing protocols, TCP/IP eventually won and the rest is history.</p><p>AI interoperability is already converging around open standards. Major technology companies including Google, OpenAI, Microsoft, and Anthropic have aligned around the Model Context Protocol (MCP) as a common interface for AI agent communication [4]. MCP defines how agents exchange context and instructions, not value, but it can support multiple payment methods, from traditional banking rails to bitcoin and stablecoins.</p><p>The rapid adoption of MCP following its introduction by Anthropic in 2024 suggests how quickly the industry can coalesce around a shared protocol. A similar convergence could occur around a “money protocol” optimized for machine-to-machine transactions.</p><p>As the machine economy ramps up, the question of what agents will use as money remains open. If money is an <em>inter-subjective fiction</em>, and the custody and exchange of value shifts from humans to autonomous AI agents, what belief system will those AI agents converge on and what will they ultimately accept as money?</p><h4>Related Blogposts</h4><ul><li><a href="https://blog.ajayshrestha.com/blockchain-trilemma-a-framework-to-evaluate-digital-assets-ed31905b0030">Blockchain Trilemma: A framework to evaluate Digital Assets</a></li><li><a href="https://medium.com/data-science/bitcoins-technical-contribution-solving-byzantine-general-s-problem-f0449973437c">Bitcoin’s technical contribution: Solving Byzantine General’s Problem</a></li></ul><p><strong>References:</strong></p><ol><li><a href="https://www.fastcompany.com/91014699/our-world-isnt-designed-for-the-growing-number-of-unbanked-people">https://www.fastcompany.com/91014699/our-world-isnt-designed-for-the-growing-number-of-unbanked-people</a></li><li><a href="https://worldpopulationreview.com/country-rankings/inflation-rate-by-country">https://worldpopulationreview.com/country-rankings/inflation-rate-by-country</a></li><li><a href="https://en.wikipedia.org/wiki/GENIUS_Act">https://en.wikipedia.org/wiki/GENIUS_Act</a></li><li><a href="https://en.wikipedia.org/wiki/Model_Context_Protocol">https://en.wikipedia.org/wiki/Model_Context_Protocol</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=455f81b516ef" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Can AI Make Mental Healthcare More Accessible?]]></title>
            <link>https://everestaijournal.com/can-ai-make-mental-healthcare-more-accessible-98f240ee773a?source=rss-658421888120------2</link>
            <guid isPermaLink="false">https://medium.com/p/98f240ee773a</guid>
            <category><![CDATA[digital-health]]></category>
            <category><![CDATA[mental-health]]></category>
            <category><![CDATA[health]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Ajay Shrestha]]></dc:creator>
            <pubDate>Tue, 07 Oct 2025 03:14:57 GMT</pubDate>
            <atom:updated>2026-01-04T22:17:43.858Z</atom:updated>
            <content:encoded><![CDATA[<h4>Millions of people lack access to mental healthcare. Can regulated AI help reduce the care gap?</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*GPZDrbU-6opefjgR" /><figcaption>Photo by <a href="https://unsplash.com/@totalshape?utm_source=medium&amp;utm_medium=referral">Total Shape</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p><strong>The Access Gap</strong></p><p>According to recent reports [1], one in four U.S. adults with a mental illness did not receive treatment between 2022 and 2023. Nearly 40% cited problems with access or a lack of awareness of treatment options [1]. These barriers for millions of Americans are being made worse by growing healthcare deserts, which have made access even more difficult.</p><p>Untreated mental health conditions can lead to devastating outcomes such as homelessness, unemployment, incarceration, and even suicide. When treatment begins late, it becomes more expensive and less effective. Improving access to care at the onset is therefore not just critical, but also an economic and public health necessity.</p><p><strong>The State of U.S. Mental Healthcare</strong></p><p>Despite spending almost one-fifth of its GDP on healthcare, the United States ranks last among high-income nations in healthcare performance [2]. Mental health accounts for 6% of total healthcare spending, i.e., one of the lowest shares among developed countries. Rising demand has not been met with adequate supply [3,4]. Provider shortages, out-of-pocket costs, lack of awareness, and stigma around mental illness continue to limit care.</p><p><strong>Can AI Help?</strong></p><p>Currently, there are no FDA-approved mental health chatbots for diagnostics or treatment [5], though promising research is underway, such as a clinical trial out of Dartmouth College [6]. FDA clearance is essential to ensure the safety and efficacy of these tools.</p><p>AI-driven mental health systems, especially when combined with tele-therapy and human oversight, could help reduce the care gap. AI can assist with triage<strong>, </strong>assessment<strong>, </strong>documentation<strong>,</strong> and personalized<strong> </strong>interventions, enabling clinicians to serve more patients efficiently.</p><p>There are a few FDA-authorized digital CBT (clinical behavioral therapy) apps [7,8]. Several other startups, including<strong> </strong>Woebot Health<strong> </strong>and<strong> </strong>NeuroFlow<strong>,</strong> are already using AI-assisted platforms to support providers in clinical decision-making. However, these systems are not yet FDA-cleared and do not directly deliver therapy.</p><p><strong>The Path Forward</strong></p><p>AI alone isn’t the answer, but regulated, ethically designed AI, integrated into human-led care, could help scale access safely. The key will be maintaining clinical oversight, ensuring data privacy, and building trust between patients, providers, and technology. With careful governance and right safeguards, AI could improve mental healthcare by making it more accessible.</p><blockquote>Disclaimer: The author doesn’t have mental healthcare background. The post is not meant to be healthcare advice.</blockquote><p><strong>References</strong></p><ol><li><a href="https://mhanational.org/wp-content/uploads/2025/09/State-of-Mental-Health-2025.pdf">https://mhanational.org/wp-content/uploads/2025/09/State-of-Mental-Health-2025.pdf</a></li><li><a href="https://www.commonwealthfund.org/publications/fund-reports/2024/sep/mirror-mirror-2024">https://www.commonwealthfund.org/publications/fund-reports/2024/sep/mirror-mirror-2024</a></li><li><a href="https://www.commonwealthfund.org/publications/explainer/2022/sep/behavioral-health-care-us-how-it-works-where-it-falls-short">https://www.commonwealthfund.org/publications/explainer/2022/sep/behavioral-health-care-us-how-it-works-where-it-falls-short</a></li><li><a href="https://www.commonwealthfund.org/publications/2024/may/mental-health-needs-us-compared-nine-other-countries">https://www.commonwealthfund.org/publications/2024/may/mental-health-needs-us-compared-nine-other-countries</a></li><li><a href="https://www.apaservices.org/practice/business/technology/artificial-intelligence-chatbots-therapists">https://www.apaservices.org/practice/business/technology/artificial-intelligence-chatbots-therapists</a></li><li><a href="https://home.dartmouth.edu/news/2025/03/first-therapy-chatbot-trial-yields-mental-health-benefits">https://home.dartmouth.edu/news/2025/03/first-therapy-chatbot-trial-yields-mental-health-benefits</a></li><li><a href="https://blogs.wysa.io/blog/research/wysa-receives-fda-breakthrough-device-designation-for-ai-led-mental-health-conversational-agent">https://blogs.wysa.io/blog/research/wysa-receives-fda-breakthrough-device-designation-for-ai-led-mental-health-conversational-agent</a></li><li><a href="https://www.psychiatryadvisor.com/features/prescription-digital-therapeutics/">https://www.psychiatryadvisor.com/features/prescription-digital-therapeutics/</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=98f240ee773a" width="1" height="1" alt=""><hr><p><a href="https://everestaijournal.com/can-ai-make-mental-healthcare-more-accessible-98f240ee773a">Can AI Make Mental Healthcare More Accessible?</a> was originally published in <a href="https://everestaijournal.com/">Everest AI Journal</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Blockchain Trilemma: A framework to evaluate Digital Assets]]></title>
            <link>https://ajay-shrestha.medium.com/blockchain-trilemma-a-framework-to-evaluate-digital-assets-ed31905b0030?source=rss-658421888120------2</link>
            <guid isPermaLink="false">https://medium.com/p/ed31905b0030</guid>
            <category><![CDATA[bitcoin]]></category>
            <category><![CDATA[cryptography]]></category>
            <category><![CDATA[digital-asset]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[systems-engineering]]></category>
            <dc:creator><![CDATA[Ajay Shrestha]]></dc:creator>
            <pubDate>Thu, 23 Jan 2025 15:55:09 GMT</pubDate>
            <atom:updated>2025-05-09T03:38:20.013Z</atom:updated>
            <content:encoded><![CDATA[<h4>The Design Trade-offs: Decentralization, Security and Scalability</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ZCmSFlRBqKbVymQC" /><figcaption>Photo by <a href="https://unsplash.com/@onefifith?utm_source=medium&amp;utm_medium=referral">Shubham Dhage</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>The goal of this post is to provide a framework based on the “Blockchain Trilemma” to differentiate the large number of crypto projects out there.</p><p>The Blockchain Trilemma, a term coined by Vitalik Buterin (co-founder of Ethereum), is derived from a well-established design constraint called the CAP Theorem in systems engineering. CAP Theorem states that distributed databases can only achieve two of the following three capabilities: Consistency, Availability, and Partition Tolerance. Below, we will explore how this idea is adapted to blockchain systems, which are cryptographically secured, immutable distributed ledger databases.</p><h4><strong>Blockchain Trilemma</strong></h4><p>The Blockchain Trilemma highlights the challenge of achieving three desired attributes of blockchain design simultaneously. Most blockchain projects make trade-offs to prioritize one or two attributes based on their intended use case.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/768/1*74hDu3O6_hN77p1MHKLhNA.png" /><figcaption>Fig1: Blockchain Trilemma in Venn Diagram</figcaption></figure><ol><li><strong>Decentralization:</strong> Ensures the network is distributed across many nodes, preventing control by a small group of entities.</li><li><strong>Security:</strong> Protects the network from attacks and ensures data integrity.</li><li><strong>Scalability:</strong> Enables the network to process a large number of transactions per second (TPS).</li></ol><h4><strong>Decentralization</strong></h4><p>Decentralization is the defining feature of blockchains, distinguishing them from traditional databases. Without decentralization, applications with regular centralized databases and software systems are typically more efficient.</p><p>The <strong>Nakamoto Coefficient</strong>, a metric originally proposed by former Coinbase CTO Balaji Srinivasan, measures the degree of decentralization in a network. It calculates the minimum number of entities (or nodes/operators) required to compromise the blockchain. E.g., a blockchain with a high Nakamoto Coefficient is more decentralized and resilient to collusion or attacks. Likewise, a project with low Nakamoto Coefficient means users have to place their trust on a smaller number of entities or people to manage the network and is therefore more vulnerable to compromise.</p><h4>Security</h4><p>Security is crucial to ensuring that incorrect entries do not corrupt the distributed ledger. This is enforced through consensus mechanisms that determine how blocks (consisting of transactions) are added to the blockchain. Common protocols include:</p><ol><li><strong>Proof of Work (PoW):</strong> Requires miners to spend compute power and compete with each other to be the first to solve a cryptographic function (SHA-256) to win the race to add the next block to the chain, and receive a reward. It is energy-intensive but highly secure, as seen in Bitcoin. This is the most secure of the three.</li><li><strong>Proof of Stake (PoS): </strong>Validators are selected based on the amount of cryptocurrency they hold and “stake.” PoS is less energy-intensive but introduces concerns about centralization and tight coupling between ownership of the underlying tokens and governance (i.e., change rules for token issuance, max token supply, voting, consensus, etc.). PoS is commonly used by many projects, including Ethereum.