<?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 Joey Anuff on Medium]]></title>
        <description><![CDATA[Stories by Joey Anuff on Medium]]></description>
        <link>https://medium.com/@joeyanuff-33180?source=rss-19349a1d4a90------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*xCwH8N4nwjii3CDBOOMPoA.jpeg</url>
            <title>Stories by Joey Anuff on Medium</title>
            <link>https://medium.com/@joeyanuff-33180?source=rss-19349a1d4a90------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 21:20:03 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@joeyanuff-33180/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[Deconstructing Kirby: GPT-3.5 vs. GPT-4]]></title>
            <link>https://joeyanuff-33180.medium.com/gpt-3-5-vs-gpt-4-illustrated-by-jack-kirby-aa3e0e7e550c?source=rss-19349a1d4a90------2</link>
            <guid isPermaLink="false">https://medium.com/p/aa3e0e7e550c</guid>
            <category><![CDATA[gpt-3]]></category>
            <category><![CDATA[jack-kirby]]></category>
            <category><![CDATA[gpt-4]]></category>
            <category><![CDATA[midjourney]]></category>
            <category><![CDATA[openai]]></category>
            <dc:creator><![CDATA[Joey Anuff]]></dc:creator>
            <pubDate>Wed, 15 Mar 2023 07:03:37 GMT</pubDate>
            <atom:updated>2023-03-16T03:31:30.352Z</atom:updated>
            <content:encoded><![CDATA[<h4>OPENAI // GPT-3.5 // GPT-4 // BING AI // MIDJOURNEY // ADDITIVE PROMPTING</h4><h4>Design schemas from classic Kirby covers with ChatGPT and Bing AI.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-YmG7LIX7oXsiRuCGOzydw.png" /><figcaption><em>(Warning: no AI art — and especially no AI Kirby — was generated for this demo.)</em></figcaption></figure><p>This post describes a chatbot sequence I ran twice, the day before and the day after GPT-4 was released (ie. Monday and Tuesday.) The “Additive Prompting” hack it’s built around is clever enough to mention on its own, but the GPT-4 before-and-afters really earn it this screengrab tour.</p><figure><a href="https://twitter.com/LinusEkenstam/status/1635270385137770496"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bmKd-4IO5-k5vifSOZXCrw.png" /></a><figcaption>Using ChatGPT to make infinite image prompts for Midjourney. <a href="https://twitter.com/LinusEkenstam">Linus Ekenstam on Twitter</a>.</figcaption></figure><p><a href="https://twitter.com/LinusEkenstam/status/1635270385137770496">The demo I’d bookmarked</a> (by <a href="https://medium.com/u/d655740094fb">Linus Ekenstam</a>) is designed to generate ornate Midjourney prompts, but that wasn’t the part I was interested in. It was instead the LLM prompt concept itself — forcing elaborate texts by populating a categorical list into a table, and then back into plain text—that struck me as worth testing right away. (I’m a sucker for Markdown-table-generating prompts as of late.)</p><p>Running my madlibs through both Bing and OpenAI was illuminating enough that I’d screenshot everything on Monday, but with GPT-4 unexpectedly deploying to Bing yesterday morning, an extended comparison was suddenly mandatory. Here’s how all three interactions transpired, in order:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lB-UY79BlGf0Hccg_P0iMw.png" /><figcaption>ChatGPT on Mar 13th, using GPT-3.5</figcaption></figure><p>Instead of a “food photography” and “hamburger,” I went with “a classic comic book cover” and “Jack Kirby.” And OpenAI’s ChatGPT web interface, driven by Monday’s GPT-3.5 version, met most of my minimum standards for task completion: describing each Kirby cover in sufficient detail to suggest a vaguely reminiscent visualization (if one was inclined to screw around with Midjourney all day.)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jYV_m8tksGUghi35Yo975g.png" /><figcaption>Bing AI on Mar 13th, using GPT-3.5</figcaption></figure><p>Monday’s Bing AI immediately outwrote Monday’s ChatGPT in descriptive output, supplying per-character pose details for each character and specific notes on perspective, costumes, and typography. (None overly correctly.)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xV7T5v5oh6VOwHSJjVKymg.png" /><figcaption>Bing AI on Mar 14th, using GPT-4</figcaption></figure><p>And finally Tuesday’s next-generation Bing AI with GPT-4: richer detail and more footnotes, more cleanly executes transformations, but still hallucinates like Hunter S. Thompson. Which will continue to be more than enough for some tasks, and nowhere close to enough on others, I guess?</p><p>And I still wouldn’t paste this stuff into Midjourney or Stable Diffusion, not because of the nonsense descriptions but because gimme a break. Actually, if it were better articulating real design elements that’d be a lot more tempting to GPU, although I’m sure it’d still look like hell. Here’s the plain-text if you want to try:</p><pre>Draw a comic book cover for Fantastic Four #1 from 1961. The cover should <br>show the four members of the Fantastic Four using their powers to fight <br>a giant green monster with yellow eyes and horns that emerges from the <br>street. Mr. Fantastic should stretch his arm to grab the monster’s horn, <br>Invisible Girl should turn invisible and run away from the monster, Human <br>Torch should fly in the air and shoot flames at the monster, and Thing <br>should lift a car over his head and throw it at the monster. The title <br>should be in large yellow letters at the top of the cover and the captions <br>should be in red boxes at the bottom. The background should show a city<br>street with cars and people fleeing in panic. The cover should have a <br>vintage style and look like it was drawn by Jack Kirby.<br><br>Draw a comic book cover for The Incredible Hulk #1 from 1962. The cover <br>should show Hulk in his original gray color lifting a jeep over his head<br>while Rick Jones tries to calm him down. Hulk should look angry and <br>muscular while Rick Jones should look scared and helpless. The title <br>should be in green letters at the top of the cover and the captions <br>should be in yellow boxes at the bottom. The background should show a <br>desert base with tanks and helicopters. The cover should have a vintage <br>style and look like it was drawn by Jack Kirby.</pre><p>Anyways, when OpenAI and MSFT unveil their multimodal interfaces, this whole tabular approach to detail accumulation probably becomes obsolete, but until then I’m gonna keep it in mind: <strong>ask the LLM to itemize the key elements of some thing, and then have that LLM populate examples according to its own schema.