<?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 Michael Bleigh on Medium]]></title>
        <description><![CDATA[Stories by Michael Bleigh on Medium]]></description>
        <link>https://medium.com/@mbleigh?source=rss-a6df102dcd62------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*TA25oUW6z-_vNf9gfSThHw.jpeg</url>
            <title>Stories by Michael Bleigh on Medium</title>
            <link>https://medium.com/@mbleigh?source=rss-a6df102dcd62------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 09:07:15 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mbleigh/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[How to add natural language AI data filters to your app]]></title>
            <link>https://medium.com/firebase-developers/how-to-add-natural-language-ai-data-filters-to-your-app-71d64a79624d?source=rss-a6df102dcd62------2</link>
            <guid isPermaLink="false">https://medium.com/p/71d64a79624d</guid>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[tutorial]]></category>
            <category><![CDATA[firebase]]></category>
            <dc:creator><![CDATA[Michael Bleigh]]></dc:creator>
            <pubDate>Wed, 17 Jul 2024 19:51:37 GMT</pubDate>
            <atom:updated>2024-07-17T23:01:07.996Z</atom:updated>
            <cc:license>https://creativecommons.org/licenses/by-nc/4.0/</cc:license>
            <content:encoded><![CDATA[<p>When building an app there’s often a tradeoff between “approachable” features for most users and “powerful” features for expert users. A good example of this is data filtering — if your app has listings with lots of different dimensions you want to expose all the options to your users, but you don’t want to overwhelm them. Let’s say you’re building a sportswear ecommerce site with many different filter options:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/249/1*PgUo3DHxUCGDhPzf-Qkgyw.png" /></figure><p>Wouldn’t it be great to offer your users the flexibility of a text search bar combined with the</p><p>fine-grained controls of a filter UI?</p><h4><strong>The Goal: Query to Filter</strong></h4><p>I’ve already implemented a rich product filtering interface inside my Next.js app. The filters that can be provided are expressed by this Zod schema:</p><pre>// example enum, there are more<br>export const SizeEnum = z.enum([&quot;XS&quot;, &quot;S&quot;, &quot;M&quot;, &quot;L&quot;, &quot;XL&quot;, &quot;XXL&quot;]);<br><br>export const FilterSchema = z<br> .object({<br>  query: z<br>   .string()<br>   .describe(&quot;a full-text search query of the name and description of the product&quot;),<br>  pickUpToday: z.boolean(),<br>  onSale: z.boolean(),<br>  gender: z.array(GenderEnum),<br>  color: z.array(ColorEnum),<br>  minRating: z.number().describe(&quot;on a scale of 0.0 - 5.0&quot;),<br>  maxRating: z.number().describe(&quot;on a scale of 0.0 - 5.0&quot;),<br>  minPrice: z.number().describe(&quot;in USD&quot;),<br>  maxPrice: z.number().describe(&quot;in USD&quot;),<br>  size: z.array(SizeEnum),<br>  brand: z.array(BrandEnum),<br>  activity: z.array(ActivityEnum),<br>  collection: z.array(CollectionEnum),<br>  features: z.array(FeatureEnum),<br>  technology: z.array(TechnologyEnum),<br>  fit: z.array(FitEnum),<br>  sort: SortEnum,<br> })<br> .partial();<br><br>export type Filter = z.infer&lt;typeof FilterSchema&gt;;</pre><p>What I want to be able to do is add a simple text search bar that can translate a natural language query, for example:</p><blockquote><em>shirt under $30 available for pickup</em></blockquote><p>into a filter that matches my existing schema:</p><pre>{<br>  &quot;query&quot;: &quot;shirt&quot;,<br>  &quot;pickUpToday&quot;: true,<br>  &quot;maxPrice&quot;: 30<br>}</pre><p>One of the most powerful capabilities of Large Language Models (LLMs) is their ability to turn unstructured language into structured data. Given enough instruction, an LLM can understand enough about a user’s query to turn it into a structured query.</p><p>You may have heard of “Retrieval Augmented Generation” or RAG…well, this is sort of the opposite! This is Generation Augmented Retrieval — we’re using Generative AI to help our users perform a “traditional” query of our product listing.</p><h4><strong>Tool Setup: Genkit and Gemini</strong></h4><p>To build our natural language filter we’re going to use <a href="https://github.com/firebase/genkit">Firebase Genkit</a>. Genkit is a library that makes it easy to add GenAI features to an application and provides a standard API that works across many model providers. For this example we’ll use Google’s <strong>Gemini 1.5 Flash </strong>model for its fast inference speed.</p><p>You can <a href="https://firebase.google.com/docs/genkit/nextjs">follow the documentation</a> to initialize Genkit in your app, or you can do it manually like so:</p><pre>npm i \<br>  @genkit-ai/core \<br>  @genkit-ai/ai \<br>  @genkit-ai/flow \<br>  @genkit-ai/googleai \<br>  @genkit-ai/dotprompt</pre><p>Once you’ve installed the packages, you’ll need to configure Genkit with the Google AI plugin (for access to the Gemini model) and the Dotprompt plugin (for the `.prompt` text prompt syntax):</p><pre>import { configureGenkit } from &quot;@genkit-ai/core&quot;;<br>import { dotprompt } from &quot;@genkit-ai/dotprompt&quot;;<br>import { googleAI } from &quot;@genkit-ai/googleai&quot;;<br><br>configureGenkit({<br> plugins: [dotprompt({ dir: &quot;./prompts&quot; }), googleAI({ apiKey: &quot;...&quot; })],<br>});</pre><p>You can get an API key from <a href="https://aistudio.google.com/">Google AI Studio</a>.</p><h4><strong>Writing the prompt</strong></h4><p>Genkit has built-in mechanisms for conforming output to a supplied schema, and since I already have my schema in Zod this is even easier. Below my <em>configureGenkit</em> block, I can register a named schema for use in my prompts:</p><pre>import { defineSchema } from &quot;@genkit-ai/core&quot;;<br>import { FilterSchema } from &quot;./schema.js&quot;;<br><br>defineSchema(&#39;Filter&#39;, FilterSchema);</pre><p>Next, I’ll create the <strong>prompts/productFilter.prompt</strong> file. Genkit’s Dotprompt library lets me define both the content of my prompt as a rich template and supply the metadata for my prompt (such as the model and output schema).</p><pre>---<br>model: googleai/gemini-1.5-flash<br>output:<br>  schema: Filter<br>---<br><br>You are helping a user filter through a selection of activewear products.<br>Given the following query supplied by the user, convert their intent into<br>the provided schema for a filter.<br><br>### Filtering Hints ###<br><br>- When a user says something like &quot;best&quot;, sort by rating.<br>- When a user says something like &quot;cheapest&quot;, sort by price.<br>- When a user says &quot;popular&quot;, sort by popularity.<br>- Only include &quot;query&quot; if the other filters are inadequate.<br>{{#if existing}}<br><br>### Existing Filters ###<br><br>The user has requested a refinement to an existing set of filters. Use the<br>existing filters as a starting point and make modifications to it. Existing<br>filters:<br><br>{{json existing}}{{/if}}<br><br>### User-Supplied Query ###<br><br>{{query}}<br><br>### Output Instructions ###<br><br>Just supply the JSON without additional text or explanation. All fields are<br>optional. Include only fields that are set to a non-null value.</pre><p>A few things to note about the prompt:</p><ul><li>Since I’ve already initialized the `googleAI` plugin I can use the Gemini 1.5 Flash model by name.</li><li>I don’t need to describe the output schema in the prompt. Genkit takes care of that for me since I provided a named schema in the frontmatter of the prompt.</li><li>If present, I provide the existing filter the user has selected (rendered as JSON) so that the LLM can refine a search as well as start from scratch.</li><li>I’ve added hints to help the LLM understand specific scenarios that I ran into while testing. You should expect prompts to evolve over time as you run into edge cases. Including high-quality examples is another great way to improve quality.