</li><li><strong>Proof of History (PoH):</strong> Utilized by Solana, this mechanism timestamps transactions to ensure chronological order and optimize speed. This is the least secure of the three.</li></ol><h4><strong>Scalability</strong></h4><p>Scalability determines whether a blockchain can process transactions quickly and cost-effectively. For blockchains to compete with traditional payment networks like Visa and Mastercard (handle thousands of TPS), achieving scalability is critical. Decentralization and security are negatively correlated to scalability.</p><h4>Summary</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*F9HvKvGqqyu_Raz0GClyMQ.png" /><figcaption>Table1: Blockchain Trilemma and common projects</figcaption></figure><h4>Potential Use Cases and Enhancements</h4><p>Projects prioritize design choices based on their intended purpose. Given the difficulty of achieving decentralization, security, and scalability simultaneously, many blockchain projects adopt a layered approach. This strategy attempts to compensate for the missing feature or extend the desired feature even more with additional layers. Potential applications could be:</p><ol><li><strong>Store of Value</strong> (Bitcoin)<strong>: </strong>Emphasis on decentralization and security makes Bitcoin ideal for store of value (also due to max supply of 21 million) or high-value financial settlements. For scalability and micropayments, Layer 2 solutions like the Lightning Network enable up to one million transactions per second.</li><li><strong>Smart Contracts </strong>(Ethereum)<strong>: </strong>Favors scalability and code functionality (Turing complete) and compromises on decentralization and security. Additional scalability is provided with layer 2 solutions.</li><li><strong>High-Frequency Trading </strong>(Solana): Focuses on scalability while compromising decentralization and security.</li></ol><p>By understanding the trade-offs inherent in each blockchain’s design, developers and investors can better align their expectations with a project’s goals and potential use cases.</p><h4>Conclusion</h4><p>The Blockchain Trilemma provides a valuable framework for evaluating the strengths, weaknesses, and trade-offs of blockchain projects. While no project can optimize all three attributes simultaneously, the layered approach allows developers to address these challenges to a certain extent. By analyzing decentralization, security, and scalability, one can better understand the potential, limitations, and risks of a blockchain project.</p><h4>Disclaimer</h4><p>This post is a very high-level review of blockchain projects from a technical design perspective and should not be taken as financial advice in any form.</p><p>For a slightly deeper dive into Bitcoin, checkout the post linked below:</p><p><a href="https://medium.com/towards-data-science/bitcoins-technical-contribution-solving-byzantine-general-s-problem-f0449973437c">Bitcoin’s technical contribution: Solving Byzantine General’s Problem</a></p><h4>References</h4><ol><li><a href="https://www.fidelitydigitalassets.com/research-and-insights/bitcoin-first-revisited">https://www.fidelitydigitalassets.com/research-and-insights/bitcoin-first-revisited</a></li><li><a href="https://en.wikipedia.org/wiki/Lightning_Network">https://en.wikipedia.org/wiki/Lightning_Network</a></li><li><a href="https://en.wikipedia.org/wiki/CAP_theorem">https://en.wikipedia.org/wiki/CAP_theorem</a></li><li><a href="https://en.wikipedia.org/wiki/Eric_Brewer_(scientist)">https://en.wikipedia.org/wiki/Eric_Brewer_(scientist)</a></li><li><a href="https://news.earn.com/quantifying-decentralization-e39db233c28e">https://news.earn.com/quantifying-decentralization-e39db233c28e</a></li><li><a href="https://river.com/learn/how-bitcoin-can-scale/">https://river.com/learn/how-bitcoin-can-scale/</a></li><li><a href="https://tokentax.co/blog/solana-vs-ethereum">https://tokentax.co/blog/solana-vs-ethereum</a></li><li><a href="https://solana.com/news/validator-health-report-october-2023">https://solana.com/news/validator-health-report-october-2023</a></li><li><a href="https://io-fund.com/crypto/blockchain/why-solana-is-outperforming-ethereum-by-26500-since-2020">https://io-fund.com/crypto/blockchain/why-solana-is-outperforming-ethereum-by-26500-since-2020</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ed31905b0030" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Top 18 books completed in 2024]]></title>
            <link>https://ajay-shrestha.medium.com/top-18-books-completed-in-2024-3242236cd50f?source=rss-658421888120------2</link>
            <guid isPermaLink="false">https://medium.com/p/3242236cd50f</guid>
            <category><![CDATA[business]]></category>
            <category><![CDATA[books]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Ajay Shrestha]]></dc:creator>
            <pubDate>Sun, 29 Dec 2024 20:45:20 GMT</pubDate>
            <atom:updated>2024-12-29T21:05:47.401Z</atom:updated>
            <content:encoded><![CDATA[<h4>Learnings in AI, Management, Business, Science and Wellbeing</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yO7xLvsSUsWSAYiF-ZUypA.png" /></figure><p>These 18 books stand out from those I completed this year. I combined reading (physical books and Kindle) with listening (on Audible). Hopefully, you’ll find a couple from this list that pique your interest and pick them up for 2025. While these books span various disciplines, they can be broadly categorized into:</p><ul><li>Technology (AI)</li><li>Management</li><li>Business</li><li>Science &amp; Wellbeing</li></ul><p>Below is a short reflection/summary of the books by category.</p><h3><strong>Technology (AI)</strong></h3><p>We are at an inflection point in human history where technology (AI and bioengineering) will have an unprecedented impact on our lives. These books explore the history, current state, and potential future of AI.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/310/1*LTYiT_oACgzEZYYMeUWkEg.png" /></figure><h4><em>The Coming Wave</em> by Mustafa Suleyman</h4><p>Mustafa Suleyman, co-founder of DeepMind and CEO of Microsoft AI, provides a thorough examination of the transformative potential and inherent risks of AI and synthetic biology. Suleyman argues that we are at a historical moment where technology is advancing at lightning speed, presenting two potential futures: authoritarian surveillance states or lawless anarchy — unless we solve the AI containment problem. Checkout this <a href="https://www.technologyreview.com/2023/07/14/1076296/mustafa-suleyman-my-new-turing-test-would-see-if-ai-can-make-1-million/">article</a> for a sense of the rate of advancement Suleyman is talking about. A recommended read for anyone seeking to understand the profound impacts of AI and biotech beyond the headlines.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/314/1*8xvTjzsX0fzua7ZBxrW70g.png" /></figure><h4>The Age of AI by Henry A. Kissinger, Eric Schmidt, and Daniel Huttenlocher</h4><p>This book offers an interdisciplinary perspective on AI’s impending impact on society, combining political, business, and scientific insights from the authors’ backgrounds. It underscores the urgent need for education and policy frameworks to navigate AI’s challenges and opportunities.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/310/1*xW3woMHPqJeokkAazpxjfQ.png" /></figure><h4>Nexus by Yuval Noah Harari</h4><p>Harari revisits his recurring theme that solving large-scale global problems requires large-scale cooperation and order, which depends on shared inter-subjective fictions — stories existing only in our minds. In Nexus, he explores how these narratives intertwine with information networks, like the printing press, mass media, and now AI, with profound implications for the future of governments (totalitarian to democratic) and societies (ranging from human flourishing and order to anarchy or even extinction). According to Harari, unlike previous technologies, AI can generate new ideas and make autonomous decisions, intensifying the challenge of aligning it with human values and goals.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/304/1*L-dIw6eqmFX1XWe07KfyMQ.png" /></figure><h4>Deep Medicine by Eric Topol, MD</h4><p>Topol delves into AI’s potential to revolutionize healthcare, enhancing diagnosis, treatment, and personalized care while allowing doctors to focus more on empathy and patient relationships.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/310/1*cijOIi0-9AHWFoyUKEj4Uw.png" /></figure><h4>The World I See by Dr. Fei-Fei Li</h4><p>Dr. Fei-Fei Li, a Stanford professor and AI pioneer, shares her inspiring journey as an immigrant and her pivotal contributions to the advancement of artificial intelligence through the groundbreaking <a href="https://en.wikipedia.org/wiki/ImageNet">ImageNet</a> project. ImageNet provided the foundation for training computer vision models and catalyzed one of the 21st century’s most significant AI breakthroughs: AlexNet’s victory in the 2012 ImageNet competition. This milestone is widely regarded as the turning point that marked the end of the so-called “<a href="https://en.wikipedia.org/wiki/AI_winter">AI winter</a>.” The book weaves together a deeply personal narrative with a historical exploration of AI’s evolution, offering profound insights into the field’s development.</p><h3>Management</h3><p>These four tech titans represent different management styles: Bezos’s customer obsession, Musk’s visionary audacity, Nadella’s focus on growth mindset, and Grove’s adaptability in the face of disruption.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/304/1*j9s9IoFeY-iR6WAnqY3WOg.png" /></figure><h4>Invent and Wander by Jeff Bezos &amp; Walter Isaacson</h4><p>A collection of Bezos’s writings, including his annual shareholder letters, this book provides insights into Amazon’s <a href="https://www.amazon.jobs/content/en/our-workplace/leadership-principles">leadership principles</a>, innovation strategies, and customer-centric philosophy.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/308/1*dOqpTey_ZNwMk9hTfKPH8A.png" /></figure><h4>Elon Musk by Walter Isaacson</h4><p>This biography delves into Musk’s journey, from his early years in South Africa to his revolutionary ventures like SpaceX and Tesla. It highlights his intense work ethic, ambitious vision, and distinctive approach to innovation. The book also outlines practical insights, such as Musk’s 5-step process, often called the “<a href="https://x.com/StartupArchive_/status/1872625977672831146">Musk Algorithm</a>,” for improving processes and outcomes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/308/1*lxURRiX-uUYYzx3fcmXf1g.png" /></figure><h4>Hit Refresh by Satya Nadella</h4><p>Nadella’s memoir reflects on his journey at Microsoft, his emphasis on empathy, and the role of innovation in revitalizing the company. Key themes include adopting a growth mindset and leading with empathy. Nadella cites making the right call to advance cloud computing and AI, and now quantum computing, as these key technologies are already making an impact and will continue to shape the future.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/322/1*-ED9BWw0OiAp9fs3y0lHgQ.png" /></figure><h4>Only the Paranoid Survive by Andrew Grove</h4><p>Grove discusses how businesses can navigate strategic inflection points — moments of profound change that can make or break a company — emphasizing vigilance, adaptability, and leadership.</p><h3>Business</h3><p>Businesses require the right alignment of vision, mission, strategy, execution, operations and market. These books offer valuable models and frameworks to guide organizations toward success.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/314/1*iWLCPqDWWweUsPO_mPJlnA.png" /></figure><h4>Blue Ocean Strategy by W. Chan Kim and Renée Mauborgne</h4><p>This book introduces the concept of Blue Ocean markets (uncontested spaces) versus Red Ocean markets (highly competitive spaces). It emphasizes creating new markets through value innovation (not necessarily a vertically integrated disruption like the one described in Peter Thiel’s book <em>Zero to One</em>), using examples like Apple’s iTunes, which de-segmented the music market by allowing users to purchase individual songs rather than albums.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/312/1*JLPXftbhEoIGqqebpuNEaA.png" /></figure><h4>Start With Why by Simon Sinek</h4><p>Sinek uses the Golden Circle framework (“Why,” “How,” “What”) to explain how successful companies inspire loyalty and achieve long-term success. Apple is cited most frequently in the book as the quintessential example, demonstrating the power of purpose-driven leadership.