</strong></p><figure><a href="https://twitter.com/nickfloats/status/1635116672054079488"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*O8czKeTBK1QHXZjC2J9DrQ.png" /></a><figcaption>ChatGPT working with Additive Prompting by <a href="https://twitter.com/nickfloats">@nickfloats on Twitter</a></figcaption></figure><p>It should be acknowledged that before Linus tweeted this prompt chain Monday, <a href="https://twitter.com/nickfloats/status/1635116679016644611">Nick St. Pierre had tweeted</a> a similar approach as “Additive Prompting” on Sunday (which I’d also bookmarked.) The major difference between their approaches is that Nick supplies his own hand-crafted schema, while Linus lets the LLM generate it.</p><pre>Please create a table that breaks down an interior architecture <br>photograph composition into the following key elements, where each of <br>these key elements is a column:<br><br>Composition, Camera Angle, Style, Room Type, Focal Point, Textures, <br>Detail, Color Palette, Brand, Lighting, Location, Time of Day, Mood, <br>Architecture<br><br>Fill the table with 10 rows of data, where:<br>Composition = &quot;Editorial Style photo&quot;<br>Room Type = &quot;Living Room&quot;</pre><p>Speaking of Nick’s great work, check out this <a href="https://twitter.com/nickfloats/status/1633342761356218368">Midjourney thread of his</a> from last week. Vivid stuff for so tiny a prompt, I was impressed! “Children are TERRIFIED of Balloons!!” should be a real poster campaign.</p><p><em>-Joey Anuff</em></p><figure><a href="https://twitter.com/nickfloats/status/1633342761356218368"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sVo0zjJSs1y3rot0cWy_Cw.png" /></a><figcaption>Propaganda Posters in Midjourney <a href="https://twitter.com/nickfloats/status/1633342761356218368">@nickfloats on Twitter</a></figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=aa3e0e7e550c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[An Important Reminder About Chat Interfaces In The Wake of ChatGPT]]></title>
            <link>https://joeyanuff-33180.medium.com/an-important-reminder-about-chat-interfaces-in-the-wake-of-chatgpt-5df21342d03d?source=rss-19349a1d4a90------2</link>
            <guid isPermaLink="false">https://medium.com/p/5df21342d03d</guid>
            <category><![CDATA[chatbots]]></category>
            <category><![CDATA[react]]></category>
            <category><![CDATA[openai]]></category>
            <category><![CDATA[google-sheets]]></category>
            <category><![CDATA[chatgpt]]></category>
            <dc:creator><![CDATA[Joey Anuff]]></dc:creator>
            <pubDate>Fri, 17 Feb 2023 01:25:18 GMT</pubDate>
            <atom:updated>2023-02-18T05:06:33.178Z</atom:updated>
            <content:encoded><![CDATA[<h4>OPENAI API // <a href="https://gptforwork.com/gpt-for-sheets/reference">GPT For Sheets</a> // PROMPT ENGINEERING // NEXT.JS 13 // <a href="https://vercel.com/templates/next.js/twitter-bio">Vercel Edge functions</a></h4><h4>Chatbots have a long history of uselessness.</h4><figure><a href="https://twitter.com/VarunJuice/status/1623062928042889216"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kSbHCjHgtwcLQ3HtGWHI3A.png" /></a></figure><p>I’m not trying to make a bold statement about web UX here, it’s just something I’ve had to remind myself of lately–how badly (but properly) maligned chatbots were until ten weeks ago.</p><figure><a href="https://twitter.com/levie/status/1622720794420985856"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Btr-2yqJkswNSNzZopz6vg.png" /></a></figure><p>It’s where I most quibble with the dismissal of ChatGPT by AI bigwigs as “nothing revolutionary”–in web UX terms at least, LLMs redeeming the chat interface is about as revolutionary as it gets.</p><figure><a href="https://twitter.com/ylecun/status/1620533783702433792"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qNGekcLjqR4x-JYCX2oEFw.png" /></a></figure><p>This is something I’ve reflected on while navigating the first crop of ChatGPT clone demos, a hot category in my January YouTube algorithm. Each new project keeps me wondering, will this project feature-match or feature-exceed OpenAI’s chatbot implementation? What even are ChatGPT’s most important features? <strong><em>How exactly does ChatGPT make chatbots tolerable?</em></strong></p><figure><a href="https://www.youtube.com/watch?v=JcE-1xzQTE0"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*CYEykUMcJN_XdsvQyWTfSg.gif" /></a><figcaption>Vercel Edge functions on the left, serverless functions on the right.</figcaption></figure><p>The first major UX upgrade I noticed about ChatGPT was its staccato text delivery: whatever ChatGPT was spewing, it was spewing in chunks. The earliest template I saw to get this right was @<a href="https://twitter.com/nutlope">nutlope</a>’s AI Twitter Bio demo, which more or less sold me on Next.js for LLM front-ends and Vercel Edge functions for LLM streaming (for the next few experiments at least.)</p><h3>Joey Anuff on Twitter: &quot;So far, @vercel&#39;s new TwitterBio template is easily the best full-stack ChatGPT starter I&#39;ve found, perfectly recreating that site&#39;s &quot;Typewriter&quot; effect (a crucial and under-appreciated UX element in ChatGPT&#39;s success, I suspect.) https://t.co/l5krG1sYYO / Twitter&quot;</h3><p>So far, @vercel&#39;s new TwitterBio template is easily the best full-stack ChatGPT starter I&#39;ve found, perfectly recreating that site&#39;s &quot;Typewriter&quot; effect (a crucial and under-appreciated UX element in ChatGPT&#39;s success, I suspect.) https://t.co/l5krG1sYYO</p><p>After tweeting the Vercel template, my old Web 1.0 comrade Greg Beato reached out suggesting a perfectly plausible customization. At some point he’d registered the domain name <a href="http://dearaibby.com/">dearaibby.com</a> and wondered if we could hijack the Vercel demo’s one-question interface to simulate an old-fashioned advice columnist.</p><figure><a href="https://github.com/januff/dear-aibby"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*crN9PSxrdpNXvAYq5ogffg.png" /></a></figure><p>With ease! Deploying an adapted prototype took maybe an hour, and applying Greg’s powerful purple design took even less. Thinking about the best wording for our prompt context took me a little bit longer.