<?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 Brian Oakley, PhD on Medium]]></title>
        <description><![CDATA[Stories by Brian Oakley, PhD on Medium]]></description>
        <link>https://medium.com/@broak?source=rss-2bc3e20a5111------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Cupvyhtct7FYCJtVvfFzUw@2x.jpeg</url>
            <title>Stories by Brian Oakley, PhD on Medium</title>
            <link>https://medium.com/@broak?source=rss-2bc3e20a5111------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 03:54:03 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@broak/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[Who TF is “Bayes” and why should I care?]]></title>
            <link>https://engineering.ferocia.com.au/who-tf-is-bayes-and-why-should-i-care-d36f28518b82?source=rss-2bc3e20a5111------2</link>
            <guid isPermaLink="false">https://medium.com/p/d36f28518b82</guid>
            <category><![CDATA[statistics]]></category>
            <category><![CDATA[rational-thinking]]></category>
            <category><![CDATA[uncertainty]]></category>
            <category><![CDATA[bayes-theorem]]></category>
            <category><![CDATA[uncertain-times]]></category>
            <dc:creator><![CDATA[Brian Oakley, PhD]]></dc:creator>
            <pubDate>Thu, 26 Jun 2025 21:50:51 GMT</pubDate>
            <atom:updated>2025-06-26T21:50:51.521Z</atom:updated>
            <content:encoded><![CDATA[<h4>A universal framework for navigating an uncertain world</h4><figure><img alt="Point-of-view illustration. You are facing two doors; one open to a red hallway that veers left, the other open to a blue hallway that veers right. In your hands is a sledgehammer, embossed with `P(x)`. Between the two doors you have smashed through the wall to reach the outside. Rolling green hills extend toward the horizon, covered in countless branching paths. You have broken through binary thinking with the power of Bayes’ Theorem." src="https://cdn-images-1.medium.com/max/1024/1*c10-OIZ8wIEtrBB6U-oeNQ.png" /><figcaption>Illustration by the Author</figcaption></figure><p>Doesn’t it feel good when you “know” something for certain?</p><p>We’ve evolved to associate <em>uncertainty</em> with danger, so unsurprisingly we tend to avoid it where possible. But the world we live in feels more uncertain than ever, and this inability to cope with uncertainty may be somewhat to blame.</p><p>In times like these I find it helpful to remember that navigating uncertainty is a challenge for which an 18th century minister found an elegant solution, but died before he knew the impact he’d have on the world…</p><blockquote><strong>A NOTE FROM THE AUTHOR:<br></strong>If you’re familiar with “Bayes’ Theorem” I want to assure you that in this (short) article I will avoid the maths-heavy explanations and cliche’d analogies that are too often used when discussing this topic. If you’ve never in your life come across the name “Bayes” then I’m very glad you’re here!</blockquote><blockquote>Either way I hope you find this piece insightful.</blockquote><p>I feel fortunate to live in a country that, relative to much of the world, is safe, stable and prosperous¹. I <em>know</em> that when I wake up tomorrow my basic human needs will be met, and that “certainty” allows me to spend time on things higher up on my hierarchy of needs. Not everyone has that luxury. In desperate times, when meeting basic human needs is <em>far from certain</em>, it’s only natural that we’d seek out what feels like a “sure thing”, or a “safe bet” — anything to ensure the lights stay on and the fridge is kept stocked.</p><p>Unfortunately for us <em>as a species</em>, we’re still a violent pack of neolithic primates². We’re clearly more intelligent than all who came before; for example, we systematically terraformed the planet (at incalculable ecological cost) primarily to enrich ourselves, and we captured lightning in our pockets to more easily broadcast our plunder. And yet, with all that intelligence, we continue to fall for the same tricks played by the same confident idiots who have something to gain³.</p><p>We have <em>always</em> lived in an uncertain world, though recently it feels like the dial has been turned up to 11. Our climate is increasingly unstable, our systems of government more polarised and erratic. On the one hand, I believe this apparent increase in uncertainty is primarily the result of an <em>increase in awareness</em>. As a species we are much more connected than we have ever been (arguably too connected), and the availability bias is real⁴.</p><p>On the other hand, how we humans deal with uncertainty can make the problems worse⁵. Instead of acknowledging our uncertainty we reach for absolutes, thinking in black-and-white in an attempt to feel in control. We prefer the tidiness of a conspiracy theory over the confusing, inconclusive language of an expert. We attach so much of our self-worth and identity to our beliefs that admitting our mistakes feels worse than learning from them.</p><p>It’s a paradox of our nature. The tools we reach for in uncertain times (belief in shared mythology, group identity, control) helped us become the dominant species on earth⁶, but recently these tools are undermining our ability to navigate uncertainty <em>wisely</em>.</p><p>It doesn’t have to be this way.</p><h3>A BRIEF HISTORY LESSON</h3><p>The Reverend Thomas Bayes described in his writings <strong>a mathematical framework for updating prior beliefs when new evidence becomes available</strong>, though he did not share his work with anyone before his death in 1761⁷.</p><p>Bayes’ interpreted probabilities as if they were degrees of subjective belief, rather than the observed frequency of some repeated observation. I’ve illustrated the distinction below:</p><figure><img alt="An illustration highlighting the differences between the “Frequentist” and “Bayesian” interpretations of probabilities. The illustration uses two scenarios: the first involving dice rolls (which can easily be repeated to find the “long-run frequency”), and the second involving an end-game battle in a Dungeons and Dragons game (for which finding the “long-run frequency is impossible)." src="https://cdn-images-1.medium.com/max/1024/1*5BsLhsyFO9UM99OfGm8rGg.png" /><figcaption><strong>Bayes’ interpretation of probabilities as “degrees of subjective belief” is particularly useful when you need to estimate the probability of an event that cannot be repeated (eg. the probability of a specific candidate winning an election).</strong> Illustrations by the Author.</figcaption></figure><p>A few years after his death a friend named Richard Price found Bayes’ theorem among his notes, and recognising its importance, published the work posthumously.</p><p>Bayes’ name has since become synonymous with probability theory, and rightly so! His philosophy is the engine behind much of modern science and technology, and has led to advancements across many areas; from financial and meteorological forecasting, to medical diagnostics and machine learning⁸.</p><h3>THREE PRINCIPLES FOR NAVIGATING AN UNCERTAIN WORLD</h3><p>Without digging into the maths⁹, I’ve outlined below <strong>three Bayesian principles that I believe are useful for <em>everyone</em></strong><em>, </em>regardless of your line of work. I’ve kept them <em>intentionally vague</em> in order to maximise the audience who might find these tips relevant (a sneaky trick that astrologers and other soothsayers often play¹⁰). But<strong> </strong>for any <strong>software engineers or product folks</strong> reading this I’ve also included some more direct advice.</p><h4>#1: FALL IN LOVE WITH BEING WRONG</h4><p>Be careful to not let your <em>identity</em> be defined by your <em>beliefs</em>. When we make this mistake, we may interpret rational criticisms of our beliefs as if they were attacks on our identity. This is a dangerous trap! Because it is only natural to resist attack, at best we may ignore evidence that contradicts our beliefs. At worst we cling to our “identity” so strongly that we justify punishing others for not sharing our beliefs.</p><p><strong>Bayesian reasoning is a framework for keeping an open-mind</strong>. It describes how we can (and should!) update our existing beliefs in light of new evidence. Unsurprisingly this is a fundamental aspect of good science¹¹. The best experiments are those that have the highest chance of falsifying your hypotheses (given they’re actually wrong).</p><blockquote><strong>FOR THE DEVS:</strong><br>You can exercise this principle by assuming <strong>the first version of anything you ship will always be sub-optimal or incomplete</strong>. Knowing that, you should design with feedback loops built in from the start. It’s much easier to iterate on something when you have good signals for what needs improvement. Bonus points if you automate the loop by building something that can self-iterate¹².</blockquote><h4>#2: LEAN INTO THE UNCERTAINTY</h4><p>Uncertainty is a fundamental feature of the universe¹³. While we might prefer a clean “yes” or “no”, <strong>we must accept that reality is not black-and-white, and we should learn to navigate with incomplete information</strong>. By ditching the absolutes and making a habit of thinking about our beliefs in “degrees of certainty”, we become more open to new evidence and less vulnerable to dogmatism.</p><p>If you’re confident that a specific outcome is likely to occur, avoid using absolute statements like “I’m <em>certain</em> this will happen”. Instead,<strong> find a number that represents the strength of your belief</strong> (eg. “I believe there is an 85% chance of this happening”). This acknowledges that rarely can we be 100% certain about future outcomes, and it leaves the door open for you to incorporate new data.</p><p>If you’re finding it hard to settle on a number, here’s a tip that works surprisingly well: <strong>imagine betting on your belief with <em>actual money</em></strong>. Find the odds where you’d be willing to take either side of the bet, and then convert those odds to a probability<em>¹⁴</em>.</p><blockquote><strong>FOR THE DEVS:</strong><br>It can be helpful to <strong>write down your estimated probabilities of success when weighing up competing options</strong>. When you invest time/money into building something, you’re implicitly making a bet that you hope will “pay out”. But if you don’t know the “pay out” you’re expecting, how can you make a good decision about which bets you should take?</blockquote><h4>#3: NEVER STOP LEARNING</h4><p>By this point this will be obvious, but I’m mentioning it anyway because I believe it’s <em>the very essence</em> of Bayes’ Theorem — <strong>we can (and should) always update our beliefs when presented with new evidence</strong>.