Notes from 20/1–27/1

The Essential Clayton Christensen Articles

It’s paywalled by HBR. Will do a separate article with the recommended articles

The threat of disruptive innovation: the core theory of why bad things happen to good companies. “Disruptive Technologies: Catching the Wave” is the big-picture “why is this a problem” article warning established companies that a seemingly rational concern with profit margins can have disastrous results. It outlines several classic examples — primarily disk drives, along with Apple and Digital Equipment Corporation — to show that there is a pattern big companies should pay attention to.

Organizational structure: “Meeting the Challenge of Disruptive Change” describes how leaders can structure their organizations to allow the kinds of innovation that stave off disruption. Here Christensen runs Digital Equipment Corporation through his framework to show how it can be used to explain that company’s infamous reversal of fortune.

Product innovation: “Marketing Malpractice: The Cause and the Cure” again asks why good managers struggle to innovate successfully, this time focusing on the discipline of product innovation itself, rather than on organizational and management structures. By understanding the tasks that customers look to a product for (the “job to be done”), a company can develop offerings — products, services, and whole brands — that customers truly value. Christensen uses the “milk shake” example to show how product developers should be considering their task.

The financial tools in the way: Established financial incentives often make it unattractive for companies to innovate. In “Innovation Killers: How Financial Tools Destroy Your Capacity to Do New Things,” Christensen and his coauthors target metrics such as discounted cash flow, net present value, and earnings per share, along with attitudes towards fixed and sunk costs. They suggest that leaders take up other methods for evaluating investments — ones that consider future value.

Business model innovation: Product innovations might be necessary, but to be truly disruptive, they often need to be delivered to the market through new business models. In “Reinventing Your Business Model,” Christensen and his coauthors describe how to determine if your company needs a new business model and what makes one successful, using examples ranging from Apple’s iTunes to CVS’s MinuteClinics.

The role of business models in M&A: To reinvent their business models, companies sometimes decide to merge with or acquire another firm. But the failure rate of M&A is somewhere between 70% and 90%. “The New M&A Playbook” explains that the failures often stem from a lack of clarity about why a merger or acquisition is being pursued. Companies need to consider whether they are really after business model reinvention or are simply looking to bolster their current model. These purposes demand very different implementations of a deal — from paying the right price to determining how employees and other resources will be handled.

Where your industry’s future growth lies: If disruption is predictable, we should be able to step back and look at markets as a whole to understand how disruption will change an industry over time. “Skate to Where the Money Will Be” describes a pattern of evolution of markets and industries that can help managers see where their next source of profits will be — so that they don’t find themselves outpaced by another company in that new sphere.

The extendable core: How do you know how big a particular threat to your business actually is? “Surviving Disruption” helps you calculate the strengths of your own potential disrupter’s business model along with your own relative advantages and determine what conditions could keep your disrupter from triumphing. Christensen and his coauthor build on the jobs-to-be-done theory and introduce the “extendable core” — the part of a disrupter’s business model that enables it to keep undercutting you as it creeps upmarket into your territory.

Disruptive innovation, revisited: The ideas summed up in the phase “disruptive innovation” have become a powerful part of business thinking in the 20 years since they were introduced — but they’re in danger of losing their usefulness, because they’ve been misunderstood and misapplied. In “What is Disruptive Innovation?” Christensen and his coauthors revisit the essential concepts, show the importance of using the term precisely, and share what they have learned from two decades’ application of the idea in the field.

What makes good management theory: By testing a business theory with the scientific method — by conducting a reality check — we can learn whether the theory will really help us predict the future. “Why Hard-Nosed Executives Should Care About Management Theory,” argues for a more rigorous testing of theories so that managers can gain a better sense of whether an idea is relevant to their specific situation.

A personal strategy: Christensen extends his examination to the personal realm, arguing that bad things sometimes happen to good people because those people lack a strategy for their lives. In “How Will You Measure Your Life?” he uses concepts from business to challenge readers to manage their careers and personal lives in a way that leads to lasting satisfaction.

AI

XLNet outperforms BERT on several NLP Tasks

Well describes the difference between the 2 models and an overview of the individual papers.

All Machine Learning Models Explained in 6 Minutes

It’s really just the basic machine learning models. Not sure if you can say “all”. I feel cheated.