</p><figure><a href="http://patterned.ai/"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TUUF-_e2c_J0bpp8ENRXEA.png" /></a><figcaption>Greg will maybe find a nice AI-generated background pattern at <a href="http://patterned.ai/">patterned.ai</a></figcaption></figure><p>Greg’s prompt (see below, left) returned a convincing Dear Abby, but I couldn’t help feeling like we were ignoring an equally impersonable advisor: <a href="https://weeklyworldnews.com/advice/179502/dear-dotti/">Dear Dotti</a>. It occurred to me whatever editorial guidelines the Weekly World News originally devised for the column in the 1980s probably looked a lot like a GPT prompt, ie. what fits on a standard-sized post-it note.</p><figure><a href="https://gptforwork.com/gpt-for-sheets/reference"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fzgiqdpYuoACozGbjv4Tpw.png" /></a></figure><p><strong>INSULTING</strong>, <strong>QUICK-TEMPERED</strong>, <strong>EASILY DISTRACTED </strong>got us a very convincing Dotti<strong>, </strong>and the last two terms were probably overkill. (Like <strong>SLIGHTLY MAD</strong>, they foreshadow my plans to prompt GPT to footnote poetry in the scholarly style of Charles Kinbote. Stay tuned!) Ultimately Greg was quite taken with the Dotti simulacrum, he thinks she has a real knack for emojis. See for yourself:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*sOkNmyThXNfUmKu8yE8R1g.gif" /><figcaption>“You should have swallowed your pride and used a net.”</figcaption></figure><p>One noteworthy takeaway from this exercise was my discovery of Google Sheets as a surprisingly functional scratchpad for GPT prompts. The same day Google’s CEO was being panned for that Bard demo, I was being blown away by this free 3rd-party extension, GPT For Sheets.</p><figure><a href="https://workspace.google.com/marketplace/app/gpt_for_sheets_and_docs/677318054654"><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Dw_dHi7zCHt_mSPC.png" /></a></figure><p>Here’s an example: I’d suggested we give each correspondent an alliterative nickname like “Sleepless in Seattle.” If we didn’t mind prompting users for location permissions, it would be easy enough to grab latitude and longitude with the browser’s Geolocation device API.</p><p>Now, there are endless APIs that’ll reverse geocode GPS data, but what if we just let text-davinci-003 figure out the whole thing for us?</p><p>Perfecting that prompt is obviously quite doable in a Python notebook or OpenAI’s playground or a HuggingFace space, but constructing it over a series of cells in Sheets is just a different level of easy. (I mean, it auto-caches <em>and</em> auto-saves.)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/686/1*DRc0rdfYt6PLay3YHDAfeQ.png" /><figcaption>The question in cell A1 is implicit, I discover later.</figcaption></figure><p>With the city name in hand, prompting for an apt alliteration was as simple as asking:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/746/1*OoiQb6XFAr9T2COqBJ_bqA.png" /><figcaption>There’s also a <a href="https://gptforwork.com/gpt-for-sheets/reference#56565ecc34664cce81c1baccea4a61d8">GPT_CREATE_PROMPT helper function</a> to cleanly concatenate cell ranges.</figcaption></figure><p>It’a hardly a streamlined prompt yet, but it’s eye-opening to see how well the spreadsheet medium works for prompt engineering in general. To me, that counts as a somewhat big deal. Google Sheets as a viable prompt IDE feels like more than a cool hack–the workflow may well become a new habit.</p><h4>TRY FOR YOURSELF</h4><ul><li><a href="https://workspace.google.com/marketplace/app/gpt_for_sheets_and_docs/677318054654">Install GPT for Sheets and Docs</a> on Chrome on Google’s Workspace Marketplace. All the functions work intuitively, but I’ve found the <a href="https://gptforwork.com/gpt-for-sheets/reference">individual explanations</a> useful to walk through alongside an open sheet. Give creator <a href="https://twitter.com/stanmarion">Stanislas Marion</a> a follow on twitter.</li><li>Along with the <a href="https://vercel.com/templates/next.js/twitter-bio">AI Twitter Bio template</a>, Vercel Developer Advocate Hassan El Mghari has <a href="https://vercel.com/blog/gpt-3-app-next-js-vercel-edge-functions">blogged the code</a> and YouTubed a short <a href="https://www.youtube.com/watch?v=9Q9_CQxFUKY">explainer vid</a>. The Edge vs. Serverless parts of that video (which I excerpt in the animated results gif above) and the speed-test in particular (<a href="https://youtu.be/9Q9_CQxFUKY?t=245">4:05–4:40</a>) are jaw-droppers.</li><li>My modifications to Hassan’s code for dearaibby.com are <a href="https://github.com/januff/dear-aibby">on GitHub</a>. Greg’s demo is currently alive and kicking at <a href="http://dearaibby.com/">dearaibby.com</a>. (Although it’ll hiccup once certain API limits are reached.)</li><li>React writers interested in AI demos should give a look to Vercel’s recent <a href="https://vercel.com/blog/deploying-ai-applications">focus on the AI category</a>, their developer advocates have been releasing a barrage of useful AI-driven templates this winter.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*ZIntRj5ity587fkBHJ_y_Q.png" /><figcaption>An excellent point by Vercel’s CEO!</figcaption></figure><p><strong><em>Joey Anuff </em></strong><em>is a first-wave web hacker, former TV developer, and current ML developer. For 2023, he’s trying to </em><a href="https://twitter.com/JoeyAnuff"><em>tweet more</em></a><em>. Currently on the lookout for interesting DevRel work.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5df21342d03d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AWS Lambdas: Easy, Easier, Easiest]]></title>
            <link>https://joeyanuff-33180.medium.com/aws-lambdas-easy-easier-easiest-823e71c02bd9?source=rss-19349a1d4a90------2</link>
            <guid isPermaLink="false">https://medium.com/p/823e71c02bd9</guid>
            <category><![CDATA[serverless]]></category>
            <category><![CDATA[nosql]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[react]]></category>
            <category><![CDATA[aws-lambda]]></category>
            <dc:creator><![CDATA[Joey Anuff]]></dc:creator>
            <pubDate>Fri, 30 Apr 2021 19:21:10 GMT</pubDate>
            <atom:updated>2021-04-30T19:21:10.737Z</atom:updated>