</li></ul><h4><strong>Integrating into the app</strong></h4><p>I’m using Next.js here, but Genkit can be used in any Node.js server environment. First, I’ll load and call my prompt as part of a Genkit <em>*flow*</em>, which you can think of as a special kind of function wrapped with tracing and observability for GenAI:</p><pre>import { defineFlow } from &quot;@genkit-ai/flow&quot;;<br>import { prompt } from &quot;@genkit-ai/dotprompt&quot;;<br>import { z } from &quot;zod&quot;;<br>import { FilterSchema } from &quot;./schema.js&quot;;<br><br>const magicFilterFlow = defineFlow({<br>  name: &quot;magicFilter&quot;,<br>  inputSchema: z.object({<br>    query: z.string(),<br>    filter: FilterSchema.optional(),<br>  }),<br>  outputSchema: FilterSchema,<br>}, (input) =&gt; {<br>  const filterPrompt = prompt(&#39;productFilter&#39;);<br>  const result = await filterPrompt.generate({input});<br>  return result.output();<br>});</pre><p>Now I can create a server action that executes the flow:</p><pre>&quot;use server&quot;<br><br>import { runFlow } from &quot;@genkit-ai/flow&quot;;<br>import { magicFilterFlow } from &quot;./flow.js&quot;;<br>import { FilterSchema } from &quot;./schema.js&quot;;<br>import { z } from &quot;zod&quot;;<br><br>export function magicFilter(input: {<br>  query: string,<br>  filter: Filter<br>}): Filter {<br>  return runFlow(magicFilterFlow, input);<br>}<br>```<br><br>and, finally, I can call the server action from my client component:<br><br>```ts<br>const [filter, setFilter] = useState&lt;Filter&gt;({});<br>const [query, setQuery] = useState&lt;string | null&gt;(null);<br><br>function handleSubmit(e) {<br>  const newFilter = await magicFilter({query, filter});<br>  setFilter(newFilter);<br>}</pre><p>and, finally, I can call the server action from my client component:</p><pre>const [filter, setFilter] = useState&lt;Filter&gt;({});<br>const [query, setQuery] = useState&lt;string | null&gt;(null);<br><br>function handleSubmit(e) {<br>  const newFilter = await magicFilter({query, filter});<br>  setFilter(newFilter);<br>}</pre><h4><strong>Wrapping up</strong></h4><p>There we have it! We’ve built an AI-powered natural language filter that works on top of an existing search/filter mechanism without having to reinvent everything from scratch. Bringing it into the app, it looks like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*bfZTw1gP1Gh0qfezGQ9fPw.gif" /></figure><p>Is this perfect? Of course not. You’ll find that some portion of the time a query is misunderstood or a filter applied incorrectly. However the stakes are low here: the user can always manually correct the LLM’s mistake.</p><p>My favorite AI features are the ones that disappear into the application and feel like a natural part of the interface. Everyone is familiar with a search bar, but thanks to GenAI our search bar can directly manipulate a complex interface by better understanding our users’ intent.</p><p>The <a href="https://github.com/mbleigh/genkitchen/tree/main/src/app/filter">full source code</a> of the demo application is available on GitHub.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=71d64a79624d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/firebase-developers/how-to-add-natural-language-ai-data-filters-to-your-app-71d64a79624d">How to add natural language AI data filters to your app</a> was originally published in <a href="https://medium.com/firebase-developers">Firebase Developers</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Idea to Game in 24 hours with Firebase]]></title>
            <link>https://medium.com/firebase-developers/pictle-idea-to-game-in-24-hours-with-firebase-a18e0da8475f?source=rss-a6df102dcd62------2</link>
            <guid isPermaLink="false">https://medium.com/p/a18e0da8475f</guid>
            <category><![CDATA[wordle]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[firebase]]></category>
            <dc:creator><![CDATA[Michael Bleigh]]></dc:creator>
            <pubDate>Wed, 19 Jan 2022 21:52:08 GMT</pubDate>
            <atom:updated>2022-01-19T21:52:08.033Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7_WQSQ2hQo7NZcVYhfWUkw.png" /></figure><p>Like seemingly everyone else on the planet, I’ve been enjoying my daily fix of <a href="https://www.powerlanguage.co.uk/wordle/">Wordle</a>. Several of my teammates at Firebase are playing it and we share our grids in a chat. In addition to being a fun puzzle, I love how Wordle represents the very best of the web — it’s just a fun little solo project, not even on its own domain. It became popular because it’s clever and because of its wholesome engagement mechanics (one puzzle a day so you keep coming back, and an absolutely genius semi-cryptic emojiful social mechanism). It’s just…lovely!</p><p>One morning, after discussing the interesting patterns our solutions had made, I came up with an idea:</p><figure><img alt="Chat transcript: omg i just thought of a wordle variant game that might actually be fun Pictle you get the correct word in advance but then it gives you a 6x6 pixel art you have to recreate by putting in wordle-valid guesses anyone want to build it with me?" src="https://cdn-images-1.medium.com/max/860/1*TWHiib7F3ACpoV4v2By4wA.png" /></figure><p>24 hours later, I launched <a href="https://pictle.web.app/"><strong>Pictle</strong></a> publicly on Twitter:</p><h3>Michael Bleigh on Twitter: &quot;Ever thought: &quot;My Wordle solution almost looks like pixel art&quot;?🖼️ Pictle is a (surprisingly challenging!) companion game to Wordle that lets you paint with your words.Built in &lt; 24h, I&#39;d love to hear what you think!https://t.co/urpxbMQi5f / Twitter&quot;</h3><p>Ever thought: &quot;My Wordle solution almost looks like pixel art&quot;?🖼️ Pictle is a (surprisingly challenging!) companion game to Wordle that lets you paint with your words.Built in &lt; 24h, I&#39;d love to hear what you think!https://t.co/urpxbMQi5f</p><p>I’ve spent most of my career building tools and platforms to make it faster and easier for developers everywhere to ship their ideas, but I don’t get as many opportunities as I’d like to go through that process myself. I thought it might be interesting to break down the approach used building Pictle.</p><h3>How to Ship Fast</h3><p>I have a young daughter at home and a busy family life; I don’t have the time or inclination to spend weeks polishing a hobby project. Fortunately, after more than a decade of hackathons, side projects, and prototypes, I’ve gotten pretty good at cutting projects down to a true MVP. A few tips:</p><ol><li><strong>Don’t just clone, make it your own.</strong> I often see prototype projects published that are clones of other products. Having a starting point is great — with Wordle as inspiration I could just “default” so many decisions around layout, color, word length, etc. But you’ll have a much more rewarding experience if you do something to put a unique spin on it. Try to come up with <em>one clever idea </em>that makes your project more than just a piece of sample code. For me, it was a visual twist on the word game.</li><li><strong>Focus on core interactions.</strong> Fake whatever you have to until the core interactions for your app/game are implemented. There’s always going to be more to add, but if you don’t get the critical mechanics right you won’t have a working product. For Pictle I hard-coded a single puzzle and iterated on the feel of entering the words. I found that the game was much more pleasant to play when the letters gave color cues as you typed instead of only on hitting submit.</li><li><strong>Are you learning or shipping?</strong> Every time I build a project from scratch I have to decide if I’m focusing on learning or focusing on shipping. If the former, I’ll try out that new framework or library that I’ve been hearing about knowing that it’ll take me longer but I’ll learn something new. If the latter, I reach for the tools I know best knowing that it’ll give me the shortest path to launch. Both approaches are fine, but you should choose intentionally!