<?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 James Cham on Medium]]></title>
        <description><![CDATA[Stories by James Cham on Medium]]></description>
        <link>https://medium.com/@jamescham?source=rss-86a8b1c606a8------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*ccfucjc5EaSYmCMZdVixMg.jpeg</url>
            <title>Stories by James Cham on Medium</title>
            <link>https://medium.com/@jamescham?source=rss-86a8b1c606a8------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 10:00:38 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@jamescham/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[Cosmic Humanity]]></title>
            <link>https://medium.com/@jamescham/cosmic-humanity-2bcfb527bfdd?source=rss-86a8b1c606a8------2</link>
            <guid isPermaLink="false">https://medium.com/p/2bcfb527bfdd</guid>
            <category><![CDATA[holidays]]></category>
            <category><![CDATA[space-travel]]></category>
            <category><![CDATA[the-long-now]]></category>
            <category><![CDATA[asteroids]]></category>
            <category><![CDATA[history]]></category>
            <dc:creator><![CDATA[James Cham]]></dc:creator>
            <pubDate>Sat, 05 Nov 2022 02:23:08 GMT</pubDate>
            <atom:updated>2022-11-08T05:59:14.810Z</atom:updated>
            <content:encoded><![CDATA[<h3>Humanity didn’t do anything cosmic until September 2022.</h3><p>We’ve managed to travel out of our atmosphere and see the Whole Earth. But it took another 60 years for us to actually change the course of an object not of this earth.</p><p>That’s because NASA launched a satellite that hit an asteroid that circled around another asteroid and changed its course. The critical thing is that we did it on purpose. We actually made a dent in the universe!</p><p>Now what we do with this ability is up to us! We could accidentally push the Earth out of orbit and doom ourselves, or create a Dyson Sphere and flourish in unimaginable ways.</p><p>Ten thousand years from now, nobody is going to remember most of what happened in 2022. But if there are people around, they will remember DART. Because for better or worse, that was the first time that Humans made a Cosmic impact.</p><h3>Notes</h3><p>Please make comments or send them to @jamescham on Twitter! I’m curious what people think.</p><p>The first one to point this out was Rodney Brooks.</p><p><a href="http://bit.ly/cosmic-humanity">Source</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2bcfb527bfdd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[James Fallows]]></title>
            <link>https://medium.com/@jamescham/james-fallows-ec3cefc49f30?source=rss-86a8b1c606a8------2</link>
            <guid isPermaLink="false">https://medium.com/p/ec3cefc49f30</guid>
            <dc:creator><![CDATA[James Cham]]></dc:creator>
            <pubDate>Mon, 08 Aug 2022 06:20:45 GMT</pubDate>
            <atom:updated>2022-08-08T06:20:45.364Z</atom:updated>
            <content:encoded><![CDATA[<p>(In honor of <a href="https://medium.com/u/adeddd83f452">Tony Stubblebine</a> taking over Medium, I’m looking for as many excuses as possible to write stories, even if they are just quick sketches and notes that I might elaborate on in the future.)</p><p>I just read this <a href="https://fallows.substack.com/p/august-recommendations">substack post</a> from James Fallows:</p><blockquote>American <em>politicians</em> and the machinery of national politics — especially the money-raising part, and the news-environment part — are indeed more and more polarized, as everyone <a href="https://fallows.substack.com/p/a-lucky-country-on-thin-ice">including me</a> has noted.</blockquote><blockquote>It’s a natural leap, but a mistaken one, to assume that every detail of national life is flattened and polarized in the same way.</blockquote><blockquote>This temptation to oversimplify is all the greater when Americans know the parts of the country where they don’t live only through red/blue Electoral College maps, or “guy in a diner” / “angry people at a rally” set-piece stories.</blockquote><blockquote>It’s a big, diverse country, which has dire problems, but most of whose people are less inflamed and crazy than you’d think by watching the worst of them on the news.</blockquote><p>And it was a reminder of the role that James Fallows played in my life. I’ve only met him once but his columns in The Atlantic on technology fundamentally shaped my view of what it was like to think about software. He’s the first who helped me imagine what a tool for thinking should be like.