            <content:encoded><![CDATA[<h4>AMPLIFY // NETLIFY // VERCEL // DATASTAX</h4><h4>These serverless functions practically write themselves.</h4><h4>Step-by-step serverless setup using Amplify, Netlify, or Vercel, connecting to a free Apache Cassandra-as-a-Service database from DataStax.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EXJxZOZlQxWGxbo3CXovVg.png" /><figcaption><a href="https://serverlesshandbook.dev/"><strong>Serverless Handbook for Frontend Engineers</strong></a><strong> by Swizec Teller (Feb 2021)</strong></figcaption></figure><p>As Swizec Teller warns in the Serverless Handbook, AWS can be a pain. My own frustrations with the AWS dashboard–as a user but especially as a developer advocate–led me away from using AWS Lambdas for the OCR function in my comic translation app, as I explained late last year:</p><figure><a href="https://joeyanuff-33180.medium.com/6-tricks-for-simpler-cloud-cv-1036b99ac791"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*v3p0ccNz-j9Xx-totk40Fw.png" /></a><figcaption><strong>And it’s still true that your app might not need Lambdas.</strong></figcaption></figure><p>Also true, though: even if your app <em>does</em> need a serverless function–and even if that serverless function was required to be an official AWS Lambda–you wouldn’t need the AWS dashboard, or an AWS account, to get it up and running. Two of the most popular Jamstack hosting platforms, Netlify and Vercel, offer idiot-proof wrappers for AWS Lambda deployments, each more developer-friendly than the next.</p><p>Meanwhile, AWS is hardly neglecting their own developer ergonomics. Their Amplify CLI goes far towards relieving many prior headaches, Cognito management in particular. Lambda set-up with Amplify, they rightly brag, takes only two commands.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*deqcU0b1tDnKqG3VaZzeFQ.png" /></figure><p>All of which presents 2021’s front-end developer with a classic high-quality problem: three distinct but compelling APIs for “real AWS Lambda functions”–Amplify’s, Netlify’s, and Vercel’s–representing three competing visions of simplicity.</p><p>Having reached the point in my app’s development where I <em>do </em>need a Lambda–to chat with the Apache Cassandra database where I’ll be storing my translations–it seemed worthwhile to feature-compare my walkthroughs of the three APIs.</p><h4>AMPLIFY vs. NETLIFY vs. VERCEL: WHICH IS THE EASIEST?</h4><figure><a href="https://github.com/januff/easy-easier-easiest-lambdas/tree/master/amplify-with-create-react-app"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tsRd_7a4H8MXnadWXdZ3Zg.png" /></a></figure><p>Putting the platforms through the same paces, I hoped, would reveal a clear favorite, or at least a clear loser–but they’re all three pretty great tools, from teams clearly focused on state-of-the-art DX. There’s some variety amongst documentation style, dev servers, and file structures, but given my previous struggles with Lambdas, far fewer gotchas than I’d expected.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/d4ba935e1710ab6dd95e5f58f6d67567/href">https://medium.com/media/d4ba935e1710ab6dd95e5f58f6d67567/href</a></iframe><p>Lined up next to one another, my crib-sheet above confirms my perception of Vercel having the fewest commands to remember. This is slightly deceptive, in that Next.js already bundles in a dev server (otherwise requiring <em>vercel dev</em> and <em>vercel deploy</em> steps.) More generally, I’d break down my most deciding factors for each platform as follows:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/3841ae48cb4183c378d53f04176158b8/href">https://medium.com/media/3841ae48cb4183c378d53f04176158b8/href</a></iframe><p><strong>Long story short: since I’m already using Vercel’s Next.js framework for my app–and Next.js gives you a free serverless API out of the box–I intend to use Vercel, but it’s a close call. (And by no means final: I started my OCR project with Python Lambdas on the AWS Dashboard and I’m certain to try other languages and APIs before I’m done.)</strong></p><p>The linked repo collects step-by-step deployment instructions for all three platforms. Where the steps occur on a web dashboard or a CLI, I’ve included screencast GIFs to illustrate the interaction.</p><figure><a href="https://github.com/januff/easy-easier-easiest-lambdas"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Xib2m2Id5G8B9DXNFUWSew.png" /></a></figure><p>Taken individually, none of these steps seem too challenging to remember, but I’m always amazed how quickly the details slip away. Stack Overflow fixes and key doc pages especially–which I either bookmark and forget or keep open as Chrome tabs for decades.</p><p><a href="https://github.com/januff/easy-easier-easiest-lambdas">GitHub - januff/easy-easier-easiest-lambdas: Quickstarts for AWS Lambda (&quot;serverless&quot;) functions in React, with and without AWS tooling. Functions connect to a serverless Apache Cassandra database.</a></p><p>Getting those out of my bookmarks and into a repo feels good, but if I can help someone get a serverless function running in a day instead of a week, I’ll feel even better, so please let me know!</p><p><strong>UP NEXT: <em>A deployed translation app to test and clone.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=823e71c02bd9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Multiverse Next Door]]></title>
            <link>https://joeyanuff-33180.medium.com/the-multiverse-next-door-444477bdaf63?source=rss-19349a1d4a90------2</link>
            <guid isPermaLink="false">https://medium.com/p/444477bdaf63</guid>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[comics]]></category>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Joey Anuff]]></dc:creator>
            <pubDate>Tue, 02 Feb 2021 19:00:39 GMT</pubDate>
            <atom:updated>2021-02-02T19:00:39.866Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*u6T4nEOg_F2k3S4MWT10GQ.png" /></figure><h4>NEXT.JS 10 // AWS AMPLIFY CLI // DATASTAX STARGATE APIs</h4><h4>Digging the global comics underground with cloud APIs &amp; JavaScript</h4><p>I’ve deployed some personally exciting ML-powered demos recently, but until now nothing personally essential: an app to help translate these growing piles of bootleg comic books.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*5YzWRadNBT9sQKkv.png" /></figure><p>The bootlegs are almost as new to me as the Machine Learning. Like ML, thrifting international flea-marketplaces (and Facebook Groups) for freak knock-offs is only recently an actually doable thing–in fact, my first big score wasn’t until late 2018:</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*HB4_CJe_ZlOIkPQFv_xuXw.png" /><figcaption>Argentinean redraw (1960) of <strong><em>Tales From The Crypt</em></strong><em> #30 (Jun 1952)</em></figcaption></figure><p>Frustratingly, the internationalization (“i18n”) improvements that make multilingual comics-buying an addicting pastime have yet to make multilingual comics-<em>reading</em> any easier.