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/312/1*MG3CUxofGuopuCJ8HYKcJw.png" /></figure><h4>Chip War by Chris Miller</h4><p>This book masterfully combines supply chain, business, and political history, explaining how the semiconductor industry became critical to global economies while exposing vulnerabilities in its reliance on a few key players.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/308/1*cs3dRGU5JWlVIggn0bd4VQ.png" /></figure><h4>The Checklist Manifesto by Atul Gawande, MD</h4><p>Gawande demonstrates the power of checklists in preventing errors and improving outcomes across industries, from surgery to aviation, making a compelling case for their simplicity and effectiveness.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/314/1*w8Xbvsf3f0rpWM2eRtmaxA.png" /></figure><h4>Accelerate by Nicole Forsgren, PhD, Jez Humble, and Gene Kim</h4><p>A data-driven guide to achieving high performance in software delivery and operations, this book identifies key metrics and practices that drive speed and stability in technology teams.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/314/1*2E31WxrRVB6xsrNjBvSMIA.png" /></figure><h4>Negotiation Genius by Deepak Malhotra and Max H. Bazerman</h4><p>This book introduces negotiation strategies like BATNA (Best Alternative to a Negotiated Agreement) and ZOPA (Zone of Possible Agreement), offering tools for achieving win-win outcomes.</p><h3>Science and Wellbeing</h3><p>These books explore advancements in science and psychology, examining how they have contributed to enhancing human well-being and overall quality of life.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/306/1*HrElxzzfscmOLLYQ3oKLpg.png" /></figure><h4>How the World Really Works by Vaclav Smil</h4><p>Smil explores the underpinnings of modern civilization from a scientific and materialistic view highlighting the critical roles of ammonia, steel, concrete, and plastics, while challenging other more subjective narratives about human progress.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/306/1*H2ZK6xkSsxOXba1MTeiRvg.png" /></figure><h4>The Body Keeps the Score by Bessel van der Kolk, MD</h4><p>This book explores how trauma impacts both the body and mind, offering a profound look at its long-lasting effects. Van der Kolk advocates for alternative healing approaches, such as mindfulness, yoga, and community engagement, to help individuals process and recover from trauma, ultimately guiding them toward healing and finding purpose.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/314/1*HNtNxZrUr98_gFzDBmyyDw.png" /></figure><h4>From Strength to Strength by Arthur C. Brooks</h4><p>Brooks provides a roadmap for transitioning from professional success to a more meaningful and purposeful life in later years. He emphasizes the fleeting nature of satisfaction derived from professional achievements and highlights the importance of gratitude, detachment, and cultivating meaningful relationships.</p><h3>Books from 2023</h3><p>Here is my list from last year.</p><p><a href="https://blog.ajayshrestha.com/books-read-in-2023-c315b9a32260">Books Read in 2023</a></p><h3>Tentative List for 2025</h3><p>List of books, I’d like to complete next year.</p><ol><li><em>Superintelligence: Paths, Dangers, Strategies</em> — Nick Bostrom</li><li><em>The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies</em> — Erik Brynjolfsson, Andrew McAfee</li><li><em>Prediction Machines: The Simple Economics of Artificial Intelligence</em> — Ajay Agrawal, Joshua Gans, Avi Goldfarb</li><li><em>Homo Deus: A Brief History of Tomorrow</em> — Yuval Noah Harari</li><li><em>Fooled by Randomness: The Hidden Role of Chance in Life and in the Markets</em>— Nassim Nicholas Taleb</li><li><em>Good to Great: Why Some Companies Make the Leap And Others Don’t</em> — Jim Collins</li><li><em>Continuous Discovery Habits: Discover Products that Create Customer Value and Business Value</em> — Teresa Torres</li><li><em>Teams That Work: The Seven Drivers of Team Effectiveness</em> — Scott Tannenbaum, Eduardo Salas</li><li><em>Radical Candor: Be a Kickass Boss Without Losing Your Humanity</em> — Kim Malone Scott</li><li><em>Think Again: The Power of Knowing What You Don’t Know</em> — Adam Grant</li><li><em>Revenge of the Tipping Point: Overstories, Superspreaders, and the Rise of Social Engineering</em> — Malcolm Gladwell</li><li><em>Breath: The New Science of a Lost Art</em> — James Nestor</li></ol><pre>Disclaimer: I don&#39;t necessarily agree with everything in these books.</pre><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3242236cd50f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Books Read in 2023]]></title>
            <link>https://ajay-shrestha.medium.com/books-read-in-2023-c315b9a32260?source=rss-658421888120------2</link>
            <guid isPermaLink="false">https://medium.com/p/c315b9a32260</guid>
            <category><![CDATA[books]]></category>
            <category><![CDATA[business]]></category>
            <category><![CDATA[personal-development]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[learning]]></category>
            <dc:creator><![CDATA[Ajay Shrestha]]></dc:creator>
            <pubDate>Sun, 31 Dec 2023 06:49:28 GMT</pubDate>
            <atom:updated>2024-08-20T16:26:45.097Z</atom:updated>
            <content:encoded><![CDATA[<h4>12 books that I enjoyed reading &amp; learned from in 2023</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/964/1*7jVfPUdeT8sqN4OL-3Eicw.png" /></figure><p>Reflecting on a year of growth and knowledge! Here are the 12 books that fueled my learning in 2023. Grateful for the insightful perspectives and inspiration these books provided. They can be grouped in the following categories:</p><ol><li>Technical</li><li>Business/Management</li><li>Social Sciences/Philosophy</li><li>Personal Growth</li></ol><p>Below is the full list along with my short take on each.</p><h3>Technical</h3><p><strong>The AI Revolution in Medicine</strong>: GPT-4 and Beyond — <em>Peter Lee , Carey Goldberg , Isaac Kohane</em></p><blockquote>2023 was a breakthrough year for Large Language Models (LLM) in AI. This book does a good job in translating what that means for healthcare providers and patient experience/outcome, now and into the future. <a href="https://www.amazon.com/AI-Revolution-Medicine-GPT-4-Beyond/dp/0138200130?&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=578b78b7eedc051272b875cf452f2ba9&amp;camp=1789&amp;creative=9325">Amazon link</a>.</blockquote><p><strong>The Deep Learning Revolution</strong> — <em>Terrence J. Sejnowski</em></p><blockquote>This is a comprehensive narration of major events (setbacks &amp; breakthroughs) and collaborations in AI -&gt; Machine Learning -&gt; Deep Learning &amp; Reinforcement Learning, since the 1950s. <a href="https://www.amazon.com/Deep-Learning-Revolution-MIT-Press/dp/026203803X/ref=tmm_hrd_swatch_0?_encoding=UTF8&amp;qid=1700435242&amp;sr=1-1&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=93e398f1ae8eeb1f5dac23096a38d092&amp;camp=1789&amp;creative=9325">Amazon link</a>.</blockquote><p><strong>Monolith to Microservices</strong>: Evolutionary Patterns to Transform Your Monolith — <em>Sam Newman</em></p><blockquote>The what, why, how and when (and when not to) to convert monolith application into microservcies based software architecture. <a href="https://www.amazon.com/Monolith-Microservices-Evolutionary-Patterns-Transform/dp/1492047848?&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=26e5056b93ac68c5985134b1eff9704b&amp;camp=1789&amp;creative=9325">Amazon link</a>.</blockquote><p><strong>Software Engineering at Google</strong>: Lessons Learned from Programming Over Time — <em>Titus Winters , Tom Manshreck , Hyrum Wright</em></p><blockquote>Comprehensive book on how Google does software engineering (people, process, tools), while maintaining a mono-repo for (almost) all codebase. <a href="https://www.amazon.com/Software-Engineering-Google-Lessons-Programming/dp/1492082791/ref=tmm_pap_swatch_0?_encoding=UTF8&amp;qid=1695694543&amp;sr=1-1&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=db073c188ea2db21f884f46a143360b0&amp;camp=1789&amp;creative=9325">Amazon link</a>.</blockquote><p><strong>Seven Brief Lessons on Physics </strong>— <em>Carlo Rovelli</em></p><blockquote>A quick read on all things physics from Newton (laws of motion) to Einstein (general relativity) to Neils Bohr &amp; Max Planck (quantum mechanics). <a href="https://www.amazon.com/Seven-Brief-Lessons-Physics-Rovelli/dp/0399184414/ref=tmm_hrd_swatch_0?_encoding=UTF8&amp;qid=1695694747&amp;sr=8-1&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=3905262167e1838ab8a2b55a6962f10c&amp;camp=1789&amp;creative=9325">Amazon link</a>.</blockquote><h3>Business/Management</h3><p><strong>Zero to One</strong>: Notes on Startups, or How to Build the Future — <em>Peter Thiel , Blake Masters</em></p><blockquote>Illuminating book on how novel value creation (that defies competition and creates monopolies) requires novel thinking and not following established processes or trying to repeat earlier/others&#39; success patterns. Amazon link.</blockquote><p><strong>Lean Enterprise</strong>: How High Performance Organizations Innovate at Scale — <em>Matthew Humble , Joanne Molesky , Barry O’Reilly</em></p><blockquote>Authors demonstrate how to incorporate DevOps and Lean-Startup principles to lead an agile organization. <a href="https://www.amazon.com/Lean-Enterprise-Performance-Organizations-Innovate/dp/1492091774/ref=tmm_pap_swatch_0?_encoding=UTF8&amp;qid=&amp;sr=&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=fb5d5149efa957cfad1b2df2d8d12177&amp;camp=1789&amp;creative=9325">Amazon link</a>.</blockquote><p><strong>Learning Agile</strong>: Understanding Scrum, XP, Lean, and Kanban — <em>Andrew Stellman, Jennifer Greene</em></p><blockquote>This book does a good job explaining the different agile methologies and principles. Amazon link.</blockquote><h3>Social Sciences/Philosophy</h3><p><strong>Outliers</strong>: The Story of Success — <em>Malcolm Gladwell</em></p><blockquote>Gladwell highlights how opportunities, external circumstances, cultural background and sheer practice (hard-work) contribute more than talent in success outcomes. <a href="https://www.amazon.com/Outliers-Story-Success-Malcolm-Gladwell/dp/0316017930/ref=sr_1_1?crid=2WVAYYVSVRGQQ&amp;keywords=outliers+book+paperback&amp;qid=1696565797&amp;sprefix=outliers+book+paper%252Caps%252C95&amp;sr=8-1&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=e51b96eed42ee6c1596a5a5c09db8811&amp;camp=1789&amp;creative=9325">Amazon link</a>.</blockquote><p><strong>21 Lessons for the 21st Century </strong>— <em>Yuval Noah Harari</em></p><blockquote>Enjoyed reading this book as much of Harari’s earlier book, <strong>Sapiens</strong>. I recommend both. While Sapiens was about the past, this is about the present. <a href="https://www.amazon.com/Lessons-21st-Century-Yuval-Harari/dp/0525512179/ref=tmm_hrd_swatch_0?_encoding=UTF8&amp;qid=1695694587&amp;sr=8-1&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=099677e0d8195f68f063bbae0a76917d&amp;camp=1789&amp;creative=9325">Amazon link</a>.</blockquote><h3>Personal Growth</h3><p><strong>Atomic Habits</strong>: An Easy &amp; Proven Way to Build Good Habits &amp; Break Bad Ones —<em> James Clear</em></p><blockquote>If you want to pick up a habit, start small, make it obvious (easy to do), attractive (appeals to you/human-bias) and rewarding (releases dopamine after doing it). Keep repeating and it becomes second nature. <a href="https://www.amazon.com/Atomic-Habits-Proven-Build-Break/dp/0735211299/ref=tmm_hrd_swatch_0?_encoding=UTF8&amp;qid=1695694667&amp;sr=8-1&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=ad37a5272d440f4afcfca1be5735d13f&amp;camp=1789&amp;creative=9325">Amazon link</a>.</blockquote><p><strong>The Startup of You:</strong> Adapt to the Future, Invest in Yourself, and Transform Your Career — <em>Reid Hoffman , Ben Casnocha</em></p><blockquote>This book is about treating your own career as a startup and how to enhance it. <a href="https://www.amazon.com/Start-up-You-Future-Yourself-Transform/dp/0307888908/ref=tmm_hrd_swatch_0?_encoding=UTF8&amp;qid=1695694171&amp;sr=8-1&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=df84bae53afbd58380fd3de7eee0e55f&amp;camp=1789&amp;creative=9325&amp;_encoding=UTF8&amp;tag=ajayshrestha-20&amp;linkCode=ur2&amp;linkId=cadc8cea43d438b2453bed316ee6c5b9&amp;camp=1789&amp;creative=9325">Amazon link</a>.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c315b9a32260" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Hard Problem of Consciousness]]></title>
            <link>https://ajay-shrestha.medium.com/hard-problem-of-consciousness-9a6a6bea6ea2?source=rss-658421888120------2</link>