</p><p>He’s even more interesting as an example of a big thinker who was consistent right about the big picture, even when it must have cost him personally: whether it was about the future of journalism, Japan, or the war in Iraq.</p><p>And he’s been in the middle of some of the biggest stories of the last 50 years, from seeing the rise and the fall of the liberal establishment while working for Charles Peter, Ralph Nader and Jimmy Carter to the seeing the rise of tech. He even worked at Microsoft for a while. He’s always felt sharp and humane to me, and it shaped the way I think about our country and the world.</p><p>I hope there’s someone working on a big biography of him. It would be fascinating!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ec3cefc49f30" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Is the metaverse really about space or time?]]></title>
            <link>https://medium.com/@jamescham/is-the-metaverse-really-about-space-or-time-541f94f3fbbb?source=rss-86a8b1c606a8------2</link>
            <guid isPermaLink="false">https://medium.com/p/541f94f3fbbb</guid>
            <dc:creator><![CDATA[James Cham]]></dc:creator>
            <pubDate>Fri, 05 Aug 2022 18:48:15 GMT</pubDate>
            <atom:updated>2022-08-05T18:48:15.324Z</atom:updated>
            <content:encoded><![CDATA[<p>I predict that I’m going to be thinking a lot about <a href="https://www.oreilly.com/radar/the-metaverse-is-not-a-place/">Tim O’Reilly’s recent piece on the metaverse</a>, and I’m going to be jotting down notes in Medium in honor of <a href="https://medium.com/u/adeddd83f452">Tony Stubblebine</a> joining the company as CEO.</p><blockquote>Using this metaphor, we see the metaverse as a continuation of a line that passes through messaging and email to “rendezvous”-type social apps like Zoom, Google Meet, Microsoft Teams, and, for wide broadcast, Twitch + Discord. This is a progression from text to images to video, and from store-and-forward networks to real time <strong>(and, for broadcast, “stored time,” which is a useful way of thinking about recorded video)</strong>, but in each case, the interactions are not place based but happening in the ether between two or more connected people.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=541f94f3fbbb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Situated AI: Taking the human component seriously]]></title>
            <link>https://medium.com/@jamescham/situated-ai-taking-the-human-component-seriously-c02ccb9da8bf?source=rss-86a8b1c606a8------2</link>
            <guid isPermaLink="false">https://medium.com/p/c02ccb9da8bf</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[economics]]></category>
            <dc:creator><![CDATA[James Cham]]></dc:creator>
            <pubDate>Wed, 13 Jul 2022 00:28:37 GMT</pubDate>
            <atom:updated>2022-07-13T00:28:37.030Z</atom:updated>
            <content:encoded><![CDATA[<p>The best summary of the state of AI ethics is Brian Christian’s new book <a href="https://brianchristian.org/the-alignment-problem/">The Alignment Problem</a>. He spent four years talking to the leading researchers and practitioners thinking about AI right now, summarizing their insights and making ties between the limitations and implications of AI models. The book highlights limitations of the current AI ethics conversation in his last chapter.</p><p>I think it points to a bigger limitation to the discussion.</p><p>The limitation is that we take AI models too seriously, treat humans as static unchanging abstractions, and think of the algorithms as the center. But AI models only exist in relation to humans. We build, inform, enforce, ignore, obey, and manipulate them. Those interactions are the actual center of the most interesting AI ethics challenges.</p><p>A few thinkers are talking about this interaction — taking humans seriously and also understanding the strengths and limitations of AI models.</p><p>Here’s a running list. I’m looking for other examples so please send them my way in the comments, via <a href="http://www.twitter.com/jamescham">Twitter</a>, or at the <a href="http://bit.ly/situated-ai">Google doc</a>.