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wRHfvX_dRVgIPWRfiS6ncA.png" /></figure><p>I keep expecting Kindle or Google Drive to add some magic “scanlation” button to perform the kind of typeset OCR I need–as we’ll see, their existing APIs make this task fairly straightforward, if not trivial–but I’ve still yet to see any such thing.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*CCzuXiux-efDQBb0oMCLrg.png" /></figure><p>A reluctance to antagonize publishers? Maybe. But then again, nobody knew world comics had so much untranslated corpora. Surely the discovery of an international brotherhood of Dr. Doom clones changes the equation somehow?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QBALRKI5fEJt2zhyuUEuOQ.png" /></figure><p>This project thus has dual goals: to introduce some crazy new archaeology to comics fandom, and to scaffold a simple tool for deciphering it, narrowly designed for personal, non-commercial use.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jJIb31dm9nnz9-ZE-xyTog.png" /></figure><h4><strong>What’s the stack?</strong></h4><p>A minimalist React app, built using the Next.js 10 web framework and chiefly dependent on two powerfully simple APIs: <a href="https://docs.amplify.aws/">Amazon’s Amplify CLI</a>, which we’ll use to access several cloud ML endpoints, and <a href="https://stargate.io/">DataStax’s Stargate APIs</a> for Apache Cassandra, the database we’ll use to store and edit our private translation libraries.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*t_CYdHkWDXbDU8NstobaoA.png" /></figure><h4>Who is this for?</h4><p>Language learners and insatiable comic readers, eventually; comic collectors and historians, who’ll enjoy the upcoming galleries and data dumps; but right now mainly developers, particularly those eager to squeeze Machine Learning functions into their full-stack web apps.</p><p>Parsing a comic book page is a challenge perfectly suited to modern ML algorithms–and there’s a lot of great new research to draw from, some with <a href="https://github.com/DRDRD18/balloons">ready code</a> like this paper:</p><figure><a href="https://arxiv.org/abs/1902.08137"><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*vE7b7GfDTEGldKRn.jpeg" /></a><figcaption><a href="https://arxiv.org/abs/1902.08137">Deep CNN-Based Speech Balloon Detection and Segmentation for Comic Books</a></figcaption></figure><p>The first and most basic ML function of this project–handwriting recognition–is one of the earliest Computer Vision tasks, one of the first things taught in CV courses, and was the subject of the first how-to of this series, <a href="https://joeyanuff-33180.medium.com/6-tricks-for-simpler-cloud-cv-1036b99ac791">6 Tricks for Simpler Cloud CV</a>:</p><figure><a href="https://joeyanuff-33180.medium.com/6-tricks-for-simpler-cloud-cv-1036b99ac791"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*70wXMJAMmNGm-X9b9fxrqw.png" /></a></figure><p>Discovering how much easier Amazon’s newer Amplify CLI was for deploying an ML endpoint (compared with Amazon’s older Lambda web console) was the major breakthrough I wanted to share in that essay.</p><h4>What’s next?</h4><p>The next installment will constitute another dramatic API upgrade, this time to our project’s database: Apache Cassandra, the database used by Apple, Spotify, Netflix and eBay. As with our OCR, we’ll be setting up some extremely concise serverless CRUD functions with the help of DataStax’s new <a href="https://github.com/kidrecursive/astrajs/tree/master/packages/collections">@astrajs/collections package</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*OkR4KTZ8ws-TO4lVL8wOBQ.png" /></figure><p>The CV and NLP models we’ll be using are often better than 90% accurate, but whatever error remains require human correction. Creating solid UX for revising our ML-annotated data–while at the same time optimizing our storage for continuous training and data analysis–is the kind of challenge Apache Cassandra is built for.</p><p>With input from the team at DataStax, I’m confident I’ll soon be explaining <em>a</em> solution, if not <em>the </em>solution, to many such database concerns common to ML-driven web apps. Some specific areas I’ll be looking into include:</p><ul><li>Using the Stargate Javascript SDK in SSR and SSG situations.</li><li>Optimizing an Apache Cassandra db for NLP analysis.</li><li>The easiest approach to non-destructive edits.</li><li>Setting up a private system for collaborative translation.</li><li>Quick-starting a pop-up store with Next.js 10 Commerce.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TAKW0nYr-QwXScokAZbZyA.png" /></figure><p><strong><em>Joey Anuff</em></strong><em> is a full-stack web developer, focused on ML applications in the comic art domain. Front-end experiments, also comics-related: my </em><a href="https://metaguru.guru/"><strong><em>2018 portfolio</em></strong></a><em> or my</em><strong><em> </em></strong><a href="https://codepen.io/januff"><strong><em>CodePen demos</em></strong></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=444477bdaf63" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[6 Tricks for Simpler Cloud CV]]></title>
            <link>https://joeyanuff-33180.medium.com/6-tricks-for-simpler-cloud-cv-1036b99ac791?source=rss-19349a1d4a90------2</link>
            <guid isPermaLink="false">https://medium.com/p/1036b99ac791</guid>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[nextjs]]></category>
            <category><![CDATA[aws-lambda]]></category>
            <category><![CDATA[computer-vision]]></category>
            <dc:creator><![CDATA[Joey Anuff]]></dc:creator>
            <pubDate>Wed, 09 Dec 2020 17:10:07 GMT</pubDate>
            <atom:updated>2021-01-30T01:19:50.907Z</atom:updated>