            <guid isPermaLink="false">https://medium.com/p/9a6a6bea6ea2</guid>
            <category><![CDATA[quantum-physics]]></category>
            <category><![CDATA[mind]]></category>
            <category><![CDATA[consciousness]]></category>
            <category><![CDATA[philosophy]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Ajay Shrestha]]></dc:creator>
            <pubDate>Thu, 08 Jun 2023 00:25:21 GMT</pubDate>
            <atom:updated>2025-03-24T01:55:05.820Z</atom:updated>
            <content:encoded><![CDATA[<h4>The elephant in the room in scientific inquiry</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*TwIAkCwqQbu7yi_W" /><figcaption>Photo by <a href="https://unsplash.com/@davidclode?utm_source=medium&amp;utm_medium=referral">David Clode</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Prompt <strong>ChatGPT</strong> or <strong>Google Bard</strong> “<em>hardest unresolved problems in science</em>” and you will see <strong>consciousness</strong> listed in the top 5.</p><p><strong>Bard</strong><em> </em>listed it at #3 as following for me, when writing this post<em>:</em></p><blockquote>“The nature of consciousness: What is consciousness? How does it arise? These are questions that have been debated by philosophers and scientists for centuries, and there is still no consensus.”</blockquote><p>There is scientific consensus on the neural correlates of consciousness, but there isn’t one on how (or if) the brain causes consciousness.</p><p>Founder of quantum mechanics and Nobel Prize winner (Physics) <a href="https://en.wikipedia.org/wiki/Max_Planck">Max Planck</a> stated [<a href="https://bigthink.com/words-of-wisdom/max-planck-i-regard-consciousness-as-fundamental/">1</a>]:</p><blockquote>“I regard consciousness as fundamental. I regard matter as derivative from consciousness. We cannot get behind consciousness. Everything that we talk about, everything that we regard as existing, postulates consciousness.”</blockquote><p>Other than being so fundamental to our experience, one reason it is tricky to pin down consciousness is because it intersects many fields, including physics, biology, neuroscience, psychology, philosophy, computer science and others.</p><p>This is a very complex topic, and it would be impossible to cover it with sufficient depth in a medium-sized article. In addition, I don’t claim to be an expert in this subject. This post is a humble attempt to touch on the high-level generalized perspectives on consciousness, and trigger curiosity in the reader’s mind (or consciousness).</p><h3>What is Consciousness?</h3><p>First, let’s define consciousness for the context of this post. In his seminal paper, “<em>What Is It Like to Be a Bat</em>”, American philosopher <a href="https://en.wikipedia.org/wiki/Thomas_Nagel">Thomas Nagel</a><em> </em>stated [<a href="https://en.m.wikipedia.org/wiki/What_Is_It_Like_to_Be_a_Bat%3F">2</a>]<em>:</em></p><blockquote>“…consciousness has essential to it a subjective character, a what it is like aspect. He writes, “an organism has conscious mental states if and only if there is something that it is like to be that organism — something it is like for the organism.”….”</blockquote><p>Consciousness is an internal and subjective first-person experience that has a qualitative characteristic. According to Nagel, any attempt to reduce it to objective/physical/materialistic characteristics would be leaving out the very definition of consciousness.</p><h3>What is hard about Consciousness?</h3><p>The term, <em>Hard Problem of Consciousness</em> was coined by Philosopher <a href="https://en.wikipedia.org/wiki/David_Chalmers">David Chalmers</a>. It captures the <em>how and why</em> parts of our subjective conscious experience. He distinguishes it from the <em>easy problem</em> of consciousness, which can explain the physical systems that enable humans/animals to process information. According to him [<a href="https://en.wikipedia.org/wiki/Hard_problem_of_consciousness">3</a>]:</p><blockquote>“…even if we have solved all easy problems about the brain and experience, the hard problem will still persist”</blockquote><h3>Quantum Mechanics &amp; Consciousness</h3><p>Ever since I took a course in Quantum Computing in graduate school, I have been intrigued by the relationship between properties of matter at the quantum level and consciousness, and its implications for the overall nature of reality. I do want to clarify the course was mostly math and didn’t dwell on the ontology of reality or philosophy. But it’s not hard to see the connection.</p><p>A notable interpretation (<a href="https://en.wikipedia.org/wiki/Von_Neumann%E2%80%93Wigner_interpretation">Neumann-Wigner interpretation</a>) of the famous <em>double-slit</em> quantum experiment [<a href="https://towardsdatascience.com/quantum-mechanics-and-its-implications-for-reality-e88e33a53c09">4</a>–<a href="https://towardsdatascience.com/what-is-a-quantum-computer-f1c44e87a379">5</a>] confirms that a conscious observer or measurement that is later observed (by a conscious observer) collapses a superposition (possibilities) into a particular state of the sub-atomic particle. In other words, a particle is rendered in a particular way by the act of observation.</p><p>I cannot overstate the implications of this for science at the sub-atomic level, as science relies on objective measurement. But if consciousness and/or the act of measurement [<a href="https://en.m.wikipedia.org/wiki/Measurement_problem">6</a>] in quantum mechanics itself influences the outcome of the measurement, then we have a major challenge.</p><p>If you are interested in learning more about the relationship between quantum mechanics and consciousness, check out the following posts on quantum mechanics linked below.</p><ul><li><a href="https://medium.com/towards-data-science/quantum-mechanics-and-its-implications-for-reality-e88e33a53c09">Quantum Mechanics and its Implications for Reality</a></li><li><a href="https://medium.com/towards-data-science/what-is-a-quantum-computer-f1c44e87a379">What is a Quantum Computer?</a></li></ul><h3>Artificial Intelligence &amp; Consciousness</h3><p>Historian and author of the best-seller book Sapiens, <a href="https://medium.com/u/8a6d8f1fc05c">Yuval Noah Harari</a> says people (including scientists) often conflate intelligence with consciousness. Consciousness is the ability to suffer, be happy, experience color, taste chocolate, etc. Intelligence, on the other hand, is the ability to solve problems without the subjective element.</p><p><a href="https://medium.com/u/8a6d8f1fc05c">Yuval Noah Harari</a> and AI pioneer and Turing Award laureate <em>Yann LeCun</em> recently debated on the impact of AI [<a href="https://amp.lepoint.fr/2519782">7</a>]. Harari stated that humans possess consciousness and intelligence and use both to solve problems, whereas AI only has intelligence. On the contrary, LeCun has stated that sentience (consciousness) is an emergent property of computational devices [8], and it is a matter of time before AI systems develop consciousness [<a href="https://amp.lepoint.fr/2519782">7</a>]. Based on his book, <em>21 lessons for the 21st Century</em>, Harari is skeptical but open to the idea of consciousness arising from information/matter, whereas LeCun is more certain.</p><p>If you are interested in learning more about problems solved by AI, checkout the post linked below.</p><p><a href="https://ajay-shrestha.medium.com/machine-learning-vs-software-development-4b23cc182742">Machine Learning Vs Software Development</a></p><h3>Dualism</h3><p>The mind-body dualism is the attempt to solve the relationship between mind (consciousness) and matter (body/brain). It states that mind (subjective experience) and matter (objective properties) are distinct and separate from each other. Essentially, one cannot be reduced and explained in terms of the other. French philosopher and scientist <em>René Descartes</em> supported dualism and linked mind with consciousness and distinguished it from the brain [<a href="https://en.wikipedia.org/wiki/Mind%E2%80%93body_dualism">9</a>].</p><h3>Non-Dualism</h3><p>Non-duality considers consciousness to be fundamental and objective attributes to be derived from consciousness. Traditions in the East have a history in non-duality [<a href="https://en.wikipedia.org/wiki/Nondualism">10</a>]. It has some overlaps with concepts in Idealism [<a href="https://en.wikipedia.org/wiki/Idealism">11</a>] in the West.</p><p>American cognitive psychologist <a href="https://en.wikipedia.org/wiki/Donald_D._Hoffman">Donald Hoffman</a> supports the <em>consciousness before matter</em> view [<a href="https://evolutionnews.org/2023/01/brain-scientist-consciousness-didnt-evolve-it-creates-evolution/">12</a>]:</p><blockquote>“Consciousness didn’t emerge from a prior physical process of evolution. Consciousness is fundamental and so we have to rethink the whole history of the universe actually from this point of view, from The Big Bang up through evolution.”</blockquote><h3>Summary</h3><p>The jury is still out on how/if the brain (physical matter) generates consciousness. To summarize the theories across the wide spectrum of disciplines from physics to philosophy, there are three camps of explanations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0Y_FhJeVT0mJX6aBB0_vGg.png" /><figcaption>Table 1: Summary of Philosophical views on Consciousness</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Y-6DOjSgSRCv4Y5ybCzC5A.png" /><figcaption>Figure 1: Simplified generalization of the three views on objectivity/subjectivity scale</figcaption></figure><h3>Resources</h3><ol><li><a href="https://bigthink.com/words-of-wisdom/max-planck-i-regard-consciousness-as-fundamental/">https://bigthink.com/words-of-wisdom/max-planck-i-regard-consciousness-as-fundamental/</a></li><li><a href="https://en.m.wikipedia.org/wiki/What_Is_It_Like_to_Be_a_Bat%3F">What Is It Like to Be a_Bat — Wikipedia</a></li><li><a href="https://en.wikipedia.org/wiki/Hard_problem_of_consciousness">https://en.wikipedia.org/wiki/Hard_problem_of_consciousness</a></li><li><a href="https://towardsdatascience.com/quantum-mechanics-and-its-implications-for-reality-e88e33a53c09">Quantum Mechanics and its Implications for Reality</a></li><li><a href="https://towardsdatascience.com/what-is-a-quantum-computer-f1c44e87a379">What is a Quantum Computer?</a></li><li><a href="https://en.m.wikipedia.org/wiki/Measurement_problem">https://en.m.wikipedia.org/wiki/Measurement_problem</a></li><li>Yuval Noah Harari (Sapiens) vs Yann LeCun (Meta) on artificial intelligence — <a href="https://amp.lepoint.fr/2519782">https://amp.lepoint.fr/2519782</a></li><li><a href="https://twitter.com/ylecun/status/1539455515822112769?lang=en">https://twitter.com/ylecun/status/1539455515822112769?lang=en</a></li><li><a href="https://en.wikipedia.org/wiki/Mind%E2%80%93body_dualism">https://en.wikipedia.org/wiki/Mind%E2%80%93body_dualism</a></li><li><a href="https://en.wikipedia.org/wiki/Nondualism">https://en.wikipedia.org/wiki/Nondualism</a></li><li><a href="https://en.wikipedia.org/wiki/Idealism">https://en.wikipedia.org/wiki/Idealism</a></li><li><a href="https://evolutionnews.org/2023/01/brain-scientist-consciousness-didnt-evolve-it-creates-evolution/">https://evolutionnews.org/2023/01/brain-scientist-consciousness-didnt-evolve-it-creates-evolution/</a></li><li><a href="https://en.wikipedia.org/wiki/Physicalism">https://en.wikipedia.org/wiki/Physicalism</a></li><li><a href="https://en.wikipedia.org/wiki/Neural_correlates_of_consciousness">https://en.wikipedia.org/wiki/Neural_correlates_of_consciousness</a></li><li><a href="https://en.wikipedia.org/wiki/Max_Planck">https://en.wikipedia.org/wiki/Max_Planck</a></li><li><a href="https://en.wikipedia.org/wiki/Thomas_Nagel">https://en.wikipedia.org/wiki/Thomas_Nagel</a></li><li><a href="https://en.wikipedia.org/wiki/David_Chalmers">https://en.wikipedia.org/wiki/David_Chalmers</a></li><li><a href="https://en.wikipedia.org/wiki/Donald_D._Hoffman">https://en.wikipedia.org/wiki/Donald_D._Hoffman</a></li><li><a href="https://en.wikipedia.org/wiki/Von_Neumann%E2%80%93Wigner_interpretation">https://en.wikipedia.org/wiki/Von_Neumann%E2%80%93Wigner_interpretation</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9a6a6bea6ea2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Tips for Scholarly Research Publication]]></title>
            <link>https://medium.com/data-science/tips-for-scholarly-research-publication-bf96b1e3ad51?source=rss-658421888120------2</link>
            <guid isPermaLink="false">https://medium.com/p/bf96b1e3ad51</guid>
            <category><![CDATA[research]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[publishing]]></category>