</p><p>Being confident in a belief is fine, as long as you remain open to new data that would adjust your level of confidence. If you are unwilling to adjust your confidence, you’re not learning — you’re <em>defending</em>.</p><p>One strategy I find helpful is to consider “what evidence would change my mind?”. By defining what evidence (if found) would falsify your belief, you’ll be more likely to find it, which gets you one step closer to the truth.</p><blockquote><strong>FOR THE DEVS:</strong><br>Remember that if a feature “flops” or an experiment “fails”, there is still insight to be gained. <strong>Keep track of your predictions and make note of the outcomes.</strong> Every result is new data that you should use to update your beliefs, which in turn will help you produce more accurate predictions in the future. <br>Tracking previous bets can help <strong>create a blame-free culture¹⁵ in your team, where unsuccessful bets are not seen as failures, but as opportunities for learning.</strong></blockquote><h3>FINAL THOUGHTS</h3><p>This post is more a love-letter to Bayes, than it is a recipe for a specific use-case of his theorem. I wanted to highlight the fact that <em>everyone</em> can find value in Bayesian reasoning, even if they aren’t “good with numbers”.</p><p>Bayes’ Theorem is <em>not just a set of equations</em>—it’s a humble but practical world-view that happens to be so robust it can be<em> proven mathematically</em>¹⁶. It feels like a fundamental law, a reminder that while the universe is inherently uncertain, <strong>by keeping an open mind and adjusting our beliefs as we go we will always find our way.</strong></p><p>Happy navigating!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6bnofWXJSuUPROYuGMAT_w.png" /></figure><h4>Footnotes:</h4><ol><li>Australia ranked 19th on the <a href="https://www.economicsandpeace.org/wp-content/uploads/2024/06/GPI-2024-web.pdf">IEP Global Peace Index in 2024</a>, and 7th on the <a href="https://hdr.undp.org/system/files/documents/global-report-document/hdr2025reporten.pdf">UN Inequality-Adjusted Human Development Index in 2025</a>.</li><li>ie. <a href="https://en.wikipedia.org/wiki/Human">Homo sapiens</a>.</li><li><a href="https://www.oxfordreference.com/display/10.1093/acref/9780191826719.001.0001/q-oro-ed4-00016553">Edward O. Wilson</a> said it best: <br><em>“The real problem of humanity is the following: We have Paleolithic emotions, medieval institutions and godlike technology.”</em></li><li>See: <a href="https://en.wikipedia.org/wiki/Availability_heuristic">Availability Bias (Wikipedia)</a>.</li><li>Tribalism thrives on moral absolutes and well-defined in-groups.</li><li>Yuval Noah Harari wrote about this (and much more) in his book <em>“Sapiens: A brief history of humankind”</em> (2015).</li><li>See: <a href="https://en.wikipedia.org/wiki/Thomas_Bayes">Thomas Bayes (Wikipedia)</a>.</li><li>See: <a href="https://en.wikipedia.org/wiki/Bayesian_inference">Bayesian Inference (Wikipedia)</a>.</li><li>For a deep-dive into the maths behind Bayes’ Theorem, I highly recommend the book <em>“Bayesian Statistics the Fun Way: Understanding Statistics and Probability with Star Wars, LEGO, and Rubber Ducks”</em> by Will Kurt (2019). If you’d prefer something much shorter, I also really like this blog post: <a href="https://medium.com/@keith-mcnulty/a-psychopathic-introduction-to-bayes-theorem-5c6e48d43a6e"><em>“A Psychopathic Introduction to Bayesian Statistics”</em></a> by Keith McNulty (2024)</li><li>See: <a href="https://en.wikipedia.org/wiki/Barnum_effect">The Barnum Effect (Wikipedia)</a>.</li><li>Thank you <a href="https://en.wikipedia.org/wiki/Falsifiability">Karl Popper</a>!</li><li>For example, by implementing a <a href="https://lazyprogrammer.me/bayesian-bandit-tutorial/">multi-armed/Bayesian bandit</a>.</li><li>See: <a href="https://en.wikipedia.org/wiki/Uncertainty_principle">Heisenberg’s Uncertainty Principle (Wikipedia)</a>.</li><li><em>An example: </em><br>Let’s say you believe that it is very likely to rain in the next week. <br> <br>Imagine you’re so certain of rain that you’d risk $80 to win just $20 (plus your original stake) if it does actually rain. The odds of this bet are 8:2 (ie. 4:1), which suggests you believe there is an 80% probability of rain.<br> <br>If that bet represents your true belief, you should by definition be equally willing to <em>take the other side</em> — betting $20 that it <em>won’t</em> rain, for the chance to win $80 if you’re correct (ie. suggesting you believe there is a 20% probability that it <em>won’t</em> rain). If the other side of the bet seems too risky, you must be less confident in rain than you originally thought!<br> <br>Simply adjust the odds until you find the odds where you would be equally willing to take either side of the bet — your “indifference odds”. This idea was introduced by <a href="https://en.wikipedia.org/wiki/Bruno_de_Finetti">Bruno de Finetti</a>, a statistician known for his work on “subjective probability”. His work suggests that the odds at which a person becomes willing to accept a bet will indicate their true beliefs about the likelihood of the event they’re betting on (because people don’t want to lose money!).</li><li>This is a value we proudly uphold in our team <a href="https://ferocia.com.au/">at Ferocia</a>, and I believe it is key factor in our success.</li><li>See: <a href="https://statproofbook.github.io/P/bayes-th.html">Bayes’ Theorem (The Book of Statistical Proofs)</a>.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d36f28518b82" width="1" height="1" alt=""><hr><p><a href="https://engineering.ferocia.com.au/who-tf-is-bayes-and-why-should-i-care-d36f28518b82">Who TF is “Bayes” and why should I care?</a> was originally published in <a href="https://engineering.ferocia.com.au/">Ferocia Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Pay up or we’ll read your mind!]]></title>
            <link>https://medium.com/@broak/how-much-are-we-willing-to-pay-for-mental-privacy-6c48fbd5a152?source=rss-2bc3e20a5111------2</link>
            <guid isPermaLink="false">https://medium.com/p/6c48fbd5a152</guid>
            <category><![CDATA[virtual-reality]]></category>
            <category><![CDATA[spatial-computing]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[apple]]></category>
            <category><![CDATA[privacy]]></category>
            <dc:creator><![CDATA[Brian Oakley, PhD]]></dc:creator>
            <pubDate>Sun, 25 Jun 2023 13:25:39 GMT</pubDate>
            <atom:updated>2025-06-26T05:46:48.610Z</atom:updated>
            <content:encoded><![CDATA[<h3>Does mental privacy still exist in the age of spatial computing?</h3><figure><img alt="An illustration of my Memoji scowling while wearing Apple Vision Pro. The background features a bold red stripe of paint, and within the red portion a repeating pattern of eyes, in a simple illustrative style. The image overall is reminiscent of the “Big Brother is Watching You” posters from George Orwell’s ‘1984’, and is meant to evoke a sense of foreboding." src="https://cdn-images-1.medium.com/max/1024/1*mm8Rj5re6ITv1NdzOxYNfw.png" /><figcaption><strong>Big Tech is watching you.</strong> (Illustration by the author)</figcaption></figure><p>Something has been worrying me <a href="https://www.apple.com/newsroom/2023/06/introducing-apple-vision-pro/">since Apple Vision Pro was announced at WWDC ’23</a>, and it’s not just how much lighter my wallet will feel once it’s available to the public.</p><p>A few years ago I wrote about <a href="https://medium.com/@broak/vr-and-ar-tech-is-headed-down-a-dark-path-a242bd6c6ea5">the surveillance-nightmare future I felt we were barreling towards</a>. A world where like bespectacled mimes we’d gleefully throw <a href="https://pokemongolive.com/?hl=en">virtual balls at hologram Pokemon</a>, allowing every moment of our lives to be recorded from our first-person-perspective, processed and directed to make the <a href="https://en.wikipedia.org/wiki/Nudge_theory">constant nudging</a> ever more effective. Where we’d spend every possible minute of attention on <a href="https://en.wikipedia.org/wiki/The_Age_of_Surveillance_Capitalism">whatever<em> the machines</em> have predicted is most profitable at the time</a>.</p><p>After consuming every bit of news I could find about Apple Vision Pro, I think I’m more frightened than ever. As many predicted, the device makes heavy use of eye-tracking to ensure your digital world looks crisp and clear, and every interaction feels “intuitive and magical”. The level of UI polish on display was admittedly stunning, and would simply be impossible without the device keeping track of your gaze at a very high frequency.</p><p>On the other hand, the privacy implications of eye-tracking are so worrisome that <a href="https://www.apple.com/newsroom/2023/06/introducing-apple-vision-pro/">Apple made specific reference to this during the announcement</a>, explaining (thankfully) that third-party developers will not have access¹. Their stance on privacy is nothing new. Apple have used it to both <a href="https://www.apple.com/privacy/">differentiate themselves</a> from the rest of “Big Tech”, and <a href="https://www.forbes.com/sites/danielnewman/2022/02/10/apple-meta-and-the-ten-billion-dollar-impact-of-privacy-changes/">to knee-cap their competitors to the tune of $10 billion dollars</a>. But in this spatial-computing era, this differentiation appears dystopian.</p><h3>The Telly-screens</h3><p>Imagine you’re shopping for a flatscreen TV. You can get a decent 49” 4K model for around A$700, but then you see <a href="https://www.freetelly.com/"><em>Telly</em> — a “dual-screen” 55-incher with an advertised price of “FREE”</a>. How could you say no?</p><p>The catch is that Telly tracks <em>everything</em> you watch.</p><p>If you weren’t already aware, <a href="https://www.theverge.com/2023/5/15/23721674/telly-free-tv-streaming-ilya-pozin-ads">this is a real product</a>. Agree to the terms, and you too can have a living room that resembles train station platform, with a giant billboard ad on the secondary screen — a persistent space for precisely targeted advertising, even if you’ve “switched off” your TV. It even has a camera facing out so you can watch old seasons of Big Brother alongside your friends remotely, the secondary screen showing the view from the camera on <em>their</em> Telly, like some Orwellian version of FaceTime.</p><p>So we now know our attention is <em>at least as valuable</em> as a 55-inch TV. Telly will provide the hardware for free as long as you agree to let them peek at everything you look at, to better understand what you find boring and what gets you going. The more they know about you, the more their <em>true customers</em> — other companies wanting to advertise to you — will pay for access to your eyeballs.