            <content:encoded><![CDATA[<h4><em>NEXT.JS 10 // AWS AMPLIFY CLI // DATASTAX DOCUMENT API</em></h4><h4>(or How Not To Build A Comic Book Parser, Pt. 1)</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1004/1*dfMVVzlHEnIBRPGKxI1RRA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/984/1*EyRN_L-82uQBWujPbf1yYg.png" /></figure><p><em>To follow along with this demo, initialize a</em><strong><em> </em></strong><a href="https://astra.datastax.com/register"><strong><em>free Cassandra database at DataStax</em></strong></a><em>, whose super-simple new </em><a href="https://www.datastax.com/dev/documents-api"><strong><em>Documents API</em></strong></a><em> we’ll be using for our backend. Make note of your username, password, cluster/database ID, region, and namespace. Init a </em><a href="https://nextjs.org/docs/api-reference/create-next-app"><strong><em>new Next.js app</em></strong></a><em> and install these extra dependencies:</em></p><pre>npm install aws-amplify axios axios-auth-refresh react-query react-query-devtools react-rough roughjs sass</pre><p>It was meant to be irresistibly simple: a quick ML app to upload comics to the cloud, perform fancy Computer Vision operations on them (OCR and translation), and correct the transcripts on a personal cloud database.</p><p>And it was simple! Well, the code was simple–my two AWS Lambda functions were maybe 10 lines each. Properly configuring permissions for the associated AWS services, though, was another story–I spent two days just trying to log my Lambda console–but mercifully, a story with which I need no longer bore you.</p><h4>1. YOU MIGHT NOT NEED THE AWS CONSOLE</h4><p>In my quest to outsource the AWS set-up nightmare, I found something absurdly better: Amplify CLI, a straight-forward command line tool that gently steps you through your AWS cloud service inits.</p><p>Consider, just for a few seconds, this unholy mess of a config file:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/a290db382c3268b7fbbdf51c2733eaf0/href">https://medium.com/media/a290db382c3268b7fbbdf51c2733eaf0/href</a></iframe><p>Amplify takes care of all of this for us. It even says “<em>WARNING: DO NOT EDIT</em>” on the config file, an order I couldn’t be happier to follow.</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*hhoRejvTTbAmm3JsTFIFog.png" /><figcaption>A full log of my project inits is in the Github repo.</figcaption></figure><h4>2. YOU MIGHT NOT NEED LAMBDAS</h4><p>A simpler Lambda set-up is what led me to the Amplify CLI in the first place. Instead of zipping all your dependencies into “Lambda Layers” and coding inside Amazon’s web dashboard, Amplify lets you code locally and commit to AWS when you’re ready. As it turned out, using Amplify meant I didn’t need Lambda functions at all anymore.</p><p>My prototype app used two custom Lambda endpoints, configured using Amazon’s Gateway API: a Node.js function to save images to S3 and a Python function to call Amazon’s Textract OCR service. With Amplify, both of these are one-liners.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/369dec440844426598a8163337148dee/href">https://medium.com/media/369dec440844426598a8163337148dee/href</a></iframe><h4>3. YOU MIGHT NOT NEED BASE64</h4><p>My original Uploader Lambda required base64-formatted images. On the client side, I needed a base64 component to do the same thing. And still I wound up fiddling with the string at both ends. Brilliantly, Amplify’s Storage object takes care of both conversions for us.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/21fd9c56db274b30fca90711b526582f/href">https://medium.com/media/21fd9c56db274b30fca90711b526582f/href</a></iframe><h4>4. DON’T USE LAST YEAR’S APIS</h4><p>My big note-to-self for all of this, I suppose, is <strong>don’t use old APIs</strong>–and when it comes to cloud infrastructure, last year’s APIs are probably too old.</p><p>It’s the same situation with DataStax database I’m using to store my transcripts–with their new Stargate APIs, I can access my scans as REST, GraphQL, or plain JSON, as I do here via their Document API:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/16c4c08a5992475503fc61d034a30e13/href">https://medium.com/media/16c4c08a5992475503fc61d034a30e13/href</a></iframe><h4>5. DISABLE SSR FOR CLIENT-SIDE LIBRARIES</h4><p>It makes perfect sense that you client-side graphics library wouldn’t render on the server side, but understanding the nature of the error (and the fix) took me longer than I’d like to admit.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zjIXDLfrkUtXfd4OxysMtw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/930/1*n6nGBgDaAjJciKbCbRzkPw.png" /></figure><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/184a95723ff12f4aae022de70d38c973/href">https://medium.com/media/184a95723ff12f4aae022de70d38c973/href</a></iframe><h4>6. MIND THIS NEXT.JS vs. AMPLIFY GOTCHA</h4><p>Next.js 10 has a surprising ability to reflect style and code changes without triggering a browser reload, but sometimes a hard reload is unavoidable. Doing that a few times, I’ve found, is liable to cough up this show-stopping error from Amplify:</p><blockquote><strong>Error: Pluggable with name AmazonAIPredictionsProvider has already been added.</strong></blockquote><p>Until they update their Predictions plug-in to handle SSR–as they’ve already done with their Storage plug-in–here’s my quickest fix: just comment out the two lines as below and save.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/91f1a0eda21d10266e048f78b3418725/href">https://medium.com/media/91f1a0eda21d10266e048f78b3418725/href</a></iframe><p>This will instantly trigger a new error:</p><blockquote><strong>Error: More than one or no providers are configured, Either specify a provider name or configure exactly one provider</strong></blockquote><p>Which is your sign to uncomment the lines, reregistering the plug-in and allowing you to proceed as normal.</p><h4>RECOMMENDED READING</h4><ul><li><a href="https://dev.to/aws/full-stack-machine-learning-on-aws-reading-text-from-an-image-with-aws-amplify-amazon-rekognition-49jf">Full Stack Machine learning on AWS: Reading Text from an Image with AWS Amplify &amp; Amazon Rekognition</a></li><li><a href="https://welearncode.com/get-started-s3/">Build a Music Sharing App with Amazon S3 and AWS Amplify</a></li></ul><h4>PROJECT REPO</h4><p><a href="https://github.com/januff/nextjs-amplify-comics">januff/nextjs-amplify-comics</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1036b99ac791" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[6 Tricks for Next-Level CRUDs]]></title>
            <link>https://joeyanuff-33180.medium.com/6-tricks-for-next-level-cruds-9a916d9e4c87?source=rss-19349a1d4a90------2</link>
            <guid isPermaLink="false">https://medium.com/p/9a916d9e4c87</guid>
            <category><![CDATA[react-query]]></category>
            <category><![CDATA[cassandra]]></category>
            <category><![CDATA[nextjs]]></category>
            <category><![CDATA[react-hook]]></category>
            <category><![CDATA[graphql]]></category>
            <dc:creator><![CDATA[Joey Anuff]]></dc:creator>
            <pubDate>Wed, 09 Dec 2020 13:02:26 GMT</pubDate>
            <atom:updated>2020-12-15T19:48:40.268Z</atom:updated>