            <category><![CDATA[academic-publishing]]></category>
            <dc:creator><![CDATA[Ajay Shrestha]]></dc:creator>
            <pubDate>Mon, 10 Apr 2023 18:46:02 GMT</pubDate>
            <atom:updated>2023-04-22T23:42:49.072Z</atom:updated>
            <content:encoded><![CDATA[<h4>If you are a graduate student trying to publish, check this out!</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*nQf49RMngLHQTh2V" /><figcaption>Photo by <a href="https://unsplash.com/@lazizli?utm_source=medium&amp;utm_medium=referral">Lala Azizli</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>The goal of this post is to share some guidelines with graduate students on publishing in journals and conferences. It is based on my past experience as a PhD student. One of my AI (deep learning) papers [<a href="https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&amp;arnumber=8694781">1</a>] recently crossed 1000 citations per Google Scholar [<a href="https://scholar.google.com/citations?user=wtBCz0QAAAAJ">2</a>]. According to Web of Science™ database [citation impact — <a href="https://en.m.wikipedia.org/wiki/Citation_impact">3</a>], only ~ 0.026% papers have over 1k citations. While it is great to see the post-publication impact the paper has had, the point up to the paper getting accepted was a lot of trial and error. This paper was rejected multiple times, took a couple of years, and underwent multiple iterations before it was accepted. I have tried to distill my learnings here into a guided process. Hopefully, the guidelines below will help you in your publishing journey.</p><h3>Problem</h3><p>Publishing is hard, especially if your school requires submission to high impact journals/conferences. Most PhD programs have publishing requirements, and it can even be the reason for holding up your graduation.</p><h4>Goals</h4><ol><li>Reduce time to publication</li><li>Amplify impact of paper</li></ol><h3>Guidelines</h3><p>Here are some of my learnings that you might find useful to reduce stress that comes with publishing. To draw an analogy, I think starting a startup company and the PhD publication process have a few things in common. Both have an uncertain future initially, but both could result in a novel contribution in their respective areas in the end. Just like a startup, your publication process needs a survey, vision, strategy, iterations, and scaling. My thoughts are organized below in the same sequence.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HOYxwmNeqDV85pCBRkI9qQ.png" /><figcaption>Figure 1: 5 milestones for publishing your paper | Image by author</figcaption></figure><h4>Survey</h4><p>Read up on current state-of-the-art in the area you are trying to publish. Remember you are trying to add to what is already out there. Many publications are open-access, so finding papers on the latest research should not be an issue. Plus, your university should also be able to provide you with necessary access. Google Scholar and ResearchGate are great open sources. Given that technology and scientific research are advancing so fast, I would also follow top researchers and companies in your area of research in LinkedIn, Twitter, and similar sites for the latest updates. If applicable, I would also explore statistics and data repositories for exploratory analysis. On the pace of the survey, recommended numbers vary from reading 1 to 7 papers per week in preparation for literature review and having 30 to 200 papers included/referenced in your survey.</p><h4>Vision</h4><p>After the survey, develop a vision of what topic you want to publish in. A startup starts with a problem it intends to solve. The founders develop a vision that aligns with them and the problem they intend to solve. To help find a match for yourself, start with the following questions:</p><ol><li>What are some of the unresolved problems in your area of study/major? E.g., <a href="https://en.wikipedia.org/wiki/Lists_of_unsolved_problems">this</a> [4] is a good reference list for unresolved problems in various disciplines.</li><li>What were some of the active research areas discussed in the survey papers?</li><li>Which course[s] excited you the most during your pre-qualifying exam coursework?</li><li>What are the areas of expertise of your advisor?</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TYMYM4_fb3TkZLn-Z0i7CA.png" /><figcaption>Figure 2: Venn Diagram to help formulate vision for publication | Image by author</figcaption></figure><p>Try to pick a topic from a region that has at least three overlaps in the Venn diagram above. Write an abstract and get feedback from your advisor. Ensure it includes your primary goal and proposed outline of the paper. The more critical the feedback, the more adjustments that will be required at this stage.</p><h4>Strategy</h4><p>During the initial stages of a startup company, its primary goal is to experiment, maximize learning, and land on building something novel the market actually needs, i.e., finding the elusive product-market fit. In your case, that would be landing on a publication, i.e., paper-publication fit. Both require strategy.</p><p>Before you come up with a strategy, consider the following:</p><ol><li>How much time and energy can you commit to publication effort? How are you going to balance other commitments?</li><li>How much time does your advisor have to guide you? Try to see if you can set up a recurring touchbase with your advisor.</li><li>What is your target time to get necessary publications to graduate? Ensure this is realistic.</li><li>How are you managing stress? Note that ~ 50% PhD candidates in North America drop out before they get their degree [<a href="https://www.wiley.com/en-us/network/publishing/research-publishing/trending-stories/improving-phd-completion-rates-where-should-we-start">5</a>]. You will need some type of life-hack to get through PhD. Unsurprisingly, two-thirds of startups fail as well, i.e., have negative return on investment [<a href="https://hbr.org/2021/05/why-start-ups-fail">6</a>].</li></ol><p>Once your figure these questions out and do some retrospective, come up with a publication strategy that includes:</p><ol><li>Paper topic/problem</li><li>Paper type, outline and target length</li><li>Target list of publications to submit to:<br>- Rank by acceptance-difficulty/impact-factor<br>- Include their turn around (review cycle) times<br>- Ensure the publication meets all criteria set by your institution</li><li>Anticipated target date to achieve publication success<br>- account for <em>n </em>iterations and time to revise paper <em>n</em> times</li></ol><h4>Iteration and Pivot</h4><p>Once you and your advisor are aligned on the strategy/focus, start your research and try to get to a point where you have something to show for (aka an MVP or minimal viable product in <em>startup</em> terms). Once you have an MVP paper, submit to journals/conferences starting at the higher difficulty range but have known shorter review cycles. Shorter review cycles let you iterate and incorporate feedback quicker and pivot strategy sooner, if needed.</p><p>Don’t get disheartened by rejection and instead use the critical (or harsh) feedback to make bold changes to your paper and/or strategy. Ensure that you are evaluating, responding to, and incorporating the feedback/gaps provided by the reviewers. Rejection with feedback is a blessing and a vital part of the process. Think of it as training an AI model. The weights of the artificial neural networks are adjusted based on feedback from labeled/training data during the training process. As shown in figure 3 below, these adjustments are initially higher and gradually taper off when you get closer to paper-publication fit (or global minima in AI/ML).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EqOHwf-7NZ-beKM0Sr8WPw.png" /><figcaption>Figure 3: <strong>Stochastic gradient descent</strong>, an iterative machine learning training algorithm to reach global/local minima (denoted with + sign). <strong>Paper-publication Fit</strong> follows similar iterations | Image by author</figcaption></figure><p>Pick journals/conferences that align with your papers research and read the submission guidelines carefully. There is nothing worse than waiting for months for feedback, only to learn the paper you submitted can not be considered because it isn’t aligned with the journal/conference’s theme or you missed an important submission step. If applicable, it will help to research past papers from journal/conference you are submitting to and cite relevant work from those papers. Also, if your co-authors (e.g., advisor) have had successful papers in certain journals, try them as well.</p><h4>Scale</h4><p>You will gradually get the hang of the process that leads you closer to publication after a few iterations, as the rejection reasons get less critical. It could take a really long time to get there. Note that your first publication could most likely be the most difficult and take the longest time.</p><p>In machine learning, there is a technique called <em>transfer learning</em> [<a href="https://en.wikipedia.org/wiki/Transfer_learning">7</a>], where you can apply knowledge gained from solving one task to another related task, with decreasing amount of learning effort. Just like that, all the validated learning from your first publication will come in very handy in speeding up your 2nd, 3rd and subsequent publications.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MP-PDQ4xF-24rxOvxsm_eQ.png" /><figcaption>Figure 4: Graph showing effort vs publication (outcome) | Image by author</figcaption></figure><p>Faster feedback cycles and iterations are the key. Use that to your advantage to get to the necessary number of publications. Stick to what works for you, and continue to learn and tweak the process for more impactful publications. Good luck!</p><h3>Disclaimer</h3><p>These tips are based on my learnings and may not work for everyone.</p><h3>References</h3><ol><li>A. Shrestha and A. Mahmood, “<em>Review of Deep Learning Algorithms and Architectures</em>,” in IEEE Access, vol. 7, pp. 53040–53065, 2019, doi: 10.1109/ACCESS.2019.2912200. (pdf download <a href="https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&amp;arnumber=8694781">link</a>)</li><li><a href="https://scholar.google.com/citations?user=wtBCz0QAAAAJ">Google Scholar Profile</a> — A. Shrestha</li><li>Citation Impact: <a href="https://en.m.wikipedia.org/wiki/Citation_impact">https://en.m.wikipedia.org/wiki/Citation_impact</a></li><li>Unresolved Problems — <a href="https://en.wikipedia.org/wiki/Lists_of_unsolved_problems">https://en.wikipedia.org/wiki/Lists_of_unsolved_problems</a></li><li>Post on PhD completion Rages — <a href="https://www.wiley.com/en-us/network/publishing/research-publishing/trending-stories/improving-phd-completion-rates-where-should-we-start">https://www.wiley.com/en-us/network/publishing/research-publishing/trending-stories/improving-phd-completion-rates-where-should-we-start</a></li><li>Why startups fail: <a href="https://hbr.org/2021/05/why-start-ups-fail">https://hbr.org/2021/05/why-start-ups-fail</a></li><li>Transfer Learning in AI/ML — <a href="https://en.wikipedia.org/wiki/Transfer_learning">https://en.wikipedia.org/wiki/Transfer_learning</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bf96b1e3ad51" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/tips-for-scholarly-research-publication-bf96b1e3ad51">Tips for Scholarly Research Publication</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cloud Basics: Container Orchestration]]></title>
            <link>https://everestaijournal.com/cloud-basics-container-orchestration-45debdf241c7?source=rss-658421888120------2</link>
            <guid isPermaLink="false">https://medium.com/p/45debdf241c7</guid>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[microservices]]></category>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[container-orchestration]]></category>
            <category><![CDATA[kubernetes]]></category>
            <dc:creator><![CDATA[Ajay Shrestha]]></dc:creator>
            <pubDate>Wed, 05 Apr 2023 05:28:59 GMT</pubDate>
            <atom:updated>2024-04-04T23:41:34.319Z</atom:updated>