</li><li><strong>Get production running first.</strong> This especially applies to actual deadline-driven scenarios like hackathons…you don’t have a product until it’s deployed live somewhere, so why wait until the end to do that? If you get your deployment process set up early you’re ready to share for feedback as soon as you decide the code is ready. I had Pictle deployed to Firebase Hosting within an hour or two of creating the project folder on my machine.</li><li><strong>It’s never too early to automate.</strong> Similar to (4) above, getting your tooling in order can save a ton of time even on quick projects. For Pictle I set up GitHub actions to automatically deploy to prod/staging based on pushes to repo branches. I also built myself a little tool for drawing the puzzles and checking to make sure they weren’t impossible (it was surprisingly easy to make impossible patterns!). More on that below.</li><li><strong>Measure something.</strong> This could be dropping in an analytics package, setting up a Twitter search feed to watch, or any number of other things, but figure out at least one way that you’re going to measure the launch of your project. If you can’t tell if anyone’s using it, how will you know if it’s worth your continued effort?</li></ol><h3>Building Less</h3><p>A big part of building something quickly is <em>not </em>building a lot of things that you might think you need. For the initial launch, I cut out everything that wasn’t immediately going to be useful to new players or necessary for me.</p><h4>Fake admin ’til you make admin</h4><p>Pictle needs a new puzzle every day based on the Wordle from 7 days earlier. I was originally going to use the Wordle solution from the day before, but realized quickly that would mean I’d have to add a new puzzle <em>every day </em>and the game would break if I forgot. Instead, I chose “one week ago” to give myself a little buffer, but how to actually craft and store the daily puzzles?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/456/1*0aDR_YzTW04xQuUn3bAuCw.png" /><figcaption>My hastily crafted “puzzle maker” tool.</figcaption></figure><p>I built an extremely simple tool that allowed me to click boxes in a grid to flip them between orange, green, and empty. This tool also takes a “seed word” (the solution) and lists the number of valid Wordle words that match the pattern based on the seed word.</p><p>I then serialize the picture into a simple string (0: empty, 1: orange, 2: green, space: new line).</p><p>As for adding it to the database? At the moment, I’m still doing it manually using the Realtime Database console. I’ll probably add more sophisticated “admin” authentication in the future, but my experience as an admin wasn’t the kind of “core experience” I needed to focus on to launch.</p><h4>Security without servers</h4><p>I wanted to make new puzzles available automatically at a specific time, and also make sure that it wasn’t possible to “peek ahead” at future puzzles. I could have built an API server for this, but why not do a little Firebase magic instead? Using Realtime Database security rules, a new puzzle automatically becomes readable once a day:</p><pre>{<br>  &quot;rules&quot;: {<br>    &quot;puzzles&quot;: {<br>      &quot;$puzzle_id&quot;: {<br>        &quot;.read&quot;: &quot;now &gt; (<br>          // timestamp for the first puzzle to start<br>          1642003200000 +<br>          (<br>            // id of the puzzle minus the start id<br>            data.child(&#39;id&#39;).val() - 200<br>          )<br>          // multiply by 24 hours in millis<br>          * 86400000<br>        )&quot;<br>      }<br>    }<br>  }<br>}</pre><p>This simple rule generates a “available time” for each puzzle based on a numeric ID, then simply denies access before that time. The puzzle data is safely out of the client, no servers required!</p><h3>The Pictle Stack</h3><p>What would a “how I built &lt;thing&gt;” post be without an obligatory rundown of the technology stack? Here’s the loadout I found useful when building the game (which you can inspect in more detail in the game’s <a href="https://github.com/mbleigh/pictle">open source repo</a>):</p><ul><li><a href="https://kit.svelte.dev/"><strong>SvelteKit</strong></a><strong>. </strong>I had never used Svelte before a few weeks ago, but decided to put it to the test in a real-world scenario. I’m not sure I’ve ever gone from zero to semi-competent with a new framework so quickly thanks in large part to the truly excellent <a href="https://svelte.dev/tutorial/basics">tutorial</a>. The batteries-included “transitional framework” SvelteKit strikes a great balance between feeling like a framework and letting me “just write JavaScript.” Honestly, I’ve been so impressed with Svelte/SvelteKit that it’s likely going to be my default tool when I start new projects.</li><li><a href="https://www.typescriptlang.org/"><strong>TypeScript</strong></a><strong>.</strong> Even for a solo project where I’m perfectly happy to cast to <em>any </em>if a variable looks at me funny, I still love to use TypeScript because it makes me more productive in the editor thanks to stronger autocomplete and auto-import capabilities when used.</li><li><a href="https://tailwindcss.com/"><strong>Tailwind CSS</strong></a><strong>.</strong> I was skeptical of Tailwind at first. How could sprinkling dozens of utility classes in HTML possibly be faster than just writing CSS (which I, unlike many others, do <strong>not </strong>hate)? After giving it a try, and especially after getting the VS Code plugin working, I was won over. Tailwind lets you rough out custom componentry in record time, and with autocomplete and a little practice it’s the fastest way I know to build UI. I’d still love to see tools that make it easy to extract groups of utilities into reusable classes, but it really does get the job done.</li><li><a href="https://firebase.google.com/"><strong>Firebase</strong></a><strong>.</strong> Working for Firebase it’s hard not to sound like a shill, but I am honestly always so happy to feel how the platform comes together when I get the chance to build an app from scratch. The game so far uses Firebase Hosting, Realtime Database, and Google Analytics for Firebase. Honestly I spent maybe 5% of the development time wrangling Firebase, but that’s because it gets out of my way and just does its job.</li></ul><p>One thing that surprised me after launching Pictle is how slim my list of runtime dependencies ends up being. With Svelte compiling the app down to vanilla HTML/JS/CSS, my package.json ended up looking like this (shoutout to the always delightful <a href="https://github.com/loonywizard/js-confetti">js-confetti</a> package!):</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/247/1*3sRuS-sshJmlElCtlU7SmA.png" /><figcaption>That’s it?!?!</figcaption></figure><h3>How It‘s Going</h3><p>Did Pictle become the same kind of viral sensation as Wordle, driving thousands of tweets and tens of thousands of new users? Nah. But it has seen several hundred visitors and some nice consistent use from a good number of them. One encouraging sign is that so far more than 35% of all visitors go on to complete at least one puzzle:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/820/1*WrjcGlDF6_WgcOUJWKe-xQ.png" /></figure><p>Since I’m hosting it all on serverless infrastructure with Firebase, there’s nothing that I have<em> </em>to do to keep the game alive beyond spending a couple minutes creating new puzzles each day. It was fun experience to build something that wasn’t strictly for developers, and if it brings a few dozen people a little bit of amusement, that’s good enough for me. After all, it was the wholesome “just for fun” nature of Wordle that inspired me to build it in the first place!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a18e0da8475f" width="1" height="1" alt=""><hr><p><a href="https://medium.com/firebase-developers/pictle-idea-to-game-in-24-hours-with-firebase-a18e0da8475f">Idea to Game in 24 hours with Firebase</a> was originally published in <a href="https://medium.com/firebase-developers">Firebase Developers</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A crash course in equity compensation at tech companies]]></title>