</p><p>But what if advertisers could <em>see more</em> than just which thumbnails you clicked and how long you spent watching? While these behavioural data are plenty invasive, they are just the footprints you leave along the <em>conscious</em> path you tread across the web. What if they could somehow peek at your <em>subconscious</em>, to know which steps you’d <em>like to take</em> even before you’ve taken them?</p><h3>Look me in the eye</h3><p>There’s a reason why poker players are often seen wearing sunglasses. You can tell a lot about what somebody is thinking by observing their gaze. <a href="https://doi.org/10.1126/science.1243091">Our ability to instinctively decode the eye motion behaviour</a> of others allows us, in a way, to <em>read the minds</em> of those around us.</p><p>Unlike the links we click or buttons we tap, our <a href="https://www.ncbi.nlm.nih.gov/books/NBK10991/">eye movements are largely reflexive</a>. Clicking through Netflix with a remote control might send a few sparse signals like `buttonClicked` and `secondsHovered`, but with eye-trackers you can record precisely <em>what a user is looking at</em> over 100 times per second.</p><p>So not only can eye-trackers reveal what (or <em>who</em>) keeps catching your eye, but <a href="https://doi.org/10.1007/978-3-030-42504-3_15">from changes in your pupil diameter <em>they can even quantify how aroused you become</em></a><em>. </em>Chuck in a Bluetooth connected smartwatch (for heart rate and blood pressure) and some earbuds that can measure EDA², and you’ve got yourself a biometric do-it-all mind-reading kit. The surveillance instruments we’ve got today seem like toys in comparison.</p><p>Forget advertising, imagine what this <em>surveillance of thought</em> could mean for oppressive governments wanting to stamp out dissent or more effectively discriminate against minorities. Imagine if this data were leaked to the public, not from any fault of your own but <a href="https://www.abc.net.au/news/2022-11-10/medibank-data-breach-latest-dark-web-leak/101632746">because it was mishandled by one of the many services you’ve entrusted with it</a>.</p><h3>Buy the product, or become the product</h3><p>As Scott Galloway puts it, the entire world is diverging into two business constructs: <a href="https://nymag.com/intelligencer/2021/06/why-subscriptions-may-be-the-wave-of-the-future.html">Apple or Android</a>.</p><p>Apple products are expensive. They’re sold at such a premium partially to offset the opportunity cost of leaving the data of their users “untapped”. Those who can afford the high up front cost (roughly the wealthiest <a href="https://www.statista.com/statistics/272698/global-market-share-held-by-mobile-operating-systems-since-2009/">28% of consumers globally</a>) are given the privilege to enter a curated garden with walls<em> just high enough</em> that <a href="http://en.wikipedia.org/wiki/The_Age_of_Surveillance_Capitalism">the surveillance capitalists</a> have trouble peering over.</p><p>For everyone else there’s Android. Those who cannot afford to turn down a free TV pay instead with their data — <em>feeding</em> the persuasive machines built to maximise consumption. Consumption that helps perpetuate their exclusion from the garden.</p><p>It’s a viscous cycle. A positive feedback loop that ensures the very people who most need to <em>stop consuming</em> are the ones who’ll be impressed upon the most to consume. We’ve seen this story before, where <a href="https://doi.org/10.1111/obr.13144">inequalities in exposure can drive inequality in outcomes</a>.</p><p>It can’t be ignored how selling us the privilege of privacy is simply good business strategy for Apple. It’s in their interests for <a href="https://www.profgalloway.com/apple-thief/">Meta and Google to be reliant on data</a> generated by Apple’s customers, because <a href="https://www.macworld.com/article/611551/facebook-app-tracking-transparency-iphone-quarterly-results.html">Apple controls the hose</a>. Privacy also helps them justify the high prices of their products, which further helps play into their positioning as a luxury brand.</p><p>They happily leave the low-end of the market open for others who fall into a race to the bottom, mining as much data as they can so they can subsidise the cost of hardware and reach the widest possible audience.</p><h3>I’m still going to buy it</h3><p>I fully intend to buy the Apple Vision Pro as close to the Australian release date as possible — not just because I’m an Apple fanboy (I fully admit I am), but because<em> I genuinely believe that spatial computing is the future.</em></p><p>Where the invention of headphones made it possible to enjoy <em>auditory content</em> in a more personal, portable way; I believe these “head-screens” will eventually do the same thing for <em>visual content</em>. Big-screen TVs and monitors won’t totally disappear, but I do think they’ll become more niche. After all, even though wireless earbuds have become the most convenient way to listen to music, you can still buy a nice set of bookshelf speakers if you’re so inclined. Headsets like Apple Vision Pro will be just like AirPods for our eyes.</p><p>But while the potential upsides excite me, I can’t help but worry about what will <em>inevitably go wrong</em>. The glacial pace of tech regulation so far terrifies me when I think about how these challenges are about to accellerate. If we’ve learned the “move fast, break things” mantra was untenable during the era of smartphones, it is even more so the case in the era of smartglasses.</p><p>I’m not worried about how much these devices will cost upfront. I’m worried that we’ll pay a much higher price down the road if we’re not careful.</p><p>I hope we know what we’re doing.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*H_IT1viMQfyXWF6WuWq5ag.png" /></figure><h4>Footnotes:</h4><p>[1] To be clear, eye-tracking can still be used to interact with third-party apps — a dedicated R1 chip will process this data at the system-level, and apps will only know where a user is look at the moment the user “clicks”.</p><p>[2] Electrodermal activity — which is used to measure changes in arousal of the sympathetic nervous system (which is in charge of your fight or flight response)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6c48fbd5a152" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Three Principles for Making Better Decisions]]></title>
            <link>https://medium.productcoalition.com/three-principles-for-making-better-decisions-cbcb3bbea956?source=rss-2bc3e20a5111------2</link>
            <guid isPermaLink="false">https://medium.com/p/cbcb3bbea956</guid>
            <category><![CDATA[bias]]></category>
            <category><![CDATA[statistics]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[risk-management]]></category>
            <category><![CDATA[decision-making]]></category>
            <dc:creator><![CDATA[Brian Oakley, PhD]]></dc:creator>
            <pubDate>Fri, 01 Oct 2021 05:33:19 GMT</pubDate>
            <atom:updated>2022-09-09T03:35:59.389Z</atom:updated>
            <content:encoded><![CDATA[<h4>Product Management, Meet Behavioural Economics</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1l-i1zqUVB9H9qqwArnvHQ.png" /><figcaption>Illustration by the Author</figcaption></figure><p>Of the countless decisions you have made throughout your life, how many of them would you say were <em>correct</em>?</p><p>I don’t actually expect you to come up with a number, but I think it’s fair to say we all want to <em>maximise the number of correct decisions</em>. Put another way, we want to <em>avoid making errors</em>.</p><p><a href="https://en.wikipedia.org/wiki/Daniel_Kahneman">Daniel Kahneman</a> is a Nobel Prize-winning Psychologist and perhaps the world’s leading expert on judgement and decision-making. In his latest book “Noise: A Flaw in Human Judgement” (co-authored by <a href="https://oliviersibony.com/about/">Olivier Sibony</a> and <a href="https://en.wikipedia.org/wiki/Cass_Sunstein">Cass R. Sunstein</a>), Kahneman describes how decisions gone awry can have devastating impacts, and outlines strategies for avoiding these errors.</p><p>According to Kahneman, errors in decision-making are comprised of two things: <em>bias</em> and <em>noise. </em>But when people (or organisations) want to improve their decision-making, they tend to focus only on <a href="https://www.mckinsey.com/business-functions/strategy-and-corporate-finance/our-insights/bias-busters">reducing biases</a>. This is because biases are much more visible when reflecting on an <em>incorrect</em> decision in hindsight — humans crave causal explanations, and biases have an “explanatory charisma” which noise lacks.</p><blockquote>“<strong>Confirmation bias</strong> is the reason they made the wrong decision”.</blockquote><blockquote>“They hired the wrong person because they fell for the <strong>halo effect</strong>”.</blockquote><blockquote>“No one spoke up beforehand because of <strong>social desirability bias</strong>”.</blockquote><p>Biases, once identified, are predictable. If a set of bathroom scales gave a reading that was <em>consistently</em> <em>higher</em> than your true weight, then you would say the scales are <em>biased</em>. Because the error is predictable, the process of eliminating it is more obvious.</p><p>In contrast, a <em>noisy</em> bathroom scale would be inconsistent, giving a <em>different</em> reading every time you stood on it. This is why noise is so often ignored — the errors it produces are less visible because they are <em>unpredictable</em>. Though less visible, the consequences of these errors are no-less costly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MzzTUx7TEHaSj3W3uiFs7Q.png" /><figcaption>Bias and noise described using a shooting range as an analogy — making good decisions is akin to shooting accurately. <strong>Biases are systematic influences that cause shared errors</strong> (like a slight bend in a gun barrel). <strong>Noise is the unwanted variability that contributes to overall error</strong> (like having shaky hands).</figcaption></figure><h3>The Cost of a Bad Decision Can Be Enormous</h3><p>The manufacturers of the Concorde supersonic airplane famously fell victim to a common bias: the <a href="https://en.wikipedia.org/wiki/Sunk_cost">sunk-cost fallacy</a>, when they made the erroneous decision to continue the project even after <a href="https://thedecisionlab.com/biases/the-sunk-cost-fallacy/">it was clear that breaking-even was impossible</a>. Their aversion to “losing” the significant investment already made ultimately caused them to lose <em>more</em> than they otherwise would have — in the end approximately <a href="https://en.wikipedia.org/wiki/Concorde">£<em>3.4 billion pounds</em></a> (adjusted for inflation).