            <content:encoded><![CDATA[<h4>Tutorial on deploying an application on Kubernetes (AWS EKS)</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*VUv_03IYMIo4inJ_" /><figcaption>Photo by <a href="https://unsplash.com/@carrier_lost?utm_source=medium&amp;utm_medium=referral">Ian Taylor</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>In this tutorial, we will deploy microservices based containerized application (retail app) on AWS’s managed Kubernetes service (EKS) cluster and expose it to the external world using loadbalancers.</p><p>Container orchestration entails automating the provisioning, networking, deploying, autoscaling, and load-balancing of containerized based applications. Kubernetes is the industry standard platform for container orchestration. EKS is AWS’s managed Kubernetes service.</p><p>This tutorial sources contents from AWS’s sites and GitHub repositories included in the resources section of this blogpost. It includes the following steps:</p><p><strong>A. </strong>Setup Cloud9 IDE<br><strong>B. </strong>Provision EKS cluster using terraform<br><strong>C.</strong> Deploy first microservice<br><strong>D. </strong>Deploy all microservices in the application<br><strong>E.</strong> Expose application using load balancer<br><strong>F.</strong> Cleanup (destroy terraform resources)</p><h3>Prerequisites</h3><p>See the Prerequisites section of the following post for setting up the following:</p><ol><li>AWS Console setup</li><li>AWS CLI setup<br>- use us-west-2 (Oregon) region for this lab (<strong>! IMPORTANT STEP!</strong>)<br>- If you already have AWS CLI installed, update to latest (<a href="https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html">link</a>)</li><li>GitHub.com and local git environment setup</li></ol><p><a href="https://towardsdatascience.com/cloud-basics-interacting-with-aws-da179d3f5829">Cloud Basics: Interacting with AWS</a></p><h3>Additional Prerequisites</h3><ol><li>Configure AWS CLI to use us-west-2 (Oregon) as the lab is not tested in the default us-east-1region. (<strong>! IMPORTANT STEP!</strong>)</li></ol><pre>aws configure<br>Default region name [us-east-1]:us-west-2</pre><h3>Application Architecture &amp; Deployment</h3><p>We will be deploying a sample retail application from AWS in this lab. The application consists of six microservices (blue circles along with the linked persistent layers in diagram below).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qLVnXcJoh75VYsNUbYqDpw.png" /><figcaption>Figure 1: Application to be deployed — source <a href="https://www.eksworkshop.com/docs/introduction/getting-started/about">EKS Workshop</a></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RI6FIB03yhSB6pGHzJwDlA.png" /><figcaption>Figure 2: Kubernetes Architecture — source <a href="https://archive.eksworkshop.com/010_introduction/architecture/architecture_control_and_data_overview/">EKS Workshop</a></figcaption></figure><p>Please following the steps below for executing the lab.</p><h4>A. IDE Setup (Cloud9)</h4><ol><li>Clone sample retail web application from AWS’s GitHub account to your local computer.</li></ol><pre>git clone git@github.com:aws-samples/eks-workshop-v2.git<br>cd eks-workshop-v2<br>git checkout main<br><br></pre><p>2. Run the following cloudformation command. It will take about 5 minutes to complete setup.</p><pre>aws cloudformation deploy --stack-name eks-workshop-ide \<br>    --template-file ./lab/cfn/eks-workshop-ide-cfn.yaml \<br>    --parameter-overrides RepositoryRef=stable \<br>    --capabilities CAPABILITY_NAMED_IAM</pre><p>Outcome in CloudFormation Console</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cHpFKH0Wim92Gt2lsdSviA.png" /><figcaption>Figure 3a: CloudFormation Stack for Cloud9 IDE in AWS</figcaption></figure><p>3. Click on the Cloud8 URL from the Outputs tab of the <em>eks-workshop-ide</em> stack</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*m0mwDbtvGwUyKNktkzxI9g.png" /><figcaption>Figure 3b: CloudFormation Stack for Cloud9 IDE in AWS</figcaption></figure><blockquote>NOTE: Ensure you are logged into aws console using IAM user and not the root user of the aws account. The IDE URL is not accessible to ROOT user. Ensure the IAM user matches the user configured in the prerequisite section (which was subsequently used to launch the CloudFormation template)</blockquote><p>Alternatively, you can also get the URL with this command.</p><pre>aws cloudformation describe-stacks --stack-name eks-workshop-ide \<br>    --query &#39;Stacks[0].Outputs[?OutputKey==`Cloud9Url`].OutputValue&#39; --output text</pre><p>4. Open <strong>Cloud9</strong> Page in your browser:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*s2mewOB7NH46oZJtiMdgNw.png" /><figcaption>Figure 3c: Cloud9 IDE in AWS</figcaption></figure><p>5. Run following command in your <strong>Cloud9 terminal</strong> to ensure setup.</p><pre>aws sts get-caller-identity<br><br>#Expected output:<br>shrestha:~/environment $ aws sts get-caller-identity<br>{<br>    &quot;UserId&quot;: &quot;AROA***********3EA:i-0c*********d9d&quot;,<br>    &quot;Account&quot;: &quot;141550591434&quot;,<br>    &quot;Arn&quot;: &quot;arn:aws:sts::1***********4:assumed-role/eks-workshop-ide-EksWorkshopC9Role-E*******v/i-0c******9d&quot;<br>}<br>shrestha:~/environment $ <br></pre><blockquote>NOTE: Unless otherwise noted, all of subsequent CLI commands in this lab needs to be run in <strong>Cloud9 terminal</strong> and not in local terminal</blockquote><h4>B. Provision EKS cluster using terraform</h4><ol><li>Download the terraform contents and execute them.</li></ol><pre><br>mkdir -p ~/environment/terraform; cd ~/environment/terraform<br>curl --remote-name-all https://raw.githubusercontent.com/aws-samples/eks-workshop-v2/stable/cluster/terraform/{main.tf,variables.tf,providers.tf,vpc.tf,eks.tf}</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ep6FRzUvZi5Y9dzS0Xjp8Q.png" /><figcaption>Figure 4: eks<em>.tf</em> inside terraform directory / Sourcecode — AWS</figcaption></figure><p>2. Initialize and create resources in AWS using terraform .</p><pre><br><br>terraform init<br>EKS_CLUSTER_NAME=eks-workshop<br>terraform apply -var=&quot;cluster_name=$EKS_CLUSTER_NAME&quot; -auto-approve</pre><blockquote><em>terraform applycommand ouput should have:</em><br><strong>Apply complete! Resources: 44 added, 0 changed, 0 destroyed.</strong></blockquote><blockquote>Note: This could take 10 to 20 minutes to complete!</blockquote><p>Once the cluster is created, execute the following for the lab to use as the cluster</p><pre>aws configure <br>#set region = us-west-2<br><br>use-cluster $EKS_CLUSTER_NAME<br><br>#Expected output<br>shrestha:~/environment/terraform $ aws configure<br>AWS Access Key ID [None]: <br>AWS Secret Access Key [None]: <br>Default region name [None]: us-west-2<br>Default output format [None]: <br>shrestha:~/environment/terraform $ use-cluster eks-workshop<br>Granting C9_USER access to the cluster via the AWS Console shrestha<br>shrestha:~/environment/terraform $ </pre><p>3. Configure your kubeclt config file to connect to your newly created EKS cluster.<br>- This command creates/update ~/.kube/config file</p><pre>aws eks update-kubeconfig --region us-west-2 --name eks-workshop<br>kubectl config view<br></pre><blockquote>The 1st command should have output like this:</blockquote><pre>Added new context arn:aws:eks:us-west-2:*********:cluster/eks-workshop to /Users/ajayshrestha/.kube/config</pre><blockquote>The 2nd command will display the config</blockquote><p>4. View the namespaces inside the EKS cluster</p><pre><br>kubectl get namespaces<br><br>#Expected output<br>shrestha:~/environment/terraform $ kubectl get namespaces<br>NAME              STATUS   AGE<br>default           Active   12m<br>kube-node-lease   Active   12m<br>kube-public       Active   12m<br>kube-system       Active   12m<br>shrestha:~/environment/terraform $ </pre><p>5. View the EKS cluster in AWS Console. Ensure you are in us-west-2 region. It could take up to 20 minutes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kmfrvw1IlzTkO7BVPXitxQ.png" /><figcaption>Figure 5a: EKS Cluster getting created in AWS</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dWBXKVp8XcIXi2m08FCI2Q.png" /><figcaption>Figure 5b: EKS Cluster getting created in AWS</figcaption></figure><h4>C. Deploy first microservice (Catalog)</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iP4DTzHNG9dUZPMZo5r0dw.png" /><figcaption>Figure 6: Catalog component deployed in pods in Kubernetes Cluster. source <a href="https://www.eksworkshop.com/docs/introduction/getting-started/about">EKS Workshop</a></figcaption></figure><ol><li>Deploy the catalog component using the YAML manifest file in manifests/base-application/catalog inside the repository.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AwHMcbkGvywY19ILPcS--g.png" /><figcaption>Figure 7: deployment.yaml manifest for Catalog component / Sourcecode — AWS</figcaption></figure><p>2. Deploy catalog Component. Ensure you are in environment directory of your repo and run following in terminal</p><pre>shrestha:~/environment $ prepare-environment introduction/getting-started<br>Refreshing copy of workshop repository from GitHub...<br><br>shrestha:~/environment $ ls ~/environment/eks-workshop/base-application/catalog<br><br>#Expected output:<br>configMap.yaml  deployment.yaml  kustomization.yaml  namespace.yaml  secrets.yaml  serviceAccount.yaml  service-mysql.yaml  service.yaml  statefulset-mysql.yaml</pre><pre>shrestha:~/environment $ kubectl apply -k ~/environment/eks-workshop/base-application/catalog<br><br>#Expected Output:<br>namespace/catalog created<br>serviceaccount/catalog created<br>configmap/catalog created<br>secret/catalog-db created<br>service/catalog created<br>service/catalog-mysql created<br>deployment.apps/catalog created<br>statefulset.apps/catalog-mysql created<br>shrestha:~/environment $ <br> </pre><p>3. Check the deployment in the newly created namespace</p><pre>kubectl get namespaces -l app.kubernetes.io/created-by=eks-workshop<br><br>#This should ouput following:<br>NAME      STATUS   AGE<br>catalog   Active   102s</pre><p>4. Verify the pods running the catalog service. There should be two pods: . Then 2nd pod (MySQL) might take a few minutes to be available.</p><pre>kubectl wait --for=condition=Ready pods --all -n catalog --timeout=180s<br>kubectl get pod -n catalog<br><br><br>#Expected Output:<br>% kubectl wait --for=condition=Ready pods --all -n catalog --timeout=180s<br>pod/catalog-6ccfd94978-hcwkx condition met<br>pod/catalog-mysql-0 condition met<br>ajayshrestha@ajays-MacBook-Air.local /Users/ajayshrestha/eks-workshop-v2/manifests/base-application/catalog [main]<br>% kubectl get pod -n catalog<br>NAME                       READY   STATUS    RESTARTS        AGE<br>catalog-6ccfd94978-hcwkx   1/1     Running   2 (3m30s ago)   3m35s<br>catalog-mysql-0            1/1     Running   0               3m35s</pre><p>5. Scale the pods to count 3 . You should now see 4 pods in total.</p><pre>kubectl scale -n catalog --replicas 3 deployment/catalog<br>kubectl wait --for=condition=Ready pods --all -n catalog --timeout=180s<br>kubectl get pod -n catalog<br><br><br>#Expected output (you should now see 3 instances/pods of catalog):<br>% kubectl scale -n catalog --replicas 3 deployment/catalog<br>deployment.apps/catalog scaled<br>ajayshrestha@ajays-MacBook /Users/ajayshrestha/eks-workshop-v2/manifests/base-application/catalog [main]<br>% kubectl wait --for=condition=Ready pods --all -n catalog --timeout=180s<br>pod/catalog-6ccfd94978-ggc57 condition met<br>pod/catalog-6ccfd94978-hcwkx condition met<br>pod/catalog-6ccfd94978-rrj2w condition met<br>pod/catalog-mysql-0 condition met<br>ajayshrestha@ajays-MacBook /Users/ajayshrestha/eks-workshop-v2/manifests/base-application/catalog [main]<br>% kubectl get pod -n catalog<br>NAME                       READY   STATUS    RESTARTS        AGE<br>catalog-6ccfd94978-ggc57   1/1     Running   0               19s<br>catalog-6ccfd94978-hcwkx   1/1     Running   2 (4m40s ago)   4m45s<br>catalog-6ccfd94978-rrj2w   1/1     Running   0               19s<br>catalog-mysql-0            1/1     Running   0               4m45s</pre><p>6. Check the cluster IP of the pods, created by service.yaml and service-mysql.yaml manifest files. The cluster IPs are not accessible outside VPC.</p><pre>kubectl get svc -n catalog<br><br>#Expected output<br>NAME            TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)    AGE<br>catalog         ClusterIP   172.20.144.70    &lt;none&gt;        80/TCP     5m57s<br>catalog-mysql   ClusterIP   172.20.202.161   &lt;none&gt;        3306/TCP   5m57s</pre><p>7. Check the catalog API is working by connecting inside the pods. A JSON payload should return.</p><pre>kubectl -n catalog exec -it deployment/catalog -- curl catalog.catalog.svc/catalogue | jq .<br><br><br>#Expected Output<br>{<br>    &quot;id&quot;: &quot;###############################&quot;,<br>    &quot;name&quot;: &quot;Gentleman&quot;,<br>    &quot;description&quot;: &quot;Touch of class for a bargain.&quot;,<br>    &quot;imageUrl&quot;: &quot;/assets/gentleman.jpg&quot;,<br>    &quot;price&quot;: 795,<br>    &quot;count&quot;: 51,<br>    &quot;tag&quot;: [<br>      &quot;dress&quot;<br>    ]<br>  },<br>  {<br>    &quot;id&quot;: &quot;###############################&quot;,<br>    &quot;name&quot;: &quot;Pocket Watch&quot;,<br>    &quot;description&quot;: &quot;Properly dapper.