            <link>https://medium.com/@mbleigh/tech-equity-compensation-crash-course-e887291bfcbc?source=rss-a6df102dcd62------2</link>
            <guid isPermaLink="false">https://medium.com/p/e887291bfcbc</guid>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[money]]></category>
            <dc:creator><![CDATA[Michael Bleigh]]></dc:creator>
            <pubDate>Fri, 10 Jan 2020 21:59:53 GMT</pubDate>
            <atom:updated>2020-01-14T21:02:22.568Z</atom:updated>
            <cc:license>https://creativecommons.org/licenses/by-nc/4.0/</cc:license>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DPMUM5GJ17D_PRbSe-_JYw.png" /></figure><p>If you’re new to the tech industry or, like me, not the kind of person who is generally attuned to the intricacies of personal finance, you may not have a great idea of how equity compensation works beyond “ooh equity, that sounds fancy.”</p><p>Other parts of your compensation are fairly straightforward: salary (how much is in your paycheck), benefits (e.g. health insurance, retirement account), and bonus (big check once a year). Equity compensation is a different beast altogether.</p><p>Equity compensation means that in addition to other things, the company is offering you some kind of stake in the company, usually in the form of shares of stock. However, all equity compensation is not created equal. Various aspects can drastically change how you should think about equity compensation.</p><p>Why should you care? Because equity scales faster and further than salary and bonus! For senior-level employees at large tech companies, <strong>equity compensation ends up eventually exceeding salary and bonus</strong>. This goes all the way to the top — a CEO might earn 5x the salary of a senior employee, but earn 100x or more in equity compensation.</p><p><strong>Note:</strong> I am not a financial expert, quite the opposite! However, I have gradually become familiar with lots of things related to equity compensation from being in tech for a long time. This post tries to summarize everything I wish I’d known about equity compensation but didn’t.</p><h4><strong>Options v. Grants</strong></h4><p><strong>Stock options</strong> are what you find most often in startups, and essentially offer you the ability to buy shares (called “exercising” your options) in the company at a certain price (called the “strike price”) at some time in the future.</p><blockquote>Example:<em> Ada has stock options for 10,000 shares with a strike price of $1.00 per share. Later, the company’s stock has risen 10x to be worth $10.00 per share. Alice exercises her 10,000 options by paying $10,000 to purchase them at the strike price. She is now holding stock worth $100,000 at current value (a $90,000 profit).</em></blockquote><p><strong>Stock grants </strong>are generally more straightforward — the company is paying you shares of stock as part of your compensation. These are typically structured as Restricted Stock Units (RSUs) and, often on a periodic basis, you will simply receive shares in the company as part of your income. Grants are much more common in larger, later-stage companies and are most common (but not exclusively) found in public companies.</p><h4>Liquidity</h4><p>Liquidity refers to how easily a given asset can be converted to cash, and is <strong>extremely important </strong>when it comes to evaluating equity compensation.</p><p>Publicly traded companies have liquid equity: at any point in time (barring time-based restrictions around insider trading) <strong>you can sell shares you hold for cash money</strong>. When evaluating an offer from a publicly-traded company, you can consider equity compensation to be nearly (but not quite) equivalent to cash compensation.</p><p>Private companies (startups et al) have illiquid equity: you own shares but there is <strong>no easy way to simply sell them for cash</strong>. You may not be able to sell any of your shares until there is a <em>liquidity event</em> (the company goes public or is bought by another company). This means you need to think about equity differently. It’s <em>not </em>nearly equivalent to cash, and depending on the stage of the company it’s very very <em>very </em>not equivalent to cash. You should discount the value of equity substantially in private companies, and drastically more so in early-stage companies.</p><p>For very early stage companies the share price and number of shares are fairly meaningless — instead, find out the total share of the company you are being offered. Remember also that this share will likely decrease over time due to dilution from new investment (where new shares are added to the total, thus reducing all existing shares total ownership percent). Unless you’re a founder or a key employee being offered, say, 2% or greater stake, you can think of this kind of equity as a lottery ticket (fun to have, but assume it will be worth nothing).</p><h4>Vesting and Cliffs</h4><p>Regardless of the type of equity compensation (stock/grant), you will likely be subject to a <strong>vesting schedule </strong>for your stock. This means that you will receive the options / granted stock gradually over time and only while you still work for the company. Equity compensation is generally used as a retention tool — you benefit from staying at the company because you still have equity that’s yet to be paid out to you.</p><p>How your stock vests can be very important, especially at large publicly traded companies. There may be a vesting <strong>cliff</strong> (most commonly one year) where you need to work at a company for a period of time before you receive anything.</p><blockquote>Example: <em>Grace receives a sign-on equity grant of 4,800 shares vesting monthly over four years with a one-year cliff. For the first 12 months, she receives nothing. After one year she has reached the vesting cliff and receives 1,200 shares all at once (a full year’s worth). Going forward, she receives 100 shares each month as they vest until the grant runs out after four years.</em></blockquote><p>If you’re well-established and have a significant savings reserve already, vesting details might not be super-important. If you don’t, however, you should educate yourself on exactly when you’ll start receiving equity and when you’ll be able to sell it.</p><p>Vesting schedules are one way companies compete in offers — recently some companies have done away with cliffs altogether such that you begin vesting on a monthly basis immediately. At the other extreme, some companies have multi-year cliffs or only vest stock annually.</p><h4>Thinking Through Job Offers</h4><p>When I received job offers, salary was <em>the number. </em>I wanted to make sure other stuff was there (e.g. health insurance), but salary was the only thing I was really looking at or trying to negotiate. This is a mistake!</p><p>Salary can and should be negotiated, but companies often have fairly tight constraints on salary. At large tech companies, there are salary bands for each level that can’t be circumvented without significant escalation. At small companies that are concerned about burn rate (how fast money goes out of the bank), salary is the hardest thing to compromise on.</p><p>Equity is where the company usually has more discretionary wiggle room, and therefore you have more space to negotiate. When making a counter-offer, you can likely get away with a much larger equity increase than salary. This is also only fair to you: equity vests over time and has risk of losing value, so $10k in extra salary is more valuable than $10k in extra equity.