            <content:encoded><![CDATA[<h4>NEXT.JS 10 // REACT-QUERY // AXIOS // DATASTAX DOCUMENT API</h4><h4>Cache more, fetch less with React Query</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MdWmU7eDcgmbj9E0I_VHLQ.png" /></figure><p><em>To follow along with this demo, initialize a</em><strong><em> </em></strong><a href="https://astra.datastax.com/register"><strong><em>free Cassandra database at DataStax</em></strong></a><em>, whose super-simple new </em><a href="https://www.datastax.com/dev/documents-api"><strong><em>Documents API</em></strong></a><em> we’ll be using for our backend. Make note of your username, password, cluster/database ID, region, and namespace. Init a </em><a href="https://nextjs.org/docs/api-reference/create-next-app"><strong><em>new Next.js app</em></strong></a><em> and install these extra dependencies:</em></p><pre>npm install axios axios-auth-refresh react-query react-query-devtools sass</pre><p>I was looking for a clean, reusable React hooks-based approach for all my endpoints, something lighter than the full Apollo Client suite–maybe just a pattern or two, some smart but minimal approach to all my JSON traffic.</p><h4>1. CACHE YOUR DATA WITH A REVALIDATION LIBRARY</h4><p>React Query more than fit the bill: a low-carb request wrapper that follows the “stale-while-revalidate” approach to data fetching. Vercel’s similar SWR hook library also looked like a good option (Vercel being the creators of the Next.js framework) but React Query author Tanner Linsley’s polished docs and devtools made his library an almost irresistible choice.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*NqljnV2XgUTBosF7.png" /><figcaption>Tanner’s <a href="https://react-query.tanstack.com/docs/comparison"><strong>Comparison Table</strong></a> is helpful. See also: <a href="https://labs.thisdot.co/blog/why-you-should-use-react-query-or-swr"><strong>Why You Should Use React Query or SWR</strong></a></figcaption></figure><h4>2. SET YOUR DEFAULTS WITH AXIOS</h4><p>Different APIs are finicky about different headers. Any day I can DRY up my boilerplate with some global defaults is a great day.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/cc6a05d670ae9f04f96b4b19e5ba5874/href">https://medium.com/media/cc6a05d670ae9f04f96b4b19e5ba5874/href</a></iframe><h4>3. REFRESH TOKENS WITH AXIOS-AUTH-REFRESH</h4><p>Authentication panic is common, what’s really refreshing is accepting help, as I did when I installed this calming token authenticator.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/b2fe039b79a9216fd75f1238e514d29b/href">https://medium.com/media/b2fe039b79a9216fd75f1238e514d29b/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Crvez4PU9qdDyEgv.png" /><figcaption>The token in local storage is just for debugging.</figcaption></figure><h4>4. PRE-LOAD YOUR QUERY CACHE (WITHOUT PRE-FETCHING)</h4><p>My favorite React Query trick: instead of requesting data for every detail view, React Query builds a cache from your first paginated GET.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/802e5b62a6439731dd1f9097a0b08bb1/href">https://medium.com/media/802e5b62a6439731dd1f9097a0b08bb1/href</a></iframe><h4>5. EXPORT SELF-DESCRIPTIVE HOOKS</h4><p>All our CRUD operations are now simple and exportable. React-Query lets us specify our cache policy during and after each query or mutation.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/ef9018360911b4663696fd9aca9336b5/href">https://medium.com/media/ef9018360911b4663696fd9aca9336b5/href</a></iframe><h4>6. SYNC EVERYTHING WITH YOUR HOOKS</h4><p>Our Items index now runs with just two hooks, an items query hook and an addItem mutation hook. Each comes with a full menu of lifecycle booleans.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/e80119dbd823c6c49ca3129cff4161e7/href">https://medium.com/media/e80119dbd823c6c49ca3129cff4161e7/href</a></iframe><p>Changes on the detail page are reflected instantly–and reverted if they fail–using the Optimistic Update strategy.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/f87c1ec03dfdf53721e1a9fc377e629c/href">https://medium.com/media/f87c1ec03dfdf53721e1a9fc377e629c/href</a></iframe><p>With React Query’s devtools drawer open, we have perfect visibility into the state of our cache.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J3lkJShPQM61Dzhy1YLDJA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aN-nQgZpiwbHFsUn8JJ2fw.png" /></figure><h4>RECOMMENDED READING</h4><ul><li><a href="https://swizec.com/blog/how-react-query-gives-you-almost-everything-you-thought-you-needed-graphql-for">How React Query gives you almost everything you thought you needed GraphQL for | Swizec Teller</a></li><li><a href="https://labs.thisdot.co/blog/why-you-should-use-react-query-or-swr">This Dot Labs: Why You Should Use React Query or SWR</a></li></ul><h4>PROJECT REPO</h4><p><a href="https://github.com/januff/nextjs-datastax-crud">https://github.com/januff/nextjs-datastax-crud</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9a916d9e4c87" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Have You Seen These Spider-Men?]]></title>
            <link>https://joeyanuff-33180.medium.com/have-you-seen-these-spider-men-fdc92418e104?source=rss-19349a1d4a90------2</link>
            <guid isPermaLink="false">https://medium.com/p/fdc92418e104</guid>
            <category><![CDATA[comic-books]]></category>
            <category><![CDATA[bootleg]]></category>
            <category><![CDATA[komik-indonesia]]></category>
            <category><![CDATA[spider-man]]></category>
            <category><![CDATA[web-animation]]></category>
            <dc:creator><![CDATA[Joey Anuff]]></dc:creator>
            <pubDate>Fri, 01 May 2020 16:48:24 GMT</pubDate>
            <atom:updated>2020-05-11T02:24:48.662Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong>Origins of the Spider-Verse: Indonesia, 1969–1979</strong></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ix-LyYtYNq-oMHMF0tuJ2w.png" /></figure><p>You may have heard of Mexico’s strange, <a href="https://io9.gizmodo.com/idw-s-president-wants-to-unearth-spider-man-s-hidden-me-1836223985">alt-canonical <em>Hombre Araña</em> comics of the 1970s</a>, but I’ll bet you didn’t know that, during the very same decade, Indonesian comic stalls were packed with off-model Spider-Clones of their own. Take a good look at the line-up below. Do any of these two dozen Spider-artist pairings look familiar?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/874/1*Dh3VPad1P7yzg4uDxlFEJQ.png" /></figure><p>Not super-likely! Unless you came of age in 1970s Jakarta, Indonesia’s superhero alt-canon was never so much as a rumor to Western fandom.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*Mix4qu2ue6-iUCnrtNxXOA.png" /></figure><p>I come from the ’70s, but not from Jakarta. So it wasn’t until very recently that I noticed the Indonesia-shaped crack in the Marvel multiverse–mid-2018 to be exact, when I first hearted the comic book page below on a Captain America Tumblr.