</p><ul><li><strong>How AIs get legitimacy: </strong><a href="https://www.judgingmachines.com/">How Humans Judge Machines</a> (<a href="https://twitter.com/jamescham/status/1311073623667474432?s=20">summary</a>). Forthcoming book by economics professor Cesar Hidalgo. A very clever study on human moral intuitions around AI. When do humans blame machines more or less? You can think of this as a “moral census” of American attitudes towards AI. Presents opportunities for when AI models have more or less legitimacy. Will change over time, and in addition to comparing human vs machines, would be interesting to add how humans judge organizations.</li><li><strong>AI should be viewed as a bureaucracy rather than God-like judge: </strong><a href="https://hci.stanford.edu/publications/2019/streetlevelalgorithms/streetlevelalgorithms-chi2019.pdf">Street Level Algorithms</a> (<a href="https://twitter.com/jamescham/status/1125504490361184256">summary</a>). Ali Alkhatib and Michael Bernstein. CS paper that deserves more attention. Looking at how laws are actually enforced, and corrected by people on the ground, and comparing that with the way that AI models are actually enforced. AI models are often blamed for being wrong but they are actually easier to correct than laws, and when corrected, they are centralized.</li><li><strong>People cheat and fool AIs (social science view): </strong><a href="https://crookedtimber.org/2019/11/25/seeing-like-a-finite-state-machine/">Seeing Like a Finite State Machine</a>. Henry Farrell. Smart blog post (and reference to James C Scott’s Seeing Like a State) by a professor of International Relations. “There is a very plausible set of mechanisms under which machine learning and related techniques may turn out to be a disaster for authoritarianism, reinforcing its weaknesses rather than its strengths, by increasing its tendency to bad decision making, and reducing further the possibility of negative feedback that could help correct against errors.” AI models are profoundly conservative: they reflect the world that we tell them about. And in that way, left unchecked, they can be manipulated and given inaccurate information.</li><li><strong>Who gets to decide? </strong><a href="https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3517287">The Allocation of Decision Authority to Human and Artificial Intelligence</a>. Joshua Gans, Susan Athey. The principal trades off an AI’s more aligned choice with the need to motivate the human agent to expend effort in learning choice payoffs. When agent effort is desired, it is shown that the principal is more likely to give that agent decision authority, reduce investment in AI reliability and adopt an AI that may be biased. Organizational design considerations are likely to impact on how AI’s are trained.</li><li><strong>People cheat and fool AIs (CS view): </strong><a href="https://a.co/9BW8ri6">Ethical Algorithms</a>, Chapter 3. Michael Kearns, Aaron Roth. “The lesson that when an app is mediating or coordinating the preferences of its users (as opposed to simply using their data for some other purpose, such as building a predictive model), the algorithm design must specifically take into consideration how users might react to its recommendations — including trying to manipulate, defect, or cheat.”</li><li><strong>Economics of AI:</strong> <a href="https://www.predictionmachines.ai/">Prediction Machines</a>. Ajay Agrawal, Avi Goldfarb, Josh Gans. Popular business book by 3 economists. AI models aren’t necessarily better. But they are much, much cheaper and consistent than the human-based judgements.</li><li><strong>People power models: </strong><a href="https://ghostwork.info/">Ghost Work</a>. Mary L. Gray and Siddharth Suri. By looking closely at the people who are providing the judgments used to build AI models and taking them seriously, you can see ways to improve those judgements and see the boundaries of the data being captured.</li><li><a href="https://twitter.com/yudhanjaya/status/1309885853544587264">The State Machine</a>. Yudhanjaya Wijeratne. Short fiction on an AI run world.</li><li><a href="https://twitter.com/fromedome/status/1314974107444813825">Adapting to the algorithm</a>. Kyle Chayka. “As I write more about how algorithms influence culture, I find myself using the word “adaptation” quite often. I analyze how creators and content have to “adapt to the algorithm” in order to find an audience on a platform.”