</p><p>This all plays into counter responses, too. If you’re applying to an early stage startup and ask for a $10k salary bump, they might come back and say “we can only do $5k salary but we’ll also do $5k equity.” Remember our massive discounting! That $5k should probably be $50k+ to offset the relative dollar-to-dollar value and risk associated.</p><p>Evaluate offers holistically and try to get a handle on your total annual compensation. Use resources like <a href="https://levels.fyi/">levels.fyi</a> to get an idea of where your compensation ought to be, and use that as a reference point. Just make sure to consider equity when considering the offer overall!</p><h4>Equity Refreshes</h4><p>After you’ve joined a company, you may receive additional stock options/grants over time as part of performance/compensation cycles. These refreshes will usually be smaller than your sign-on offer. It works this way because refreshes stack on top of your still-vesting equity.</p><blockquote>Example:<em> Grace received a sign-on equity grant of 4,800 shares vesting over four years and is now one year into her new job. She receives a refresh grant of 960 shares that also vest over four years (but unlike her sign-on grant, there is no one-year cliff). She now vests 120 shares per month. Without additional grants, she will vest at 120 shares per month in years 2, 3, and 4, and then in year 5 she will drop to 20 shares a month.</em></blockquote><p>An important side effect of the stacking of refresh grants is that there can be a significant impact on compensation when the sign-on grant is fully vested. This is often called the <strong>four-year cliff </strong>(since most equity grants vest over four years). In large tech companies, it’s not unusual for your total compensation to <em>go down</em> after four years <em>even if you’ve been performing well.</em></p><p>Why does this happen? For a few reasons:</p><ol><li>Sign-on grants are usually significantly larger than individual refresh grants to serve as an incentive to join and an incentive to stay for a decent amount of time.</li><li>If your company is doing well, the stock price may be going up significantly. Equity refresh grants are often allocated in <em>dollars </em>as opposed to <em>shares</em>, so it can be possible that the dollar value of a grant is larger but the number of shares is smaller.</li><li>Because reasons? 🤷</li></ol><p>Honestly, it’s a bit surprising to me that companies allow four-year cliffs to happen as much as they do. You’d think that retaining seasoned employees would be a top priority, but the system doesn’t seem to be set up to make that happen well. This can be why you’ll often see employee departures either after one year (first cliff) or four years (sign-on grant fully vests).</p><h4>Wrapping Up</h4><p>Companies have all kinds of ways to attract and retain employees: the work, the mission, the team, the free food. Equity compensation is just one piece of that puzzle, but it’s one that will continue to grow in importance throughout your career and may eventually become the single biggest factor in your overall compensation.</p><p>No one sat me down and told me any of this. I learned most of it after the fact, sometimes realizing I had been lucky and other times realizing I screwed up (oops! giant tax bill!). My hope is to demystify it all a little bit to put others on a more level playing field that I found myself.</p><p>Good luck out there. May your vesting schedule be rapid and your refreshes ever compounding!</p><p><em>Postscript: The following applies to stock grants in publicly-traded companies. If you’re looking at a private company or stock options, parts may not apply and feel free to skip.</em></p><h4>To Hold or to Sell?</h4><p>Once your stock has vested, you need to decide what to do with it — do you hold onto it or sell it? There are some who ascribe to immediately selling equity as it vests, even enrolling in “autosale” programs that will do this automatically. The theory goes that because you have a significant investment in the company already (they pay your salary, you have unvested stock), it is wise to immediately sell to diversify your holdings.</p><p>You may also decide not to sell any (if you don’t need the money) and hold it as an investment that will hopefully go up over time. You may also choose to hold some and sell some. Here there isn’t (in my opinion) a single right answer but you should be intentional about whether you’re holding or selling (don’t just ignore the accumulating stock in your account!).</p><h4>Taxes</h4><p>When your stock vests, the company may sell a portion of the vested shares immediately to use for tax withholding. Stock grants count as regular income for tax purposes, so you need to report it just like you do your salary and bonus. This has implications come tax time!</p><p>Since your stock compensation is added to your salary and bonus, taken all together you might be in a higher tax bracket than you expect. You should make sure to adjust your voluntary withholding so you don’t end up with a massive surprise bill come tax season (been there, done that!). You may also be able to increase the withholding of the stock vesting.</p><p>Another important tax consideration for equity is capital gains tax. This is the tax you pay on the <em>increased value </em>of your stock when you sell it vs. when it vested. This will be either <strong>short-term </strong>or <strong>long-term</strong> capital gains tax, depending on whether you’ve held the stock for more than a year before selling it. This means that when you sell shares, the specific shares that you sell matter! An approach to picking shares to sell that I use is:</p><ol><li>Sell any shares that you’ve held for more than a year and will therefore be subject to long-term capital gains tax.</li><li>Sell shares that vested at a higher price than the current stock price, if any (you won’t have to pay any capital gains tax on this).</li><li>Sell either the shares that vested at the highest value or the most recently vested shares, depending somewhat on the specifics.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e887291bfcbc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[When should I Server-Side Render?]]></title>
            <link>https://medium.com/@mbleigh/when-should-i-server-side-render-c2a383ff2d0f?source=rss-a6df102dcd62------2</link>
            <guid isPermaLink="false">https://medium.com/p/c2a383ff2d0f</guid>
            <category><![CDATA[product]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[engineering]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Michael Bleigh]]></dc:creator>
            <pubDate>Wed, 01 Nov 2017 17:34:11 GMT</pubDate>
            <atom:updated>2017-11-10T02:44:04.930Z</atom:updated>
            <cc:license>http://creativecommons.org/licenses/by/4.0/</cc:license>