Microsoft Introduces Project Petridish to Find the Best Neural Network for your Problem

• PHASE 0: Petridish starts with some parent model, a very small human-written model with one or two layers or a model already found by domain experts on a dataset.
• PHASE 1: Petridish connects the candidate layers to the parent model using stop-gradient and stop-forward layers and partially train it. The candidate layers can be any bag of operations in the search space. Using stop-gradient and stop-forward layers allows gradients with respect to the candidates to be accumulated without affecting the model’s forward activations and backward gradients. Without the stop-gradient and stop-forward layers, it would be difficult to determine which candidate layers are contributing what to the parent model’s performance and would require separate training if you wanted to see their respective contributions, increasing costs.
• PHASE 2: If a particular candidate or set of candidates is found to be beneficial to the model, then we remove the stop-gradient and stop-forward layers and the other candidates and train the model to convergence. The training results are added to a scatterplot, naturally creating an estimate of the Pareto frontier.

Blockchain

Blockchain and Asset Management Part 2: The Digital Asset Manager

Audit trails — Hyperledger Besu, Pegays Plus, Quorum

ICOs shifting to STOs instead

Atomic Delivery Versus Payment (DvP) so transactions that are settled once payment is made. Reduce systemic risk in clearing

Despite the appetite for asset managers to engage with crypto assets, they must be aware that the process of buying, selling and holding digital assets today, provide no immediate operational efficiency. Based on legacy systems and evolving regulation, trading crypto assets can be complex alongside how to select the assets, how to mark them to market and how to custody them. Based on this, you must have investors who are seeking exposure to this class in search of alpha and or as a hedge.

Hacker Noon, a popular tech site, goes on the blockchain

Uses GUN (github)instead of Civil

There’s a whole world outside of Ethereum

Introduction to Multisig Contracts

Mist, Consensys, Gnosis, Argent, BitGo, Partiy (depreciated after it was hacked)

Doesn’t provide code for multisig contracts

Ethereum

Eth 2.0 Dev Update #42 — “Mainnet-Capable Testnet + Now Hiring!”

There’s a lot that happened recently and it seems much closer to Eth 2.0.

Hiring an experience software dev

Tokens,Gas and Gas limit in Ethereum

Simple explainer of what gas is like

You want to set the Gas Price high enough so that a miner includes your transaction in a block. If you are in a hurry, you can set the Gas Price higher, so that you jump ahead of everyone in line. If you are not in a hurry, you just need to set a number high enough so that someone eventually includes your transaction.

Dapp.com 2019 Annual Dapp Market Report

Pretty useful cross-chain comparisons

Data

Natural Language Processing for Web Developers

Case in point — Hugging Face’s Write With Transformer project, which allows you to write in a text editor using ML-powered autocomplete:

Now, however, machine learning is powering support bots that can do more than answer pre-selected questions. These new support bots can field a user’s question and, using your FAQs and documentation, respond with a personalized answer using AllenNLP’s ELMo-BiDAF:

To build a text summarizer that monitors your niche and produces short, relevant summaries, you might need:

• A scraper to parse reports from relevant outlets
• An API that can analyze a body of text and return a summary
• A dashboard or feed to display your summaries

Google’s BERT is a popular NLP model, and more importantly, it is implemented in a community project called the BERT Extractive Summarizer which uses BERT for text summarization. You can implement the summarizer in a few lines of code, and instantly have a text summarization API.

Development

How to Build an Energy Market on a Blockchain

One contract wonder — good to consider for course

Go’s Tooling is an Undervalued Technology

Makes me want to try out Go. Everything else has been such a pain to learn

Build a trading simulator in Python

But what about your trades making the prices change?

Dashboards in Python for Beginners and Everyone Else using Dash

Simple tools to make life easier

Leadership

This Is What Impactful Engineering Leadership Looks Like

Once a quarter, I’ll say, ‘Okay next week is career growth week,’ and that’s what we’ll talk about during one-on-ones.

It’s a shame when companies force people to move into management as the only option for career growth. It’s a clear mistake.

There’s an inflection point when someone moves from engineer to manager, and it can feel very uncomfortable — like you’re only in meetings and not getting anything done.

You need to develop local experts within teams, projects and platforms who can think deeply and guide architectural discussions.

You want to be leveling up everyone on your team all the time. Execution will follow.

You want systems that both catch when a comma is missing and get people to think more broadly about what good architecture looks like.

A thoughtful engineering manager recognizes all hard work, especially if it’s not super flashy.

What I Learned Raising Capital

Market Risk: If you are young and have no skills, and no relevant experience, then take on market risk. Do something where no one is there yet, and see if you can create a market. Justin.tv (24-hour streaming POV) became Twitch (watch gamers play). There was no market for streaming media consumption at the time.

Execution Risk: If you are experienced and well networked, then take execution risk. Compete on how well you can build, market, sell, hire, and iterate. Atrium is execution risk on making legal services better, a well established market.

Security

20 Hours, $18, and 11 Million Passwords Cracked

Basically an advertisement fo Hashcat