</p><p>Decision errors are <a href="https://www.mckinsey.com/business-functions/risk-and-resilience/our-insights/the-business-logic-in-debiasing">more common than you think</a>, and because they accumulate over time, the financial impacts can be staggering. Therefore, it’s not surprising that lot of work has been done to minimise the number of errors made in organisational decisions.</p><p>Many organisations now actively use <a href="https://www.mckinsey.com/business-functions/risk-and-resilience/our-insights/the-business-logic-in-debiasing"><em>debiasing</em> techniques</a> to reduce the business risks associated with biased decision-making. The methods they employ are based on learnings from<a href="https://www.mckinsey.com/industries/financial-services/our-insights/all-in-the-mind-harnessing-psychology-and-analytics-to-counter-bias-and-reduce-risk"> the fields of psychology and behavioural economics</a>, to which Kahneman’s work has contributed significantly. In his latest book he builds upon this work by highlighting the contribution of noise to the error equation:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*m6WYvEHI_I3xfxzSCpWY_A.png" /><figcaption>Decision errors are a function of <strong>both bias and noise</strong>. Reducing either one reduces overall error, however most of the focus to date has been only on reducing biases, while the contribution of noise has been ignored.</figcaption></figure><p>Kahneman outlines a set of principles for making better decisions, which he calls <em>decision hygiene. </em>Just as practicing good personal hygiene can prevent illness, practicing good decision hygiene can prevent errors in decision-making.</p><p>In this article I’ve summarised these principles below, through the lens of a product manager. However I’ve kept the scenario somewhat vague, since I believe these principles are universally valuable and can be applied in a variety of contexts. I’ve also included a short summary at the end for those who are already somewhat familiar with these principles — think of it as a “decision hygiene cheat-sheet”.</p><blockquote>“Wherever there is judgement, there is noise — and more of it than you think.”</blockquote><blockquote>— from “Noise: A Flaw in Human Judgement” (p.12)</blockquote><h3>Three Principles of Decision Hygiene</h3><p><strong><em>Let’s imagine the following scenario:</em></strong></p><p>You’re a product manager and you want to build a feature which will decrease the time to value for your customers. After some discovery work, you’ve narrowed down your options to five possible features — all of which seem <a href="https://medium.com/whispir/what-phases-do-product-ideas-go-through-at-whispir-5bbcd0add3">desirable, usable, feasible and viable</a> — but you can only build one.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1jYliV-oUDA1Jj3FdwIsOw.png" /><figcaption>Imagine five options, labelled A, B, C, D, and E.</figcaption></figure><p>How do you decide which one is the right choice? In an ideal world you would choose the one that will produce the greatest benefit with the lowest cost, but how can you tell which that is without being able to see the future?</p><h3><strong>Principle One — “</strong>Resist premature intuitions.”</h3><p><strong><em>Structure your judgements using several independent criteria.</em></strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zIs4ovk2DIx-eJatFYDpXQ.png" /><figcaption>Structuring your judgements allows you to focus on the important factors, and compare different options on a level playing field.</figcaption></figure><p>When you look at the five options, you might find that one immediately appears more attractive than the others — this is your intuition talking, and you <em>must ignore it</em> (for now, it will come in handy at the end!).</p><p>Don’t get me wrong, your intuition is a great time-saver but it is quick to jump to conclusions, and it often makes mistakes (If you’ve read Kahneman’s previous book “<a href="https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow">Thinking, Fast and Slow</a>”, this is <a href="https://suebehaviouraldesign.com/kahneman-fast-slow-thinking/">System 1</a>). We can reign in our intuitions by <em>giving our judgements some structure</em>.</p><p>Instead of evaluating each option as a whole, define 3–5 specific criteria which are <em>relevant to the decision being made</em>. Then give each option a rating along each of these specific dimensions. For example, you might want to compare options based on technically complexity, Or maybe how much of an impact each would have on a specific organisational metric. The important point is to <em>consider each criteria in isolation</em>, to avoid letting a good score on one criteria blind you to a bad score on another.</p><p>By focusing on specific criteria, you force yourself to compare the options on a level playing field. This principle is particularly powerful in the hiring process (<a href="https://careers.google.com/how-we-hire/#step-interviews">Google use it!</a>) as it can help organisations avoid common biases, such as the <em>halo effect</em> — where a positive first impression casts a “halo” that can lead a hiring manager to overestimate a candidate’s ability, or even disregard a clear red flag.</p><blockquote>“We have a structured process to make hiring decisions. Why don’t we have one for strategic decisions? After all, options are like candidates.”</blockquote><blockquote>— from “Noise: A Flaw in Human Judgement” (p.324)</blockquote><h3><strong>Principle Two — “</strong>Think like a statistician.”</h3><p><strong><em>Take the “outside view”, because everything regresses to the mean. Try to use relative judgements rather than absolute ones.</em></strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UN5AZ4wGT9_gc41hD4aZWA.png" /><figcaption>Stepping back to take the <strong>outside view</strong> means putting the current situation in context, and considering the statistics of each option.</figcaption></figure><p>You don’t need to have formally studied statistics to think like a statistician (though bonus points if you have!), the key here is to <em>take the outside view</em> when evaluating your options.</p><p>In our example, building <strong>Option B</strong> might be highly popular within the organisation, and so you might be tempted to choose it based solely on that fact. However this is taking the <em>inside view</em> — you have only considered the information from <em>within </em>your current situation.</p><p>Imagine instead that you looked more broadly at what happened when <em>other organisations </em>chose to build something similar to <strong>Option B</strong>. It would not be surprising to find that within your own organisation, the expected benefits of <strong>Option B</strong> are overly optimistic when put in context.</p><p>In short, taking the outside view means taking into account the statistics of a <em>class of cases</em> similar to the present one. Anchor your expectations to the <em>average — </em>a good rule-of-thumb given the inevitable <a href="https://en.wikipedia.org/wiki/Regression_toward_the_mean">regression to the mean</a>.</p><p>Another important thing to consider is the scale on which you are evaluating your options. Where possible, use relative scales as opposed to absolute ones. It has been shown that people are much better at ranking options than they are at scoring them on an abstract scale.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Igttn1GJZdltepsqP73YZA.png" /><figcaption>In this example, two people have very different perceptions of the absolute scale, though their relative rankings are identical.</figcaption></figure><p><a href="https://www.mckinsey.com/business-functions/strategy-and-corporate-finance/our-insights/sounding-the-alarm-on-system-noise">A relative scale is <em>less noisy</em> than an absolute one</a> because pairwise comparisons are more reliable both between individuals, and within an individual at different points in time.</p><p>For example, if you were asked to assign <strong>Option B</strong> a number between 0 and 10 today, and then again in a few weeks time, you’d likely choose a different number. However your pairwise comparisons (“B is better than A, but not as good as E”) will be much more consistent over time.</p><p><em>(Author’s Note: I am not suggesting to avoid absolute scales entirely! However, when you are comparing the relative values of a number of different options, ratings along relative scales are more statistically reliable).</em></p><blockquote>“We effortlessly invoke causes for the events we observe, but thinking statistically about them must be learned and remains effortful. Causes are natural; statistics are difficult.”</blockquote><blockquote>— from “Noise: A Flaw in Human Judgement” (p.219)</blockquote><h3><strong>Principle Three — “</strong>Accuracy, not individual expression.”</h3><p><strong><em>Aggregate multiple independent judgements. Use the wisdom of the crowd to anchor decisions to a reliable estimate.</em></strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zCNWFBl50lHrJja1eHONMA.png" /><figcaption>A single judgement is vulnerable to errors (both bias and noise). Averaging multiple independent judgements produces<strong> more reliable estimates</strong>.</figcaption></figure><p>The goal of judgement is to be <em>accurate</em>, not to express the opinion or intuition of a single individual. Therefore, it is almost always a good idea to leverage the combined experience/knowledge of a group of stakeholders.</p><p>If you’ve ever watched the game show “Who Wants to be a Millionaire”, you’ve witnessed this principle in action whenever a contestant chooses to “<a href="https://millionaire.fandom.com/wiki/Ask_the_Audience">ask the audience</a>”. Every audience member votes for the answer they believe is correct, and the contestant then uses the aggregate results to anchor their own decision. This is a powerful lifeline to use, in fact the US host has claimed the aggregated votes of the audience <a href="http://www.freestak.com/blogposts/who-wants-to-ask-the-audience-the-benefits-and-pitfalls-of-social-media/">are correct 95% of the time</a>.</p><p>This principle works because of the phenomenon known as the “<a href="https://en.wikipedia.org/wiki/Wisdom_of_the_crowd">wisdom of the crowd</a>”. A group of people will naturally hold a variety of viewpoints, and therefore a variety of biases or preconceptions. From a statistical perspective, aggregating multiple viewpoints “averages out” the error within the group — leading to a more accurate estimate of the true value.