&quot;,<br>    &quot;imageUrl&quot;: &quot;/assets/pocket_watch.jpg&quot;,<br>    &quot;price&quot;: 385,<br>    &quot;count&quot;: 33,<br>    &quot;tag&quot;: [<br>      &quot;dress&quot;<br>    ]<br>  },<br>  {<br>    &quot;id&quot;: &quot;###############################&quot;,<br>    &quot;name&quot;: &quot;Chronograf Classic&quot;,<br>    &quot;description&quot;: &quot;Spend that IPO money&quot;,<br>    &quot;imageUrl&quot;: &quot;/assets/chrono_classic.jpg&quot;,<br>    &quot;price&quot;: 5100,<br>    &quot;count&quot;: 9,<br>    &quot;tag&quot;: [<br>      &quot;dress&quot;,<br>      &quot;luxury&quot;<br>    ]<br>  },<br>  {<br>    &quot;id&quot;: &quot;###############################&quot;,<br>    &quot;name&quot;: &quot;Wood Watch&quot;,<br>    &quot;description&quot;: &quot;Looks like a tree&quot;,<br>    &quot;imageUrl&quot;: &quot;/assets/wood_watch.jpg&quot;,<br>    &quot;price&quot;: 50,<br>    &quot;count&quot;: 115,<br>    &quot;tag&quot;: [<br>      &quot;casual&quot;<br>    ]<br>  },<br>  {<br>    &quot;id&quot;: &quot;###############################&quot;,<br>    &quot;name&quot;: &quot;Smart 3.0&quot;,<br>    &quot;description&quot;: &quot;Can tell you what you want for breakfast&quot;,<br>    &quot;imageUrl&quot;: &quot;/assets/smart_1.jpg&quot;,<br>    &quot;price&quot;: 650,<br>    &quot;count&quot;: 9,<br>    &quot;tag&quot;: [<br>      &quot;smart&quot;,<br>      &quot;dress&quot;<br>    ]<br>  },<br>  {<br>    &quot;id&quot;: &quot;###############################&quot;,<br>    &quot;name&quot;: &quot;FitnessX&quot;,<br>    &quot;description&quot;: &quot;Touch of class for a bargain.&quot;,<br>    &quot;imageUrl&quot;: &quot;/assets/smart_2.jpg&quot;,<br>    &quot;price&quot;: 180,<br>    &quot;count&quot;: 76,<br>    &quot;tag&quot;: [<br>      &quot;dress&quot;,<br>      &quot;smart&quot;<br>    ]<br>  }</pre><h4>D. Deploy all microservice</h4><p>Now lets deploy rest of the components of the application.</p><ol><li>Go to manifests/base-application/ in your IDE to view the components. Open kustomization.yaml.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gn1wsE-j2JmVFqwMnheMNQ.png" /><figcaption>Figure 8: kustomization.yaml manifest for base application/ Sourcecode — AWS</figcaption></figure><p>2. Go to your local repo folder and create rest of the components.</p><pre><br>kubectl apply -k ~/environment/eks-workshop/base-application<br><br><br>#Expected Output<br>shrestha:~/environment $ kubectl apply -k ~/environment/eks-workshop/base-application<br><br>namespace/assets created<br>namespace/carts created<br>namespace/catalog unchanged<br>namespace/checkout created<br>namespace/orders created<br>namespace/other created<br>namespace/rabbitmq created<br>namespace/ui created<br>serviceaccount/assets created<br>serviceaccount/carts created<br>serviceaccount/catalog unchanged<br>serviceaccount/checkout created<br>serviceaccount/orders created<br>serviceaccount/rabbitmq created<br>serviceaccount/ui created<br>role.rbac.authorization.k8s.io/rabbitmq-endpoint-reader created<br>rolebinding.rbac.authorization.k8s.io/rabbitmq-endpoint-reader created<br>configmap/assets created<br>configmap/carts created<br>configmap/catalog unchanged<br>configmap/checkout created<br>configmap/orders created<br>configmap/dummy created<br>configmap/ui created<br>secret/catalog-db unchanged<br>secret/orders-db created<br>secret/rabbitmq created<br>secret/rabbitmq-config created<br>service/assets created<br>service/carts created<br>service/carts-dynamodb created<br>service/catalog unchanged<br>service/catalog-mysql unchanged<br>service/checkout created<br>service/checkout-redis created<br>service/orders created<br>service/orders-mysql created<br>service/rabbitmq created<br>service/rabbitmq-headless created<br>service/ui created<br>deployment.apps/assets created<br>deployment.apps/carts created<br>deployment.apps/carts-dynamodb created<br>deployment.apps/catalog configured<br>deployment.apps/checkout created<br>deployment.apps/checkout-redis created<br>deployment.apps/orders created<br>deployment.apps/orders-mysql created<br>deployment.apps/ui created<br>statefulset.apps/catalog-mysql unchanged<br>statefulset.apps/rabbitmq created<br>shrestha:~/environment $ </pre><p>3. Wait for all components to complete</p><pre>kubectl wait --for=condition=Ready --timeout=180s pods -l app.kubernetes.io/created-by=eks-workshop -A</pre><p>4. See the namespaces, deployments and services for each component. Note that the cluster IPs are not accessible outside VPC.</p><pre>kubectl get namespaces -l app.kubernetes.io/created-by=eks-workshop<br>kubectl get deployment -l app.kubernetes.io/created-by=eks-workshop -A<br>kubectl get svc -l app.kubernetes.io/created-by=eks-workshop -A</pre><p>Output of the four commands above:</p><pre>shrestha:~/environment $ kubectl get namespaces -l app.kubernetes.io/created-by=eks-workshop<br>NAME       STATUS   AGE<br>assets     Active   53s<br>carts      Active   53s<br>catalog    Active   5m17s<br>checkout   Active   53s<br>orders     Active   53s<br>other      Active   53s<br>rabbitmq   Active   53s<br>ui         Active   53s<br>shrestha:~/environment $ kubectl get deployment -l app.kubernetes.io/created-by=eks-workshop -A<br>NAMESPACE   NAME             READY   UP-TO-DATE   AVAILABLE   AGE<br>assets      assets           1/1     1            1           58s<br>carts       carts            1/1     1            1           58s<br>carts       carts-dynamodb   1/1     1            1           58s<br>catalog     catalog          1/1     1            1           5m23s<br>checkout    checkout         1/1     1            1           58s<br>checkout    checkout-redis   1/1     1            1           58s<br>orders      orders           1/1     1            1           57s<br>orders      orders-mysql     1/1     1            1           57s<br>ui          ui               1/1     1            1           57s<br>shrestha:~/environment $ kubectl get svc -l app.kubernetes.io/created-by=eks-workshop -A<br>NAMESPACE   NAME             TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)                                 AGE<br>assets      assets           ClusterIP   172.20.232.38    &lt;none&gt;        80/TCP                                  63s<br>carts       carts            ClusterIP   172.20.61.199    &lt;none&gt;        80/TCP                                  63s<br>carts       carts-dynamodb   ClusterIP   172.20.124.105   &lt;none&gt;        8000/TCP                                63s<br>catalog     catalog          ClusterIP   172.20.200.90    &lt;none&gt;        80/TCP                                  5m28s<br>catalog     catalog-mysql    ClusterIP   172.20.229.241   &lt;none&gt;        3306/TCP                                5m28s<br>checkout    checkout         ClusterIP   172.20.220.231   &lt;none&gt;        80/TCP                                  63s<br>checkout    checkout-redis   ClusterIP   172.20.25.218    &lt;none&gt;        6379/TCP                                63s<br>orders      orders           ClusterIP   172.20.200.220   &lt;none&gt;        80/TCP                                  63s<br>orders      orders-mysql     ClusterIP   172.20.220.30    &lt;none&gt;        3306/TCP                                63s<br>rabbitmq    rabbitmq         ClusterIP   172.20.254.241   &lt;none&gt;        5672/TCP,4369/TCP,25672/TCP,15672/TCP   63s<br>ui          ui               ClusterIP   172.20.91.187    &lt;none&gt;        80/TCP                                  63s</pre><h4>E. Expose application using load balancer</h4><p>Now lets expose the service externally.</p><ol><li>Run following command in your <strong>Cloud9 terminal</strong></li></ol><pre>use-cluster eks-workshop<br>prepare-environment exposing/load-balancer<br><br>#Expected Output<br>shrestha:~/environment $ use-cluster eks-workshop<br>Granting C9_USER access to the cluster via the AWS Console shrestha<br>Updated context arn:aws:eks:us-west-2:14*******4:cluster/eks-workshop in /home/ec2-user/.kube/config<br>shrestha:~/environment $ prepare-environment exposing/load-balancer<br>Refreshing copy of workshop repository from GitHub...<br><br>Resetting the environment, please wait<br>Waiting for application to become ready...<br>Cleaning up previous lab infrastructure...<br>Creating infrastructure for next lab...<br>Environment is ready<br>shrestha:~/environment $  </pre><p>This installs the AWS Load Balancer Controller in our Amazon EKS cluster:</p><p>You can view the Terraform that applies these changes <a href="https://github.com/aws-samples/eks-workshop-v2/tree/stable/manifests/modules/exposing/load-balancer/.workshop/terraform">here</a></p><p>2. Go to ../modules/exposing/load-balancer/nlb/nlb.yamlin your IDE to view the components</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ShU4bEGrqxEaB-MVnnRkWQ.png" /><figcaption>Figure 9: nib.yaml / Sourcecode — AWS</figcaption></figure><p>3. Check microservices are only accessible internally. Note that none have “External-IP”.</p><pre>kubectl get svc -l app.kubernetes.io/created-by=eks-workshop -A<br><br>#Expected Output:<br>shrestha:~/environment $ kubectl get svc -l app.kubernetes.io/created-by=eks-workshop -A<br>NAMESPACE   NAME             TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)                                 AGE<br>assets      assets           ClusterIP   172.20.232.38    &lt;none&gt;        80/TCP                                  6m43s<br>carts       carts            ClusterIP   172.20.61.199    &lt;none&gt;        80/TCP                                  6m43s<br>carts       carts-dynamodb   ClusterIP   172.20.124.105   &lt;none&gt;        8000/TCP                                6m43s<br>catalog     catalog          ClusterIP   172.20.200.90    &lt;none&gt;        80/TCP                                  11m<br>catalog     catalog-mysql    ClusterIP   172.20.229.241   &lt;none&gt;        3306/TCP                                11m<br>checkout    checkout         ClusterIP   172.20.220.231   &lt;none&gt;        80/TCP                                  6m43s<br>checkout    checkout-redis   ClusterIP   172.20.25.218    &lt;none&gt;        6379/TCP                                6m43s<br>orders      orders           ClusterIP   172.20.200.220   &lt;none&gt;        80/TCP                                  6m43s<br>orders      orders-mysql     ClusterIP   172.20.220.30    &lt;none&gt;        3306/TCP                                6m43s<br>rabbitmq    rabbitmq         ClusterIP   172.20.254.241   &lt;none&gt;        5672/TCP,4369/TCP,25672/TCP,15672/TCP   6m43s<br>ui          ui               ClusterIP   172.20.91.187    &lt;none&gt;        80/TCP                                  6m43s<br>shrestha:~/environment $ </pre><p>4. Create Loadbalancer</p><pre><br>kubectl apply -k ~/environment/eks-workshop/modules/exposing/load-balancer/nlb<br><br>#Expected Output:<br>shrestha:~/environment $ kubectl apply -k ~/environment/eks-workshop/modules/exposing/load-balancer/nlb<br><br>service/ui-nlb created<br><br> </pre><p>Verify creation. ui was already there before. Now you should also see ui-nlb with an external endpoint/IP.</p><pre><br>kubectl get service -n ui<br><br>#Expected Output<br>NAME     TYPE           CLUSTER-IP      EXTERNAL-IP                                                            PORT(S)        AGE<br>ui       ClusterIP      172.20.91.187   &lt;none&gt;                                                                 80/TCP         8m48s<br>ui-nlb   LoadBalancer   172.20.70.108   k8s-ui-uinlb-b38f8921bb-54d54e341ee561bb.elb.us-west-2.amazonaws.com   80:30991/TCP   42s<br>shrestha:~/environment $ </pre><p>5. Get URL of the NLB. It might take a few minutes to provision NLB</p><pre><br>shrestha:~/environment $ kubectl get service -n ui ui-nlb -o jsonpath=&quot;{.status.loadBalancer.ingress[*].hostname}{&#39;\n&#39;}&quot;<br><br>#Expected Output:<br>k8s-ui-uinlb-b38f8921bb-54d54e341ee561bb.elb.us-west-2.amazonaws.com<br>shrestha:~/environment $ <br>shrestha:~/environment $ <br></pre><pre>kubectl get deployment -l app.kubernetes.io/created-by=eks-workshop -A<br><br>#Expected Output<br>shrestha:~/environment $ kubectl get deployment -l app.kubernetes.io/created-by=eks-workshop -A<br>NAMESPACE   NAME             READY   UP-TO-DATE   AVAILABLE   AGE<br>assets      assets           1/1     1            1           11m<br>carts       carts            1/1     1            1           11m<br>carts       carts-dynamodb   1/1     1            1           11m<br>catalog     catalog          1/1     1            1           15m<br>checkout    checkout         1/1     1            1           11m<br>checkout    checkout-redis   1/1     1            1           11m<br>orders      orders           1/1     1            1           11m<br>orders      orders-mysql     1/1     1            1           11m<br>ui          ui               1/1     1            1           11m</pre><p>6. Checkout the URL in your Browser</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_85reenULCa1OwB-hyxmhg.png" /><figcaption>Figure 10: UI for the Base Application from AWS</figcaption></figure><h4>F. Cleanup (destroy terraform resources)</h4><ol><li>Navigate to the directory where you cloned the repository and run the following commands in Cloud9 terminal.