            <content:encoded><![CDATA[<p>One of the most common debates I see surrounding modern web tech revolves around server-side rendering. Some people will tell you it’s absolutely critical, others will say it’s nice-to-have, other still will tell you it’s downright harmful. So which is it? Well, all three. Like everything else in programming and life, <strong>it depends.</strong></p><p>When you’re sitting down to decide on the architecture for your modern, performant, Progressive Web App, you will need to decide whether or not to use Server-Side Rendering (henceforth SSR). First, <strong>you should avoid SSR if you don’t need it. </strong>Most modern web apps require sophisticated interaction with the UI that has to be driven by non-trivial amounts of JavaScript. If you need to write all of that JS anyway, and you <em>don’t </em>need the first page load benefits that SSR gives you, you’ll be better off just building a <a href="https://developers.google.com/web/fundamentals/architecture/app-shell">static app shell</a> and avoiding the headaches of client-server code reuse, data re-hydration, and dynamic content cache invalidation.</p><p>To know whether or not you need SSR, just answer these questions:</p><ol><li><strong>Does your app require a sign-in to view most content (e.g. GMail)?<br></strong> If so, <strong><em>you don’t need SSR</em>. </strong>Build a static app shell that pops up the login page with as fast of Time-To-Interactive as you can, and preload all of your app’s assets and scripts while the user types in their credentials. Use a service worker to cache your app shell so subsequent loads remain super-speedy.</li><li><strong>Is your app more oriented around reading or doing (content-focused vs. interaction-focused)?</strong><br>If most users are going to be <em>doing something </em>soon after they land on the page (e.g. a note-taking app), <strong><em>you don’t need SSR. </em></strong>SSR can, in some cases, put your app into an “uncanny valley” state where you can <em>see </em>the page but you can’t <em>interact with </em>the page. This ends up being a worse experience than a spinner or loading screen, as clicks or taps will simply stall out or do nothing.</li><li><strong>Is SEO / social sharing important to your app (e.g. ecommerce)?<br></strong>If so, <strong><em>you need (at least limited) SSR</em>. </strong>While some search engines now execute JavaScript and are able to parse and index JS-only sites, all crawlers are capable of reading content returned from the server. In some cases (if you only care about social sharing of your content, for example), you can get away with limited SSR that only injects &lt;meta&gt; tags into the head of your HTML, with the actual page render still happening in JavaScript. Related to this are technologies like <a href="https://www.ampproject.org/">AMP</a>, which mandate that content be on-page to be considered valid.</li><li><strong>Are first-time users likely to visit via deep content links? (e.g. news sites)<br></strong>Related but not identical to [3], if users are most likely to land on a deep content link (and reading is more important than interacting for your app), <strong><em>you probably want SSR. </em></strong>SSR will <em>always be </em>the fastest way to get a first paint because the HTML is sent over the wire in the first response and isn’t blocked on JS parsing or additional requests.</li></ol><p>That’s it! These four rules of thumb are not going to properly guide you in 100% of situations, but by and large you can think of SSR as being necessary for <strong>content-heavy sites with deep links as the first page load</strong>. If you do decide to add SSR to your site, make sure to optimize critical-path CSS, defer all JS, and make your first render appropriately disable interactive bits that aren’t going to be usable until some scripts load.</p><p>Best of luck, and hope this was helpful!</p><p><em>If you look at my bio, you might see I work on </em><a href="https://firebase.google.com/"><em>Firebase</em></a><em> and be surprised that I’d advocate SSR. The reality is that sometimes you </em><strong><em>need </em></strong><em>that extra bit of performance on first page load, and for some cases SSR is really the only way to make it happen.</em></p><p><em>Better yet, you may not have known that you can do SSR with Firebase! David East has a new video series that’s just kicked off that shows you </em><a href="https://www.youtube.com/watch?v=nbTn1czE2L8&amp;list=PLl-K7zZEsYLkbvTj8AUUCfBO7DoEHJ-ME"><em>how to do Server-Side rendering on Firebase Hosting with Cloud Functions</em></a><em>, you can find out more there.</em></p><p><em>👋🏾 Get to know the people and ideas shaping the products we use every day. </em><a href="https://usejournal.typeform.com/to/YDVRgX?utm_source=usejournal.com&amp;utm_medium=blog&amp;utm_content=michael_bleigh&amp;utm_campaign=guest_post_community"><strong><em>Subscribe to Noteworthy</em></strong></a><em> — the product &amp; design newsletter written by the </em><a href="https://usejournal.com/?/utm_source=usejournal.com&amp;utm_medium=blog&amp;utm_content=michael_bleigh&amp;utm_campaign=guest_post_community"><em>Journal</em></a><em> team.</em></p><figure><a href="https://usejournal.typeform.com/to/YDVRgX?utm_source=usejournal.com&amp;utm_medium=blog&amp;utm_content=michael_bleigh&amp;utm_campaign=guest_post_community"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*spegX3SQAdzTh74cW7qW8w.png" /></a></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c2a383ff2d0f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Fear and Feminism]]></title>
            <link>https://medium.com/@mbleigh/fear-and-feminism-9dbb9a40fdec?source=rss-a6df102dcd62------2</link>
            <guid isPermaLink="false">https://medium.com/p/9dbb9a40fdec</guid>
            <dc:creator><![CDATA[Michael Bleigh]]></dc:creator>
            <pubDate>Tue, 30 Sep 2014 23:17:56 GMT</pubDate>
            <atom:updated>2014-09-30T23:24:54.701Z</atom:updated>
            <content:encoded><![CDATA[<p>Are you a man? Have you ever heard or read a conversation about feminism, women’s rights, women in tech, etc. and felt uncomfortable? Had an opinion and wanted to contribute to the conversation but hesitated, afraid of how you might be perceived? Did you not say anything at all because of that fear? Did it leave you feeling pretty crappy about the whole experience?</p><p><strong>Good. </strong>You’ve now experienced the <em>tiniest taste</em> of the doubt, social pressure, and second-guessing our culture puts upon women every day of their lives.</p><p>Women who want to exist in a professional context have to constantly worry about appearing too bossy, too opinionated, too “mean.” Women are sometimes judged just for <em>appearing to try too hard</em>. I’ve seen it happen first-hand to the people I care most about in the world.</p><p>Remember that uncomfortable feeling from the first paragraph? Now apply it to every professional conversation you’ve ever had, from asking for a raise to just planning out the week’s work with your team. It’s the world we’ve made for women, and it sucks.</p><p>I’m terrified every time I speak about feminism. I’m worried I’ll say something unintentionally insensitive or worse, hurt the cause by misunderstanding how to be helpful. This fear makes me extraordinarily cautious about choosing my words, and makes me carefully consider my statements from others’ perspective. In other words, it’s healthy. Women don’t need me to come to their conversational rescue. I can almost always do more good by listening to and amplifying the voice of women than I can by trying to make an argument of my own.</p><p>The fear women experience is not healthy. Unlike women, I can choose to stop being afraid. I can decide I don’t care and the repercussions will be minor. Women have to care, have to consider every word and action or the repercussions can be devastating. Women have to don the mask of “just one of the guys” or risk being excluded altogether. Until women don’t have to be afraid, I’m going to choose to stay afraid.</p><p>It’s the least I can do.</p><p>I don’t want to write a post about women without taking the opportunity to amplify a woman’s voice. I came across <a href="https://medium.com/@bougie/notes-on-im-a-feminist-and-so-can-you-7b9c278a0b7c">Caroline Drucker’s talk/notes on “I’m a Feminist and so can you”</a> and it has links to tons of resources that I’ll be reading in the next few days. Really good stuff, go check it out!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9dbb9a40fdec" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Naming URL-Friendly Strings]]></title>
            <link>https://medium.com/@mbleigh/naming-url-friendly-strings-fe8674e7873d?source=rss-a6df102dcd62------2</link>
            <guid isPermaLink="false">https://medium.com/p/fe8674e7873d</guid>
            <dc:creator><![CDATA[Michael Bleigh]]></dc:creator>
            <pubDate>Thu, 27 Mar 2014 01:11:09 GMT</pubDate>
            <atom:updated>2014-03-27T01:11:09.014Z</atom:updated>