</p><p>It is important to note this can backfire spectacularly if the crowd you choose is too homogenous (i.e. not enough diversity of knowledge/expertise). If everyone in the group shares the same biases, averaging their judgements will not average out the bias.</p><h4>The key to choosing a wise crowd</h4><p>Find a <em>diverse</em> group of stakeholders with experiences and knowledge that <em>vary</em>. Make sure each person has the time and space to make their judgements independently, and early on (using the first two principles!), <em>before</em> the group comes together for a round-table discussion. Too much cross-talk will cause the group to become more homogenous in their thinking — because people will naturally <a href="https://en.wikipedia.org/wiki/Authority_bias">influence each others decisions</a>.</p><p>Finally, <em>aggregate the independent ratings </em>so that you have a single score for each item. There are many ways to do this, but the easiest method is a simple average (<em>For the maths nerds reading this, follow me on Medium if you’re interested in reading a follow-up article describing more specific methods!</em>)</p><p>Aggregating the ratings has the added benefit of anonymising the individuals judgements. If people know that their judgements will be visible, they may alter their judgements to avoid scrutiny — this is called <a href="https://en.wikipedia.org/wiki/Social-desirability_bias">social desirability bias</a>. This observer effect is a key reason why we use voting booths in elections and anonymous surveys when gathering employee feedback — because otherwise we cannot expect the responses to be honest.</p><blockquote>“Unfortunately, most roundtables are run in a way that makes it much too easy to achieve agreement, because participants quickly converge on the opinions stated first or most confidently.”</blockquote><blockquote>— from <a href="https://hbr.org/2016/10/noise">Harvard Business Review</a></blockquote><h3>Making the Final Decision</h3><p>At the beginning of this article, I told you to <em>temporarily ignore your intuitions</em>. It is at this point that you can <em>unleash it</em>.</p><p>If you followed the principles of decision hygiene described above, you will have a reliable rating for each of the five options, and these ratings will reflect the structured, independent judgements of a wise crowd.</p><p>It is important that your process <em>does not prescribe</em> the final decision! The goal of decision hygiene is to <em>reduce decision errors by delaying intuition</em>, not banning it. If you simply let the highest aggregate score determine which option is chosen, stakeholders may feel their hands have been tied and reject the process entirely. Instead, take the wisdom of the crowd to the table and use it as an anchor in the final discussion.</p><p>Without an anchor it is more likely that a single loud or prominent voice will unconsciously steer a group off-course. But someone would need a particularly strong and reasoned argument to dispute a decision reliably anchored by structured considerations, reliable statistics, and a wise crowd.</p><h3>A Short Guide to Practicing Decision Hygiene</h3><ul><li>Structure your judgements into several <em>independent</em> criteria. This will help you consider the different facets of each option more clearly, and therefore more objectively.</li><li>Avoid using absolute rating scales. Instead, <em>rank</em> your options from best to worst, considering each criteria <em>in isolation</em>. That means if you have three different criteria in your evaluation, you will have three separate rankings.</li><li>Consider the statistics of similar situations when making your judgements. Don’t fall into the trap of thinking “because X happened once, it is likely to happen again”— anchor your estimates to the <em>mean</em>.</li><li>Use the wisdom of the crowd. Gather a diverse group of trusted stakeholders, and avoid discussing the decision before every person has evaluated all options <em>independently</em>.</li><li>Aggregate the scores into a single metric for each option. Use this aggregated score as an anchor when making the final decision.</li></ul><p>With time, practicing decision hygiene becomes second-nature, and it will feel as simple (and necessary!) as washing your hands. Now go forth and take these principles with you as you navigate the many decisions that lie ahead!</p><p>Choose wisely!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*CNf-6cCBPr_b4TMswmh3Yg.png" /></figure><p><strong><em>If you’re curious…</em></strong></p><p>I’m a computational neuroscientist working in fintech, and I use my background in quantitative research, statistics and behavioural science to help guide product development.</p><p>This article was heavily inspired by the book “Noise: A Flaw in Human Judgement” by <a href="https://en.wikipedia.org/wiki/Daniel_Kahneman">Daniel Kahneman</a>, <a href="https://oliviersibony.com/about/">Olivier Sibony</a> and <a href="https://en.wikipedia.org/wiki/Cass_Sunstein">Cass R. Sunstein</a> (2021). I believe the principles and practices it outlines have immense value, and I want to share the things that I have found helpful in my own line of work.</p><p>We all make multiple decisions every day. Granted most of them aren’t high-stakes organisational bets, but regardless — decision hygiene isn’t a process, it’s a <em>set of principles</em> I strive to apply to every judgement I make.</p><p>I hope I’ve inspired you to do the same.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cbcb3bbea956" width="1" height="1" alt=""><hr><p><a href="https://medium.productcoalition.com/three-principles-for-making-better-decisions-cbcb3bbea956">Three Principles for Making Better Decisions</a> was originally published in <a href="https://medium.productcoalition.com/">Product Coalition</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Visualising Trade-Offs in Software Development]]></title>
            <link>https://medium.com/whispir/visualising-trade-offs-in-software-development-82cd0b00072d?source=rss-2bc3e20a5111------2</link>
            <guid isPermaLink="false">https://medium.com/p/82cd0b00072d</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[opportunity-cost]]></category>
            <category><![CDATA[decision-making]]></category>
            <category><![CDATA[technical-debt]]></category>
            <dc:creator><![CDATA[Brian Oakley, PhD]]></dc:creator>
            <pubDate>Tue, 07 Sep 2021 01:07:39 GMT</pubDate>
            <atom:updated>2021-09-08T23:48:34.268Z</atom:updated>
            <content:encoded><![CDATA[<h4>Marshmallows for Product Managers</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*If8oqjcUgV6YIRNh0nkQUA.png" /><figcaption>Illustration by the Author</figcaption></figure><p>You might have heard of the famous <a href="https://en.wikipedia.org/wiki/Stanford_marshmallow_experiment">Stanford marshmallow experiment</a> from 1972, where children were given a single marshmallow on a plate, with the promise of a second one if they could wait some time before eating it. If the child couldn’t bear to <a href="https://en.wikipedia.org/wiki/Delayed_gratification">delay gratification</a> for 15 minutes (and many could not!) they would forfeit the second marshmallow</p><p>This experiment highlights something we all understand intuitively — sometimes it is difficult to forgo short-term gains in exchange for some delayed, but much greater, long-term gain. Even when the benefits are clearly defined and unambiguous, like compound interest on retirement savings, we often find ourselves instead prioritising the <em>here-and-now. </em>I suspect this is why <a href="https://www.abc.net.au/news/2021-02-26/kim-kardashian-afterpay-millennials-zip-buy-now-pay-later/13194484">AfterPay is so popular among jaded millennials</a>, myself included.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jT9cmgo7HVTMHz7J9h7GWA.png" /><figcaption><strong>Human beings kinda suck at delaying gratification, even if the opportunity cost is large and unambiguous.</strong> Image modified by the Author. Source: <a href="https://knowyourmeme.com/memes/distracted-boyfriend">https://knowyourmeme.com/memes/distracted-boyfriend</a></figcaption></figure><h3>The tech-debt bog</h3><p>The speed at which a company can develop software is a function of <em>resources</em>, <em>requirements</em> and <em>tech-debt</em>. That third parameter <a href="https://adhoc.team/2021/06/02/tech-debt-interest-rate/">slowly and inevitably increases with time</a>, and like driving through a muddy bog, the deeper in it you are, the slower you move.</p><p>This leads to a common dilemma for PMs:</p><blockquote>Continue delivering features (albeit slowly, due to accumulated technical debt)?</blockquote><p><em>OR</em></p><blockquote>Press pause on delivering features and focus on addressing the problem, which will increase our velocity long-term?</blockquote><p>The <em>right</em> answer varies business to business, and also changes with time depending on available resources, level of tech-debt, etc. The answer today may be different to the answer in 6 months time, and it’s likely to create some tension within the business — where the priorities of different functions pull the decision-maker in multiple, competing directions.</p><p>Perhaps your devs are fully on-board with multiple sprints dedicated to code refactoring and clean-up (<em>“it will make my life so much easier going forward”</em>), while other stakeholders may balk at the idea of an extended feature drought (<em>“how do I justify delaying this feature yet again?”</em>). If you’re the one tasked with deciding, how do you choose a path? And perhaps more importantly, how do you convince people that your choice was the right one?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*STi9ejl1rSYnuUqA-ZyhXA.png" /><figcaption>Source: <a href="https://vincentdnl.com/drawings/">https://vincentdnl.com/drawings/</a></figcaption></figure><h3>Q<em>uantify and visualise</em> the trade-off</h3><p>While it may be much easier to argue for one option or another in the abstract, it is often said “a picture is worth a thousand words” — so <em>let’s draw a picture</em>.