</li></ol><pre><br>delete-environment<br><br>cd ~/environment/terraform<br>terraform destroy -var=&quot;cluster_name=$EKS_CLUSTER_NAME&quot; -auto-approve<br><br><br><br>#Expected Output<br>shrestha:~/environment $ delete-environment<br>Refreshing copy of workshop repository from GitHub...<br><br>Resetting the environment, please wait<br>Waiting for application to become ready...<br>Cleaning up previous lab infrastructure...<br>Environment is ready<br>shrestha:~/environment $ cd ~/environment/terraform<br>shrestha:~/environment/terraform $ terraform destroy -var=&quot;cluster_name=$EKS_CLUSTER_NAME&quot; -auto-approve<br>.....<br>.....<br>Destroy complete! Resources: 44 destroyed.</pre><blockquote>This could take 20 minutes or more and should destroy/delete all resources created earlier. Ensure the cluster eks-workshop and related Loadbalancers/EC2s are not visible in the AWS Console (correct region) as well. Check the CloudFormation stack as well. Also, ensure your AWS billing alerts are configured at desired threshold.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VstrN2GudiIb3QARLP4eew.png" /><figcaption>Figure 9: AWS Console showing no EKS cluster</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/894/1*SWnkdOLJWC6A2FbSOG34Tw.png" /><figcaption>Lab related ec2 instances should also be terminated in the correct region</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*r6CLabWezV8wEFHQgiYRmQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IntD_p3vsKTu1ntkh6U-dA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PdkMLLLUB7jd-BnjK_ilmg.png" /></figure><p>2. Feel free to revert back your default AWS region to us-east-1 in you <strong>local terminal</strong>.</p><pre>aws configure<br>Default region name [us-east-2]:us-east-1</pre><h3>Revision History</h3><ol><li>Published April 6, 2023 <br>- Original</li><li>Updated November 26, 2023<br>- Updated “region” and misc steps to adapt to new version of code</li><li>Updated November 28, 2023<br>- Updated “Cloud9” steps.</li></ol><h3>Resources</h3><ol><li>Sample Microservices-based Application — <a href="https://github.com/aws-containers/retail-store-sample-app">https://github.com/aws-containers/retail-store-sample-app</a></li><li>Current AWS EKS Workshop — <a href="https://www.eksworkshop.com/">https://www.eksworkshop.com/</a><br>- Related Git Repo: <a href="https://github.com/aws-samples/eks-workshop-v2">https://github.com/aws-samples/eks-workshop-v2</a></li><li>Archived AWS EKS Workshop — <a href="https://archive.eksworkshop.com/">https://archive.eksworkshop.com/</a><br>- Related Git Repo: <a href="https://github.com/aws-samples/eks-workshop">https://github.com/aws-samples/eks-workshop</a></li><li><a href="https://aws.amazon.com/getting-started/hands-on/break-monolith-app-microservices-ecs-docker-ec2/">https://aws.amazon.com/getting-started/hands-on/break-monolith-app-microservices-ecs-docker-ec2/</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=45debdf241c7" width="1" height="1" alt=""><hr><p><a href="https://everestaijournal.com/cloud-basics-container-orchestration-45debdf241c7">Cloud Basics: Container Orchestration</a> was originally published in <a href="https://everestaijournal.com/">Everest AI Journal</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cloud Basics: Pipeline as Code]]></title>
            <link>https://everestaijournal.com/cloud-basics-pipeline-as-code-eb60e92bb757?source=rss-658421888120------2</link>
            <guid isPermaLink="false">https://medium.com/p/eb60e92bb757</guid>
            <category><![CDATA[git]]></category>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[cloud-computing]]></category>
            <dc:creator><![CDATA[Ajay Shrestha]]></dc:creator>
            <pubDate>Tue, 21 Feb 2023 04:06:52 GMT</pubDate>
            <atom:updated>2023-10-19T01:07:50.500Z</atom:updated>
            <content:encoded><![CDATA[<h4>Launch application with deployment pipeline configured as code</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*cgPvrg7rrs47G_7W" /><figcaption>Photo by <a href="https://unsplash.com/@carlosaranda?utm_source=medium&amp;utm_medium=referral">carlos aranda</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>In this tutorial we will be launching WordPress application in AWS cloud using Deployment Pipeline configured as code. Pipeline as code (PaC) is similar to “infrastructure as code (IaC)”, where we launch and configure infrastructure as code. In PaC, we configure build, test and deployment steps in the form of code. The code is organized in stages, jobs and actions sections in a declarative manner. The automated build and deployment described in PaC is usually triggered by a code commit to a designated branch in central version control code repository. PaC enables pipeline configuration to follow standard software development best practices.</p><h4><strong>Benefits of PaC -</strong></h4><ol><li>Version Control and change tracking</li><li>Standardization / Reliability</li><li>Automation / Cost savings / Shorter Lead Time</li><li>Collaboration / Code Review / Learning</li></ol><h4><strong>Tutorial Outline</strong></h4><p>A. Create GitHub.com OAuth token</p><p>B. Create secrets in AWS SecretsManager</p><p>C. Clone GitHub repo with Pipeline &amp; Application Code and update</p><p>D. Launch Pipeline Code</p><p>E. View Application get deployed automatically</p><h4>Related Posts</h4><p>See related posts on infrastructure as code and db migrations as code —</p><ul><li><a href="https://ajay-shrestha.medium.com/cloud-basics-infrastructure-as-code-b1027a07e118">Cloud Basics: Infrastructure as Code</a></li><li><a href="https://ajay-shrestha.medium.com/cloud-basics-intro-to-db-migrations-3e84967a8123">Cloud Basics: Intro to DB Migrations</a></li></ul><h3>Prerequisites</h3><p>See the Prerequisites section of the following post for setting up the following:</p><ol><li>AWS Console setup</li><li>AWS CLI setup</li><li>GitHub.com and local git environment setup</li></ol><p><a href="https://towardsdatascience.com/cloud-basics-interacting-with-aws-da179d3f5829">Cloud Basics: Interacting with AWS</a></p><p><strong>A. GitHub.com OAuth Token</strong></p><p>Create GitHub.com OAuth token with access to your repositories</p><p>Follow step here: <a href="https://docs.github.com/en/authentication/keeping-your-account-and-data-secure/creating-a-personal-access-token">https://docs.github.com/en/authentication/keeping-your-account-and-data-secure/creating-a-personal-access-token</a></p><p><strong>B. Create secrets in AWS SecretsManager to store</strong></p><ul><li>DB password</li></ul><pre>aws secretsmanager create-secret --name wordpress-lab-db-password --description “DB Password for Wordpress Lab” --secret-string &quot;your password here&quot;<br><br>=========Example Output==========<br>{<br>    &quot;ARN&quot;: &quot;arn:aws:secretsmanager:us-east-1:134:secret:wordpress-lab-db&quot;,<br>    &quot;Name&quot;: &quot;wordpress-lab-db-password&quot;,<br>    &quot;VersionId&quot;: &quot;dfaed4&quot;<br>}</pre><ul><li>GitHub OAuth token</li></ul><pre>aws secretsmanager create-secret --name github-oauth-token --description &quot;Secret for GitHub&quot; --secret-string &quot;insert your GitHub OAuth token&quot;<br><br><br>=========Example Output==========<br>{<br>    &quot;ARN&quot;: &quot;arn:aws:secretsmanager:us-east-1:1934:secret:github-oauth-t&quot;,<br>    &quot;Name&quot;: &quot;github-oauth-token&quot;,<br>    &quot;VersionId&quot;: &quot;fdc414d&quot;<br>}</pre><p>You can see two secrets in AWS Console &gt; SecretsManager</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nLn8eKR8qkLmR3NGnKcp6g.png" /><figcaption>Figure 1: secrets stored in AWS Secrets Manager</figcaption></figure><p><strong>C. Clone Repo and make Changes</strong></p><ol><li>Clone GitHub repo that includes the Pipeline &amp; Application Code</li><li>Go to <a href="https://github.com/shrestha-ajay/cloud-basics4">https://github.com/shrestha-ajay/cloud-basics4</a></li><li>Fork this repository by clicking on “Fork” in the GitHub UI</li><li>Clone this repository to your local computer by running following:</li></ol><pre>$ git clone git@github.com:YOUR-USERNAME/YOUR-REPOSITORY.git<br>$ cd YOUR-REPOSITORY-Folder</pre><p>5. Update parameter values for following in you local branch (in your computer) using your IDE and save.</p><p>a. codepipeline.yaml</p><ul><li><em>GitHubOwner</em></li><li><em>RepositoryName</em></li></ul><p>b. wordpress-app.yaml</p><ul><li><em>KeyName</em></li></ul><p>6. Validate the CloudFormation files with following command</p><pre>% aws cloudformation validate-template --template-body file://wordpress-app.yaml<br>% aws cloudformation validate-template --template-body file://codepipeline.yaml</pre><p>7. Push changes to your GitHub.com account</p><pre>git add .<br>git commit -m &quot;updating parameter values&quot;<br>git push</pre><p><strong>D. Launch Pipeline Code (CloudFormation)</strong></p><pre>./create-codepipeline.sh<br><br>=========Example Output==========<br>{<br> &quot;StackId&quot;: &quot;arn:aws:cloudformation:us-east-1:14154:stack/github-wordpress-codepipeline/499dcfdf&quot;<br>}<br><br>========for Windows users=========<br>create-codepipeline.sh (command prompt)<br>sh create-codepipeline.sh (git bash)</pre><p>1. View pipeline creation events in CloudFormation Stack</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/468/1*tRRfb9jvVkhokl_2cANwKA.png" /><figcaption>Figure 2: CloudFormation Stack status in AWS CloudFormation Console</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bK7Reatv-Wp4S1Q6nu9ISA.png" /><figcaption>Figure 3: CloudFormation Stack events in AWS CloudFormation Console</figcaption></figure><p>2. The 2nd stage (<em>Deploy</em>) of pipeline will fail initially. Either push a new commit to main branch in GitHub or simply click on <em>retry</em> button next to <em>Deploy</em> stage in AWS CodePipeline to trigger the deploy.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/918/1*IDdq1N3gAyrYlMuQGnB2XA.png" /><figcaption>Figure 4: CodePipeline stages in AWS CodePipeline Console</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GSeFultvL_QPRkOtwlrL7g.png" /><figcaption>Figure 5: CodePipeline status in AWS CodePipeline Console</figcaption></figure><p><strong>E. View Application get deployed automatically</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cC8OK4DPbMkgut5RP1i_SQ.png" /><figcaption>Figure 6: Events of CloudFormation Stack triggered by CodePipeline. The Stack launches WordPress Application</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZfyV2ahrtONcZM991YrG1A.png" /><figcaption>Figure 7: WordPress Application EC2 (webserver) created by earlier CloudFormation stack</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8j6XdaxTCA5BE1UZ_1ItkQ.png" /><figcaption>Figure 8: Ouput (Web server URL) of CloudFormation Stack that launched the WordPress Application</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WrpduTEC_bUfb0ZkdUHHPA.png" /><figcaption>Figure 9: WordPress Setup Interface</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2j_6xwvS0DF6IK7M07nMPQ.png" /><figcaption>Figure 10: WordPress main page</figcaption></figure><h3>Clean Up</h3><p>Gentle reminder to terminate all resources (2 CloudFormation stacks) created during the tutorial to avoid being charged extra for them. You will need to empty the S3 bucket (created from the CloudFormation stack) to successfully delete the stack.</p><h3>Resources</h3><ol><li><a href="https://aws.amazon.com/codepipeline/">https://aws.amazon.com/codepipeline/</a></li><li><a href="https://about.gitlab.com/topics/ci-cd/pipeline-as-code/">https://about.gitlab.com/topics/ci-cd/pipeline-as-code/</a></li><li><a href="https://github.blog/2022-02-02-build-ci-cd-pipeline-github-actions-four-steps/">https://github.blog/2022-02-02-build-ci-cd-pipeline-github-actions-four-steps/</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eb60e92bb757" width="1" height="1" alt=""><hr><p><a href="https://everestaijournal.com/cloud-basics-pipeline-as-code-eb60e92bb757">Cloud Basics: Pipeline as Code</a> was originally published in <a href="https://everestaijournal.com/">Everest AI Journal</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>