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tz1bm32Vg46crAuQOIEY5A.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4FaJaXlmn7qv7_kFyYyH0w.jpeg" /></figure><p>A less mindful Marvelite might blink and miss it, but the splash page pictured above does not, in fact, faithfully reproduce Jim Steranko’s iconic story-opener from Captain America #110. A side-by-side comparison illuminates what this Steranko nerd grasped in an instant:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/606/0*1AVMpnD1gdZwYdKo.jpg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/822/1*jB9gVQsncLBB3aW0xHIpJA.png" /><figcaption>Captain America #110 (Marvel Comics, Feb 1969) // Serial Kapten Amerika (Sanggar Karya, Feb 1979)</figcaption></figure><p>As you can see, the Indonesian edition is neither a photo-stat nor a tracing but a freehand recreation, not just translated but redrawn entirely.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/650/1*CnnenqaJEpsD8A-S1-AbzQ.png" /></figure><p>Such a “dirty Steranko” would’ve shook up my want-list any year of my life, but in 2018 it busted in like the Kool-Aid Man.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fgiphy.com%2Fembed%2FcdME6kF1p5Mw0LMG5t%2Ftwitter%2Fiframe&amp;display_name=Giphy&amp;url=https%3A%2F%2Fgiphy.com%2Fgifs%2FcdME6kF1p5Mw0LMG5t&amp;image=https%3A%2F%2Fmedia.giphy.com%2Fmedia%2FcdME6kF1p5Mw0LMG5t%2Fgiphy-downsized-large.gif&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=giphy" width="435" height="341" frameborder="0" scrolling="no"><a href="https://medium.com/media/8366bc95e9b2a84b9472ce4a60ab033c/href">https://medium.com/media/8366bc95e9b2a84b9472ce4a60ab033c/href</a></iframe><p>Ever the playful front-end developer, just right then I also was taking liberties with Steranko’s late-’60s layouts, which I’d been tickled to discover existed in wide contemporary translation.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fcodepen.io%2Fjanuff%2Fembed%2Fpreview%2FLJJNjj%3Fheight%3D600%26slug-hash%3DLJJNjj%26default-tabs%3Dhtml%2Cresult%26host%3Dhttps%3A%2F%2Fcodepen.io&amp;display_name=CodePen&amp;url=https%3A%2F%2Fcodepen.io%2Fjanuff%2Ffull%2FLJJNjj&amp;image=https%3A%2F%2Fs3-us-west-2.amazonaws.com%2Fcodepen-thumbnails%2FPen%2F28109744%2Fthumbnail.small.jpg%3Fv%3D1584369189&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=codepen" width="800" height="600" frameborder="0" scrolling="no"><a href="https://medium.com/media/5d4e26a25ece1a11547f542f078c446b/href">https://medium.com/media/5d4e26a25ece1a11547f542f078c446b/href</a></iframe><p>Alas, by the time I managed to track down my own scanning copy of <em>Serial Kapten Amerika</em>, I wasn’t really obsessing over MorphSVG-driven transitions anymore.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_H1MJfGRLLcQwsuCKETULA.png" /></figure><p>But my quest for <em>Komplotan Hydra</em> opened my eyes to the bigger picture. I discovered that there’s more to Bronze Age Indonesian comic history than freehand Steranko clones–there’s also Spider-clones.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/854/1*MBM7yjdh1C09YowW-NI6yw.png" /></figure><p>Of course, Indonesian comics go much deeper than fake Spider-Men–or fake anything–but there’s only so much one can cram into a single Medium article.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/864/1*nT-FHjAJQuD0MhSJcUnIkA.png" /></figure><p>So be patient while I get another spreadsheet ready to share–or be impatient, drill into the links below and commence digging your own cross-dimensional treasure.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PMGNRRJyXkMejtAVQbxcHw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0TZ5YafevoEJyPP1eRBHtg.png" /></figure><h4><strong><em>FURTHER READING</em></strong></h4><p>Essential overview of “apocryphal” Marvel history, published by Italian comics scholars in early 2020, perfectly intelligible in auto-translation:</p><p><a href="https://www.fumettologica.it/2019/03/fumetti-marvel-apocrifi/5/">Storia dei fumetti Marvel non prodotti dalla Marvel - Pagina 5 di 8 - Fumettologica</a></p><p>My first major exposure to Indonesian comics was the Flickr pages of Komic Kazi, a father-son team from Brazil who really know how to collect international comics. Skye Ott, the son, spent a year on the ground gathering the set presented here:</p><p><a href="https://www.flickr.com/photos/110108265@N07/albums/72157673486752796/">Indonesian Comics</a></p><p>Ott also amassed a preposterous trove of original Indonesian comic art, with exhibition-ready overviews of the industry’s most iconic talents:</p><p><a href="https://www.flickr.com/photos/110108265@N07/albums/72157654140874895">International Published Art</a></p><p>Indonesian comic collectors took to Facebook Groups for Buy/Sell/Trade quicker than many regional subcultures, accumulating searchable archive for years’ worth sales and scholarship:</p><p><a href="https://www.facebook.com/groups/240790932718333/?multi_permalinks=1923174691146607%2C1923985167732226%2C1923177001146376%2C1924108651053211%2C1923992764398133&amp;notif_id=1587655708212595&amp;notif_t=group_activity">Jual dan Lelang Komik Indonesia</a></p><p><em>Foreign Comic Collector Magazine</em>, briefly a physical magazine, now the Facebook home of hardcore international comics collectors:</p><p><a href="https://www.facebook.com/groups/274684996585171/">Foreign Comic Collector Magazine (Official Group)</a></p><p>One of many 2019 news items on the viral re-emergence of Mexico’s non-canonical <em>Hombre Arana:</em></p><p><a href="https://io9.gizmodo.com/idw-s-president-wants-to-unearth-spider-man-s-hidden-me-1836223985">IDW&#39;s President Wants to Unearth Spider-Man&#39;s Hidden Mexican History</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*g-KRXoDbP91E1EFhPthm9A.png" /></figure><p><strong><em>Joey Anuff</em></strong><em> is a full-stack </em><strong><em>web developer</em></strong><em>, focused on ML applications in the comic art domain. For more comic-related demos, please visit my </em><a href="https://metaguru.guru/"><strong><em>2018 portfolio</em></strong></a><em> or my</em><strong><em> </em></strong><a href="https://codepen.io/januff"><strong><em>CodePen demos</em></strong></a><em>.</em></p><p><em>Should I be prototyping your tech? I’m looking for interesting Dev Rel assignments involving: GraphQL, headless CMS, cloud ML (CV/NLP), Gatsby/Django.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fdc92418e104" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>