</p><h4>Consider the current situation</h4><p>We can visualise <em>Option One</em> (the path where development continues, but slowly due to tech-debt) by drawing a simple line using the equation:</p><blockquote><strong>y = mx</strong></blockquote><p>… where <strong><em>y</em></strong> is your <em>measure of the outcomes/output</em> (for example, “average number of features delivered”), <strong><em>x</em></strong> is your <em>measure of time </em>(e.g. “months”), and <strong><em>m</em></strong> is your current<em> speed</em> (“the number of features delivered per month”). These are just examples, use whichever measures make the most sense for your given situation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NjcJstsB2ynw0eHWrnoXQg.png" /><figcaption><strong>Even small changes in speed (<em>m</em>) result in large differences in outcomes over time.</strong></figcaption></figure><p>Using this equation, you can very easily calculate the <em>expected</em> number of features delivered after <strong><em>x</em></strong> months, by multiplying by <strong><em>m</em></strong><em>.</em> If, on average, you deliver <strong>2.5 features per month</strong>, in <strong>24 months</strong> you should have delivered <strong>60 features</strong> (2.5 <em>multiplied by</em> 24 = 60).</p><h4>Consider the alternative</h4><p>Drawing the second option (the one where development is temporarily halted) is slightly more complex, since we have a few extra parameters to consider.</p><p>First you need to estimate how long the “pause” will last — let’s call this parameter <strong><em>p</em></strong>. This will depend on your particular situation, but I’d suggest leaning towards <em>over</em>-estimating here (you’ll see why later).</p><p>For this example, let’s say it will take <strong>4 months</strong> to sort out the tech-debt (therefore <strong><em>p</em></strong> = <strong>4</strong>), and during that time you expect <strong>zero</strong> features to be delivered (<strong><em>m</em></strong> = <strong>0</strong>). Because anything multiplied by zero equals zero, this period of time looks like a horizontal line from the origin to where <strong><em>x </em></strong>= <strong>4</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Sc96qm63Pg4PkYX7teHAKA.png" /><figcaption><strong>Immediately it is clear the green path comes at at a cost, at least in the short-term.</strong></figcaption></figure><p>It may be the case that some proportion of the team <em>can</em> continue to develop new features while the tech-debt is being addressed. If so, you can adjust the slope of the line accordingly (<em>see left</em>). You can even use a negative value to represent work that will need to be re-done (i.e., you’ve gone backwards a bit) as a result of the pause. In any case, make sure to cut the line off at <strong><em>p</em></strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bKb38DV0eYsC-R65K7EgVw.png" /><figcaption><strong>Outputs during the pause don’t need to be zero.</strong></figcaption></figure><p>Finally, you need to estimate your <em>speed after the pause</em>. This is challenging because you’re essentially trying to predict how the work done during the pause will change productivity in the future. I’m not going to get into the details of how to find this number (maybe in a future post if anyone is interested?), but again I’d suggest using a conservative estimate here.</p><p>For now, lets say you’ve spoken to your tech leads and they expect that after the tech-debt is dealt with, average development time per feature should halve (therefore, speed should double). In this scenario, the slope of the line, <strong><em>m</em></strong>, is double that of <em>Option One</em> (2 * 2.5 = 5, so <strong><em>m = </em>5</strong>).</p><p>Draw <em>this</em> line originating at the end of the line representing the pause, and you <em>should</em> see that it crosses over the line for <em>Option One</em> — <em>this is the “break-even point”</em>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MOotxw8iLBfV4rtlk1WSGg.png" /><figcaption><strong>Option One in pink, Option Two in green. The break-even point is highlighted in blue.</strong></figcaption></figure><p>In the chart above, you can clearly see the short-term losses associated with <em>Option 2</em> are paid back in full at the 8 month mark. From that point onwards, the increase in speed <em>compounds</em>.</p><p>This is particularly dramatic when taking the long-term view:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ivUzE2Xmg2R6BBFEeh8DJQ.png" /><figcaption><strong>Zoom-out a bit to see the long-term gains (i.e. the opportunity cost of choosing Option One).</strong></figcaption></figure><p>At the 3 year mark, choosing <em>Option One</em> would result in around 90 features being delivered. In contrast, choosing <em>Option Two</em> would mean developing 90 features would take only 22 months (just under 2 years), and at the 3 year mark, 160 features will be delivered.</p><p>In this example case, the long term compounding benefits are dramatic. This is why earlier I suggested using conservative estimates. Another approach is to draw two lines for <em>Option Tw</em>o— a “best-case” and a “worst-case” scenario — and consider the range of possible break-even points when making your decision.</p><h3>So which is the right option?</h3><p>After visualising both options, you will still need to weigh the costs and benefits before coming to a decision. If your situation <em>requires</em> some outcome before development can resume, then you may need to stick with Option One. If the expected short-term losses can be tolerated, and the long-term gains are attractive — Option Two is the way to go. If you’ve followed these steps and found that the two paths<strong><em> do not</em></strong> cross over (i.e. no break-even point), congratulations! Your decision has been made for you and should be plainly apparent in your chart.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4k1Fz40rclqvZZLIMixwSQ.png" /><figcaption><strong>Your decision will depend on how your business can tolerate the short-term losses.</strong></figcaption></figure><h3><strong>TL;DR:</strong></h3><blockquote>The closer the break-even point, the stronger the argument for taking Option Two.</blockquote><blockquote>If no break-even point exists, choose Option One.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/683/1*8GspI-Fesy0pJKymDboTCQ.jpeg" /></figure><p><strong><em>If you’re curious…</em></strong></p><p>I work in the <strong>Product Intelligence</strong> team at Whispir, and my job is to help our teams make better decisions using my expertise in statistics, neuroscience and data visualisation.</p><p>This blog post came into being because recently some of our teams were considering a large technical project which might delay features, and we needed a quick and easy way to consider the options, and communicate the impact. I built a simple trade-off visualisation tool so that teams could “see” the long-term benefits (or lack thereof!) and have a conversation anchored by numbers, rather than gut-feel.</p><p>Admittedly, the method I’ve described here is pretty basic, and could easily be extended by adding non-linearities (to account for accelerating/decelerating productivity) or time-dependent confidence intervals (because near-term projections are more reliable than those further out).</p><p>However, in the spirit of Tim Peters’ “<a href="https://www.python.org/dev/peps/pep-0020/">The Zen of Python</a>”:</p><blockquote>If the implementation is hard to explain, it&#39;s a bad idea.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=82cd0b00072d" width="1" height="1" alt=""><hr><p><a href="https://medium.com/whispir/visualising-trade-offs-in-software-development-82cd0b00072d">Visualising Trade-Offs in Software Development</a> was originally published in <a href="https://medium.com/whispir">Whispir</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[It’s time for a new player in VR]]></title>
            <link>https://medium.com/@broak/its-time-for-a-new-player-in-vr-a14c70bb6d0e?source=rss-2bc3e20a5111------2</link>
            <guid isPermaLink="false">https://medium.com/p/a14c70bb6d0e</guid>
            <category><![CDATA[apple]]></category>
            <category><![CDATA[vr]]></category>
            <category><![CDATA[surveillance]]></category>
            <category><![CDATA[facebook]]></category>
            <category><![CDATA[privacy]]></category>
            <dc:creator><![CDATA[Brian Oakley, PhD]]></dc:creator>
            <pubDate>Wed, 19 Aug 2020 06:07:17 GMT</pubDate>
            <atom:updated>2021-09-07T06:01:51.834Z</atom:updated>
            <content:encoded><![CDATA[<h3>We need a new player in VR</h3><h4>Someone needs to play the ‘privacy’ card… and fast.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*fTO91guGcerlCrDS" /><figcaption>Photo by <a href="https://unsplash.com/@rev3n?utm_source=medium&amp;utm_medium=referral">Michał Kubalczyk</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>This morning, <a href="https://twitter.com/oculus/status/1295769648470753281?s=20">Oculus (a division of Facebook) announced a policy change</a> which will require new VR users to log in using their Facebook accounts, and existing users to log in by 2023, or face losing functionality.</p><p>And the internet has responded… <a href="https://gizmodo.com/facebook-just-told-oculus-users-to-go-get-fucked-1844765573">poorly</a>.</p><p>While the official line was that the policy change would be beneficial because it would “make it easier to find, connect, and play with friends in VR”, people called out the move for <a href="https://twitter.com/Grummz/status/1295903777359192064?s=20">what it clearly is</a> — a move to ensure the data collected from users in VR is more reliably tied to Facebook’s <a href="https://www.businessinsider.com/explainer-what-exactly-is-the-social-graph-2012-3?op=1">enormous social-graph</a>.</p><p>Other companies exist in the VR space (such as Valve, HTC, and HP, among others), however Facebook is currently the only major company offering a <em>stand-alone </em>headset, meaning you do not need to own a beefy PC to use it. This is an extremely important distinction to make because while other headsets may be more technically capable, they also cost up to three times as much, and that is without taking into account the price of a gaming computer.</p><p>For the vast majority of the public, the Oculus Quest is the only viable option to get into VR at a reasonable price, and with little technical know-how. Facebook knows this, and so while there are no viable competitors in the stand-alone VR space, they can risk alienating their consumer base with moves like this, because after all, <em>where else are you going to go</em>?