            <content:encoded><![CDATA[<h4>I’ve had enough. I’m calling them “refs”</h4><p>Every application I build, there’s a moment I dread. It’s that time when I’m creating a data model for something and I need that dreaded field. You know the one, the “short, human-readable URL-friendly string” that can be used as an identifier for the data.</p><p>OK, well <strong>id</strong> is already taken, that’s the unique auto-generated identifier. Some people call it a <strong>slug</strong>, but that seems wrong. I could use <strong>username</strong> or <strong>login</strong>, but what about when it’s not a user? Maybe <strong>param</strong> or <strong>segment</strong>, but those sound so technical. What about <strong>handle</strong> or <strong>nick</strong>? Eh.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/300/1*iO6uAEdVfukevOJAzO4oQQ.gif" /><figcaption>Image by <a href="http://ashlyna.com/">Ashlyn Anstee</a></figcaption></figure><p>I’ve put this question to Twitter several times and no one seemed to have any better idea than I did. There doesn’t seem to be any commonly accepted term for this even though it must appear in the majority of web applications everywhere.</p><p>Enough. I sat and thought about it for a while, and from here on out I’m going to call it a <strong>ref</strong>. Here’s my reasoning:</p><ol><li>It’s short, just three characters. That makes it quick to type and also makes it representative of what’s desired: a short, URL-friendly name.</li><li>It’s at least semi-descriptive. In context, “ref” is obviously a short-hand for “reference.” The finer points (URL-friendly, etc.) may not be obvious but we’re at least in the same ballpark.</li><li>Other than Git/GitHub, I don’t know of wide usage of the term, which means that it shouldn&#39;t run into too many collisions with existing libraries/code.</li></ol><p>So why bother writing a post about this? Because the more people decide to agree with me, the more <strong>ref</strong> will become part of the common user vocabulary and the less it will require explanation. Naming will always be hard, but maybe this part of it can get a bit easier.</p><p><em>Don’t like </em><strong><em>ref</em></strong><em>? Have something better? </em><a href="http://twitter.com/mbleigh"><em>Let me know on Twitter</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fe8674e7873d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[‘Gravity’ is Sexist]]></title>
            <link>https://medium.com/i-m-h-o/gravity-is-sexist-bc089d643d34?source=rss-a6df102dcd62------2</link>
            <guid isPermaLink="false">https://medium.com/p/bc089d643d34</guid>
            <dc:creator><![CDATA[Michael Bleigh]]></dc:creator>
            <pubDate>Thu, 17 Oct 2013 17:49:20 GMT</pubDate>
            <atom:updated>2013-10-17T17:50:04.224Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/700/0*fjb-Xxxx9jh_njip.jpeg" /></figure><p><strong>Spoiler Alert:</strong> This post will ruin the closest thing to a ‘twist’ that Gravity has. It’s not a huge deal, but if you haven’t seen the movie and don’t want any plot spoilers, best to move along.</p><ol><li>Sandra Bullock’s character is the most panicky, inept person ever sent into space. It’s <a href="https://twitter.com/neiltyson/status/386986752914624513">never explained</a> why a medical doctor with only six months of training is somehow adding equipment to the Hubble, probably because there’s no world in which this would happen.<strong><br></strong>Science fiction movies have certainly had more panicky people in space (<a href="http://www.youtube.com/watch?v=dsx2vdn7gpY">Bill Paxton in Aliens?</a>) but Gravity isn’t a science fiction movie. It takes place in present times with present technology, in space. It’s realistic fiction, comparisons should be more to Apollo 13 than any sci-fi.</li><li>For the first half of the movie, Sandra Bullock basically has no agency and takes no useful action to save herself. George Clooney unflappably holds her hand through every action and swoops in with his magic jetpack whenever she inevitably does something wrong. She spends a good deal of time literally tethered to him as he does all the work.</li><li>By all appearances, George Clooney’s character would have survived the entire ordeal easily had he not been forced to frequently expend limited resources and take gigantic risks to save Sandra Bullock’s “damsel in distress.”</li><li>The only back-story for her character is how emotionally damaged she is because she is a mother who lost her child. There’s no screen time given to how awesome at her job she must be to have landed this incredibly rare and prestigious opportunity, nor is any given to showing her being more competent than anyone else at any task she performs in the movie.</li><li>There is a point in the movie where she has given up all hope and just decided to die. Rather than taking this moment for her to find the strength <em>within herself</em> to carry on, she instead hallucinates George Clooney back into the picture to be Space Daddy and tell her exactly what she needs to do.</li></ol><p>Whatever else you may think of Gravity, don’t think that it is doing a good job of portraying a strong female role. It’s the opposite. This is just one more story reinforcing “in catastrophic situations, men are cool as cucumbers and women freak out and get emotional” stereotypes.</p><p>Alfonso Cuarón may have <a href="http://blogs.indiewire.com/womenandhollywood/alfonso-cuaron-defends-having-female-lead-in-gravity">fought to have a female lead</a> in Gravity, but having a female lead and having an admirable female character are two totally different things.</p><p>And now for some pre-facto rebuttals:</p><ol><li><strong>The circumstances were unprecedented, any man or woman could have panicked in that way.</strong> That’s true, but not just anybody did. The reference points we are given in the movie are the cool, collected man and the panicky, inept woman.</li><li><strong>But ultimately she saves herself!</strong> Yes, but there’s both the problem of hallucinating a man to tell her what to do and that we are never shown a single thing that she does that seems like it is better or different from what our male character would have done, and he’s only dead because of her.</li><li><strong>Her inexperience was central to the tension of the plot, it had to be this way.</strong> I disagree. This isn’t Apollo 13 where we’re recounting some historical actual space disaster. This is a movie about everything in space blowing up. There would have been tension and doubt as to her survival even if she were exactly as cool and collected as George Clooney.</li><li><strong>So how should it have been, then?</strong> How about this: just swap the roles. Sandra Bullock (or someone maybe a bit older) plays the cool, collected experienced astronaut who is paired with a rookie astronaut on his first mission. This probably can’t be George Clooney because I can’t really imagine him playing a rookie or panicky anything. How about Bill Paxton?</li><li><strong>Who cares, it’s just a movie.</strong> A movie that is garnering universal critical acclaim, huge box office success, likely award nods, and is also talked up for having a female lead. We’re showering praise and success and reinforcing the idea that damsels in distress are a great way to tell stories.</li></ol><p>I believe that media and entertainment hold immense sway over the attitudes and behaviors of our culture, and I find it frustrating how few opportunities are seized to portray fantastic, strong female characters, especially in engineering or scientific roles. The world needs more Ellen Ripleys, not more Ryan Stones (that’s the character’s name, if you’ve forgotten. I had).</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bc089d643d34" width="1" height="1" alt=""><hr><p><a href="https://medium.com/i-m-h-o/gravity-is-sexist-bc089d643d34">‘Gravity’ is Sexist</a> was originally published in <a href="https://medium.com/i-m-h-o">I. M. H. O.</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Violent Video Games are Lazy]]></title>