</p><h4>A privacy-focused alternative</h4><p><a href="https://medium.com/@broak/vr-and-ar-tech-is-headed-down-a-dark-path-a242bd6c6ea5?source=friends_link&amp;sk=5562cc30009ffa286b091fa76132c996">Two weeks ago I wrote about</a> the concerns I have for VR if businesses like Facebook continue to focus on treating users not as customers, but as sources of data for advertisers. A hypothetical future Oculus Quest might gather so much bio-metric and behavioural data that it could make Facebook’s current recommendation and ad-serving algorithms look antiquated by comparison. This is the future that would likely eventuate if no privacy-focused competitor to the Quest arrives in the short term.</p><p><em>But what if there was another option?</em></p><p>At it’s core, the Oculus Quest is an Android-powered smartphone you strap to your face. It features a battery, ARM-based processor, internal storage, a display, cameras and some WiFi and Bluetooth modules. What separates the Quest from the smartphone world (other than its form-factor) is that if you want to play VR games on a stand-alone VR headset, your only option is to buy a headset from Facebook.</p><p>The obvious company to make a competitor for the Quest would be Apple — and <a href="https://www.roadtovr.com/apple-ar-vr-job-listings-new-products/">there is</a> <a href="https://vrscout.com/news/apple-patent-mr-headset-retina-projection/">evidence</a> <a href="https://www.roadtovr.com/apple-magic-leaps-former-employees/">they are</a> <a href="https://www.roadtovr.com/report-apple-nextvr-acqusition/">well on</a> <a href="https://www.roadtovr.com/report-apple-acquires-motion-capture-firm-ikinema/">their way</a> towards announcing one. They have a long history of focusing on consumer privacy, and have completely rejected using ad-based business models for any of their products. They also pride themselves on their design and ease-of-use, which is a major selling point of the Quest. Most importantly, however, is Apple’s track-record for building reliable hardware and well-supported app ecosystems. Just as games and apps are routinely ported between iOS and Android in the phone world, so too would we likely see stand-alone VR games ported to a future Apple-made VR headset. They have even started making moves into the gaming space with their Apple Arcade subscription service, which hosts hundreds of games exclusive to the iOS platform.</p><p>There will be downsides. A headset that doesn’t rely on selling data and serving ads will likely cost more. But what is important is that consumers have <em>options</em>. I’m not saying that Apple is the <em>only</em> company capable of competing with Facebook in standalone VR, but if users want a privacy-focused alternative sooner rather than later, they <em>might be our best bet</em>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a14c70bb6d0e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why XR tech could lead us down a dark path]]></title>
            <link>https://medium.com/@broak/vr-and-ar-tech-is-headed-down-a-dark-path-a242bd6c6ea5?source=rss-2bc3e20a5111------2</link>
            <guid isPermaLink="false">https://medium.com/p/a242bd6c6ea5</guid>
            <category><![CDATA[vr]]></category>
            <category><![CDATA[privacy]]></category>
            <category><![CDATA[facebook]]></category>
            <category><![CDATA[capitalism]]></category>
            <category><![CDATA[surveillance]]></category>
            <dc:creator><![CDATA[Brian Oakley, PhD]]></dc:creator>
            <pubDate>Fri, 07 Aug 2020 05:44:47 GMT</pubDate>
            <atom:updated>2021-09-08T03:18:22.910Z</atom:updated>
            <content:encoded><![CDATA[<h4>Wearable panopticons for everyone!</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*TOofTtSmmCuwjJzI" /><figcaption>Photo by <a href="https://unsplash.com/@ciabattespugnose?utm_source=medium&amp;utm_medium=referral">Lucrezia Carnelos</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Companies have a fiduciary duty to their shareholders to stay profitable and continually grow. This fact is not necessarily a bad thing, however a combination of factors – business model, leadership, and morality – determine the externalities that will ultimately burden the public.</p><p>Right now, when the general public think of AR, they likely think of “Pokemon GO”. This isn’t surprising, given the clear absence of killer apps in the space. Even the latest iPad Pro, featuring a combination of computer vision and LIDAR, provides a merely <em>acceptable </em>level of presence with rudimentary scene recognition and object occlusion. And even then, the current hardware is at least a year ahead of meaningfully useful software. End users don’t care <em>how </em>their device does what it does, they only care whether what it does is valuable to them. But AR tech is moving quickly, and with companies like Facebook (with their Oculus division) and Apple investing vast amounts into this space, we’re going to see rapid developments unlike any seen since the first iPhone. Whether these developments end up being a net-positive or negative for our society is yet to be seen.</p><h3>Two competing business models</h3><p>If you want to start a tech company, you have two business models to choose from: sell <em>products</em>, or sell <em>data</em>.</p><p>Selling products is hard — it is difficult to predict what people want. It’s even more difficult to navigate the geopolitical labyrinth required to produce these products cheaply, quickly and reliably at scale. But the biggest challenge is to keep the market coming back for more (or in <a href="https://en.wikipedia.org/wiki/Juicero">Juicero’s case</a>, buying the damn thing in the first place).</p><p>Selling data is somewhat easier. What your public-facing “product” actually <em>is </em>doesn’t matter, as long as you gather<em> behavioural data</em>. In the words of <a href="https://www.bookdepository.com/Age-Surveillance-Capitalism-Shoshana-Zuboff/9781781256855">Shoshana Zuboff</a>, users are the rocks from which the raw material of behavioural data is mined. The <em>real </em>products these companies sell are the prediction algorithms that are highly sought after by anyone looking to market a their new soft drink, diet fad or political ideology. The difficult thing in this case of these businesses, is <em>how do you keep people using</em> (and therefore providing data)? Here we see the <a href="https://www.humanetech.com/what-we-do#problem">same techniques recycled time and time again</a> — never-ending scrolling feeds, strategic randomisation, recommendation-algorithms, and more, all derived from psychological studies about addiction and eerily similar to the <a href="http://sitn.hms.harvard.edu/flash/2018/dopamine-smartphones-battle-time/"><em>strategies used by the gambling industry.</em></a> The goal is to keep you scrolling, liking, commenting, subscribing and engaging — because <em>this is where the money is</em>.</p><p>As of today, augmented reality is essentially in its infancy. Neither business model has used this technology to its full potential, though both are well suited to. The most robust headset as of today is perhaps the Microsoft HoloLens 2 — a head-mounted display aimed at industry applications like construction and manufacturing, and sells very few units compared to traditionally consumer-focused products. Presumably, the high unit cost covers the R&amp;D costs + manufacturing, logistics and profit (if any). This is essentially the first model — <em>selling a product</em>. An exemplary of this model is Apple, a company that repeatedly displays its aptitude at producing high-margin electronics, paired with frighteningly effective marketing that ensures each product (see: iPod, iPhone, iPad, AirPods, etc) become a middle-class staple and veritable status symbol. A successful AR headset using this model <em>will </em>be expensive. It <em>will </em>be highly sought after, but it <em>will</em> remain out of reach for many people — such is the fundamental downside of such a business model.</p><p>The alternative is a cheaper headset, one much more accessible to the vast majority of the public (especially in developing nations), where the costs are subsidised by the sheer amount of data they will collect. However, while the Facebook, YouTube, and Netflix apps on your phone may <em>infer </em>your preferences, dislikes and motivations from what you tap, swipe and hover over, AR and VR headsets are <em>fundamentally different</em>. A fully featured headset will <a href="https://en.wikipedia.org/wiki/Foveated_rendering">track your gaze to ensure what you see is as realistic as possible</a>, likely also collecting information about your <a href="https://en.wikipedia.org/wiki/Mydriasis">pupil size</a> over time. It is not unreasonable to expect such a headset to connect to peripherals like your smartwatch which will gather periodic bio-metrics such as <a href="https://support.apple.com/en-au/guide/watch/apda88aefe4c/watchos">heart rate</a>, <a href="https://www.samsung.com/au/support/mobile-devices/measure-bp-on-galaxy-watch-active/">blood pressure</a> or <a href="https://www.wareable.com/health-and-wellbeing/zenta-vinaya-specs-price-features-release-date">galvanic skin response</a>. Furthermore, an AR headset requires cameras facing <em>outwards </em>to be able to place virtual elements in space, and ensure they are coherent and aligned with your surroundings. This combination of technologies results in a device that can sample, at multiple times per second, your behavioural and physiological response to any naturalistic stimulus — or in layman’s terms, <em>read your thoughts about everything you see</em>*.</p><p>Let’s imagine you’ve got the latest set of smart glasses from Facebook.</p><p>Say you were wearing these glasses as you walked down an aisle of a store. Every subtle glance at an object would be recorded, quantified and correlated with your bio-metrics and past behaviour. Stare at a pack of <a href="https://en.wikipedia.org/wiki/Tim_Tam">Tim-Tams</a> for a split-second longer than you otherwise would have, and the headset correlates that statistically significant increase in <em>gaze duration</em> with a simultaneous <em>micro-dilation of your pupils</em>, and tiny, but meaningful increases in<em> heart rate</em> and <em>skin conductivity</em>. A few short seconds later you get an AR notification of a discount on Tim-Tams if you scan your loyalty card at the checkout (yet another form of behavioural data collection). In essence, what we have is a <em>nightmarishly </em>effective tool for behavioural modification — <a href="https://www.wired.co.uk/article/the-age-of-surveillance-capitalism-facebook-shoshana-zuboff">a task our smartphones currently serve</a>, but on steroids.