            <link>https://medium.com/i-m-h-o/violent-video-games-are-lazy-cb55410aa22e?source=rss-a6df102dcd62------2</link>
            <guid isPermaLink="false">https://medium.com/p/cb55410aa22e</guid>
            <dc:creator><![CDATA[Michael Bleigh]]></dc:creator>
            <pubDate>Tue, 18 Jun 2013 23:54:33 GMT</pubDate>
            <atom:updated>2013-06-18T23:54:33.695Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/400/0*kgQsuUZ6hb2HR2tM.png" /></figure><p>I’m tired of violent video games. More specifically, I’m tired of combat being the core mechanic of the vast majority of AAA titles. We aren’t building arcade games in the 1980&#39;s anymore; we have the technology to create compelling narratives, believable characters, and we’ve only scratched the surface of what non-violent gameplay could be.</p><p>L.A. Noire was the first game that made me start thinking about combat as a lazy cheat. Gathering clues, interrogating suspects, it felt like a real detective simulation. And then, for some reason, I’d be dumped into a shootout in which I killed 10-20 people. <em>What?!?</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/620/0*iWtfD2GA23nHHFEz.jpeg" /></figure><p>It felt awkward. It felt tacked on. Most of all, it felt like the developers who had otherwise created innovative and compelling gameplay were afraid to ship a game that didn’t have a player-created body count.</p><p>More recently I felt the same frustration with Bioshock Infinite. The original Bioshock is one of my favorite games of all time, and the combat feels like a natural part of the dystopian life of post-Fontaine Rapture. Exploration would be punctuated by a splicer here and there. In Infinite, however, you feel a constant pressure to move forward without exploring. The combat is constant, long-range, and against large groups of opponents. The game was telling me that I shouldn’t take my time, shouldn’t explore, I needed to rush to the next big battle.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Pp8kZDVDsm1CvJK7.jpeg" /></figure><p>I would love to see Bioshock Infinite with exploration and investigation as the core mechanics. Instead of fighting everyone you see, imagine if your search for Elizabeth involved detective work, gathering clues and following them around a (more) open world Columbia. You’d have a chance to truly experience the twisted floating city without feeling like it’s nothing more than a big, beautiful set piece for standard FPS fare.</p><p>I’m not implying that every game should eschew combat. When it fits with the narrative, when it’s innovative, when it helps the game realize its vision, combat is great. What I am saying is that it would be great to flip the standard, for the shooters to become the outliers instead of the other way around. Games like L.A. Noire and Heavy Rain give us a glimpse of what non-combative games could look like, but we’ve only really scratched the surface.</p><p>If studios are going to spend tens or hundreds of millions of dollars developing AAA titles, why can’t that budget go towards funding game mechanics and experiences that can’t be achieved on a smaller scale? We have the technology to create immersive worlds, so why don’t we?</p><p>Of course, those are (somewhat) rhetorical questions. I know that the answer is that betting tens or hundreds of millions on unproven gameplay or unknown IP is riskier than retreading popular ground. I believe, however, that the gaming audience is ready and willing to explore uncharted territory if there are developers and publishers willing to make a bet on the unknown.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cb55410aa22e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/i-m-h-o/violent-video-games-are-lazy-cb55410aa22e">Violent Video Games are Lazy</a> was originally published in <a href="https://medium.com/i-m-h-o">I. M. H. O.</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Microsoft is Right(ish)]]></title>
            <link>https://medium.com/i-m-h-o/microsoft-is-right-ish-23195e01e628?source=rss-a6df102dcd62------2</link>
            <guid isPermaLink="false">https://medium.com/p/23195e01e628</guid>
            <dc:creator><![CDATA[Michael Bleigh]]></dc:creator>
            <pubDate>Thu, 13 Jun 2013 18:31:59 GMT</pubDate>
            <atom:updated>2013-06-13T18:31:59.119Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*AL3kvxiveyeFnhyb.jpeg" /><figcaption>Photo via Edge-Online.com</figcaption></figure><h4>Xbox One licensing policies aren’t crazy, they’re the future.</h4><p>Microsoft is taking a pummeling in the public sphere for their new game licensing scheme with the Xbox One.</p><blockquote>What do you mean I can’t just loan a disc to a friend? No rentals? This is madness! Persistent internet connection required? FFFFUUUUUUU!</blockquote><p>I’ve seen the rage, and it’s not 100% unjustified. The always-on requirement won’t be a problem until it’s a huge problem. We just have to look back a few months to <a href="http://kotaku.com/simcitys-creator-calls-simcity-launch-inexcusable-493225797">the launch of Sim City</a> to see why people are more than a little skeptical. Microsoft will need to have god-like uptime for Xbox Live, but they’re one of few companies that have the infrastructure knowledge to pull that off.</p><p>Let’s not talk about persistent connections, let’s talk about game licensing. Here Microsoft is actually making a brave and bold move that may do more to shape the future of gaming than anything else that is happening at E3.</p><p><strong>Physical copies of software won’t exist in ten years.</strong> This trend is obvious and doesn&#39;t really need to be justified, but you only have to look at the rise of the App Store, Steam, or any of a hundred other signals to see that it’s the case. But there are huge problems with our current systems of digital distribution: there’s no resale, no “loaning to a friend,” no rental, none of the post-purchase activities that have existed throughout all of gaming’s history.</p><p>You might think that game publishers are thrilled with this (and perhaps some are, I can’t speak for them). Publishers don’t see a dime of used game sales, so eliminating that market can only mean more money for them, right? Wrong. Oh, so very wrong.</p><p>Many gamers finance their purchase of new games by trading in their old games. They don’t see the price of a game as $60, they see it as $20 plus a couple trade-ins. If you take away their ability to sell used games, you’ll see a kind of sticker rage completely unlike the $10 bump when we entered the last generation.</p><p>Microsoft is the first big company that I’ve seen actually try to tackle this problem. Sure, the Xbox One licensing scheme isn’t perfect and there will be growing pains, but look at what they are accomplishing:</p><ol><li>The ability to share digital purchases with close family and friends without having to have some kind of password-shared account.</li><li>The ability for developers and publishers to get a piece of used game sales, making it a new market opportunity instead of a black hole.</li><li>A console where if you lose your game disc, you still have the game.</li></ol><p>There may be a number of these policies that don’t pan out or have to substantially change to work in a way that is friendly both to consumers and the game industry, but I for one at least appreciate that Microsoft was willing to take a risk and start thinking about some of the harder questions of our digital gaming future.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=23195e01e628" width="1" height="1" alt=""><hr><p><a href="https://medium.com/i-m-h-o/microsoft-is-right-ish-23195e01e628">Microsoft is Right(ish)</a> was originally published in <a href="https://medium.com/i-m-h-o">I. M. H. O.</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>