</p><h3>It’s going to happen whether we like it or not</h3><p>I described these two alternative business models earlier as if they were somehow mutually exclusive — but we know this is not the case. The world we live in is rife with systemic inequalities and this is yet another example. Those fortunate enough to be able to afford the most expensive options are likely to own devices or subscribe to service tiers that provide increased privacy in the form of greatly reduced data collection and ad-targeting. You could buy an iPhone if you can afford the premium, or you can save money but accept Google’s default settings which keep track of your <a href="https://www.wired.com/story/google-tracks-you-privacy/">location, browsing, search and messaging histories</a>. You could pay a monthly fee for your apps, or instead use the free tier, which is paid for by tracking your searches and movements across the web, and targeting you with advertisements .</p><p>I’m not convinced that there is much we can do to ensure that such a future does not eventuate. We could forgo modern technology entirely, but that is becoming increasingly unrealistic. Our devices are gradually supplanting our wallets, car keys, ID’s and passports. They have already made themselves invaluable to those with disabilities, and have served to democratize access to information in underprivileged and marginalised populations. In ten to fifteen years, it might be more unusual to own a pair of glasses that<em> don’t</em> provide turn-by-turn navigation, automatic text translation and puppy face-filters.</p><p>At the end of the day, the companies in this space will do what is best for their shareholders, even if it means becoming ever more invasive.</p><p><em>*</em><a href="https://www.theverge.com/2019/7/16/20697123/elon-musk-neuralink-brain-reading-thread-robot"><em>I’m exaggerating just a bit, unless of course Elon has anything to say about it…</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a242bd6c6ea5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Facebook’s Oculus is blazing a fresh path]]></title>
            <link>https://medium.com/@broak/oculus-aims-for-comfort-over-power-6a9df3102b1f?source=rss-2bc3e20a5111------2</link>
            <guid isPermaLink="false">https://medium.com/p/6a9df3102b1f</guid>
            <category><![CDATA[gaming]]></category>
            <category><![CDATA[oculus]]></category>
            <category><![CDATA[virtual-reality]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[consumer-technology]]></category>
            <dc:creator><![CDATA[Brian Oakley, PhD]]></dc:creator>
            <pubDate>Tue, 19 May 2020 09:40:39 GMT</pubDate>
            <atom:updated>2021-09-08T03:17:28.781Z</atom:updated>
            <content:encoded><![CDATA[<h4>Choosing comfort over horsepower</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XFB6SIlqqpqQZdbpmA29Ag.jpeg" /><figcaption>Photo by <a href="https://www.pexels.com/@thisisengineering?utm_content=attributionCopyText&amp;utm_medium=referral&amp;utm_source=pexels">ThisIsEngineering</a> from <a href="https://www.pexels.com/photo/woman-in-white-tank-top-using-black-laptop-computer-with-vr-headset-3861458/?utm_content=attributionCopyText&amp;utm_medium=referral&amp;utm_source=pexels">Pexels</a></figcaption></figure><p>This week Facebook <a href="https://www.oculus.com/blog/celebrating-the-one-year-anniversary-of-oculus-quest-and-rift-s/">announced it would start accepting productivity apps in the Oculus store</a> for their popular Oculus Quest virtual reality headset. Until now, their store has primarily served games and video-streaming applications, but their latest announcement points to some of their broader goals — a push for their flavor of VR to be <em>the next computing platform</em>, which could not have come at a better time given the current circumstances.</p><p>The global pandemic has forced many people to swiftly adapt to a life of working, studying and socializing remotely. So far this meant video conferencing has become a primary mode of communication (<a href="https://www.marketwatch.com/story/zoom-microsoft-cloud-usage-are-rocketing-during-coronavirus-pandemic-new-data-show-2020-03-30">a boon for companies such as Zoom</a>), though how effective this is as a substitute for the “real thing” is debatable. Virtual reality (and it’s close cousin augmented reality) offer an experience of ‘presence’ not possible with the traditional computing platforms most commonly used today. <a href="https://www.youtube.com/watch?v=RCB_mfGmh9w&amp;t=3179s">This is the fundamental pitch Mark Zuckerberg made at Oculus Connect 6 in 2019</a>, where he explains a primary goal of Oculus is to create “a future that defies distance” — an idea so eerily fortuitous I’m surprised I haven’t heard his name being thrown around with the numerous COVID-19-related conspiracy theories we’ve been seeing lately. In any case, the fact that Oculus is aiming to be the next computing platform is telling, and highlights a fundamental distinction between their strategy, and that of their closest competitor in the virtual reality space, Valve.</p><h3><strong>Low-friction is key.</strong></h3><p>Both Oculus and Valve make HMDs (head-mounted displays), and both run digital distribution platforms serving software to be bought and used on their hardware. At present, the vast majority of this software is aimed at gamers and tech early-adopters, which makes sense given the origins of both of these companies. But while Valve has continued to make their headsets better for the gaming market, Oculus has been gradually pivoting their hardware away from it’s PC-reliant roots. The first generation of headsets required both a powerful PC, and external tracking beacons to produce a convincing, six-degrees-of-freedom (room-scale) experience — and this is still the case for Valve’s latest — the Index (albeit with more accurate tracking, higher display refresh-rates and a larger field-of-view).</p><p>In contrast, Oculus has traded the tracking accuracy afforded by external beacons for the “good-enough” approach of integrated tracking cameras, simplifying the set-up experience significantly. While Valve’s flagship headset still requires a beefy gaming PC to work, the Oculus Quest headset relies on integrated mobile hardware, removing the need for a PC entirely. Again, this represents a trade-off — the processor is much less powerful, but it opens up the ability for the headset to work almost anywhere wirelessly. More importantly for Facebook, it expands the potential consumer base significantly — a clear priority, repeatedly mentioned during Oculus Connect.</p><p>These hardware trade-offs all result in one major advantage for Oculus — <em>reducing friction</em>. You put the headset on, and it turns on immediately. You can be standing, sitting, lying in bed, it doesn’t matter — the headset senses you are wearing it and switches on immediately, no cables to plug in or PC to boot up. Recently they even updated the software so you don’t need to use the bundled controllers to navigate the interface, a potential point of friction for users unfamiliar with the buttons, thumb-sticks and triggers commonly featured on game controllers. Instead, the headset recognizes your hands automatically, and you can browse the web, watch videos, and in the near future, play certain games without ever touching the controllers. The aim to reduce friction is a strategy that ensures the product is <em>accessible to the broadest demographic </em>possible, and it has worked wonders for companies such as Apple (the iPod, iPhone and AirPods), Amazon (Alexa, Kindle), and perhaps most recently, Zoom (with their popular, no-install video conferencing platform), just to name a few. These products are <em>simple to use by design</em>, and as a result they have each become ubiquitous.</p><h3><strong>Cheaper, Lighter, Faster.</strong></h3><p>Creating a new computing platform is not easy, even if you have billions to pour into the project (just ask Microsoft and their disastrous Windows Phone operating system). At the core of this is a chicken-and-egg problem. To entice consumers into buying your hardware, you need great software and experiences — but to attract the developers needed to build this software, you need a large active consumer base. For years VR has struggled with this problem (and it arguably still is struggling), but Oculus appears to have finally overcome the hurdle somewhat. The all-in-one Oculus Quest costs $399, a far-cry from the $999 Valve Index which also requires a gaming PC to use. The Quest features a large library of software available for purchase, <a href="https://www.roadtovr.com/oculus-quest-app-revenue-100-million-sales-first-year/">has sold over $100 million worth of games and apps in its first year</a> (a great sign for developers) and the headsets are selling “as fast as [they] can make them”, <a href="https://investor.fb.com/investor-events/event-details/2019/Facebook-Q2-2019-Earnings/default.aspx">according to Mark Zuckerberg, himself</a>.</p><p>Oculus is rumored to be releasing an updated version of the Quest headset at some point in 2020 (though now likely to be in 2021, given the global pandemic), with improvements focusing on primarily on comfort. This is the trajectory they need to stay on if they want this new platform to succeed. The next, and future Quest’s <a href="https://www.oculus.com/blog/half-dome-updates-frl-explores-more-comfortable-compact-vr-prototypes-for-work/">will get increasingly small, light and more comfortable</a>. Many of the most popular games in VR (and crucially, those with the broadest appeal) such as Beat Saber have users sweating within half an hour of use. VR is fundamentally a more active experience, and so while Oculus was ahead of the curve by axing the tether to the PC (the wireless experience is much more comfortable and user-friendly), the faster they get to their projected future light-weight, glasses-like form-factor, the better — even if this means sticking with lower-powered processing in the short-term.</p><p>Increasing comfort and lowering the barrier to entry will result in increased consumer interest and increased sales, therefore producing economies of scale which will result in lower prices (not to mention the potential for new streams of data collection and advertising revenue for Facebook — a somewhat worrying topic I will cover in a future article). Lower prices tend to equal more consumers, which tend to equal more (and better) software. The announcement of new productivity-focused applications clearly suggest Oculus has its eye on much broader use-cases for their VR platform, but these improvements in comfort can not come quickly enough.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6a9df3102b1f" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>