</li><li><a href="https://www.techatbloomberg.com/blog/a-conversation-about-human-computation-with-bloombergs-human-computation-architect-walter-s-lasecki/">A Conversation about Human Computation</a>. Walter Lasecki. “Two things are exciting to me. First is the ability to look at how the traditional organizational structure intersects with how we think about human computation at the task level — that is, incorporating both the micro and macro structures into the design of “hybrid intelligence” systems. The second thing is real-time applications of human computation methods — thinking about ways to create new systems and tools to provide human-in-the-loop answers within seconds of the data becoming available.”</li><li><a href="https://en.wikipedia.org/wiki/Her_(film)">Her</a>: You should probably add the film “Her”2013 as a reference to human — AI interface. It’s contributed to my vision for <a href="http://timeos.co/">http://timeOS.co</a> and steered my deep thinking about the potential of a Personal OS. <a href="https://twitter.com/rmishiev/status/1405278067413508100">@rmishiev via Twitter.</a></li><li><a href="https://www.dropbox.com/s/3wy36804jhlc3qa/Farrell-Fourcade%20High-Tech%20Modernism.pdf?dl=0">High Tech Modernism</a> Henry Farrell and Marion Fourcade. “Instead, we claim that algorithms are much better understood as a manifestation of hierarchy than as an extension of market competition.”</li><li><a href="https://www.lightspeedmagazine.com/fiction/the-perfect-match/">The Perfect Match</a>, Ken Liu. “You thought Centillion was just an algorithm, a machine. But now you know that it’s built by people — people like me, people like you.” via Ajay Agrawal</li></ul><p>More links and running notes at the <a href="http://bit.ly/situated-ai">Google doc</a>.</p><p>(I’m posting this because my friend Tony is now running Medium so I figured I’d take a look at publishing. Still a great experience and looking forward to seeing what happens now!)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c02ccb9da8bf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I love Asimov but I would stop reading here. I wish you started with the zoox story?]]></title>
            <link>https://medium.com/@jamescham/i-love-asimov-but-i-would-stop-reading-here-i-wish-you-started-with-the-zoox-story-b5aaca666ae2?source=rss-86a8b1c606a8------2</link>
            <guid isPermaLink="false">https://medium.com/p/b5aaca666ae2</guid>
            <dc:creator><![CDATA[James Cham]]></dc:creator>
            <pubDate>Wed, 01 Dec 2021 20:45:11 GMT</pubDate>
            <atom:updated>2021-12-01T20:45:11.172Z</atom:updated>
            <content:encoded><![CDATA[<p>I love Asimov but I would stop reading here. I wish you started with the zoox story?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b5aaca666ae2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Do customers respond to last mile? I don’t really get it.]]></title>
            <link>https://medium.com/@jamescham/do-customers-respond-to-last-mile-i-dont-really-get-it-323bc15835b3?source=rss-86a8b1c606a8------2</link>
            <guid isPermaLink="false">https://medium.com/p/323bc15835b3</guid>
            <dc:creator><![CDATA[James Cham]]></dc:creator>
            <pubDate>Wed, 01 Dec 2021 20:43:24 GMT</pubDate>
            <atom:updated>2021-12-01T20:43:24.117Z</atom:updated>
            <content:encoded><![CDATA[<p>Do customers respond to last mile? I don’t really get it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=323bc15835b3" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Notes from Founder Conversations]]></title>
            <link>https://medium.com/@jamescham/notes-from-founder-conversations-666b896b39b9?source=rss-86a8b1c606a8------2</link>
            <guid isPermaLink="false">https://medium.com/p/666b896b39b9</guid>
            <dc:creator><![CDATA[James Cham]]></dc:creator>
            <pubDate>Thu, 19 Mar 2020 06:03:24 GMT</pubDate>
            <atom:updated>2020-03-19T06:03:24.250Z</atom:updated>
            <content:encoded><![CDATA[<p>A personal running list.</p><p><em>During Early Days of Corona Virus before human and economic impact was totally understood:</em></p><p>“But this is the time when you also win loyal customers for the next decade. And those will not just be companies but individual users who will want to take you to their next job.”</p><p>“2008 was the year that Airbnb took off, and when hotels approached Booking.com to list inventory (which they then never took off).”</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=666b896b39b9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Your Next Management Guru]]></title>
            <link>https://medium.com/@jamescham/were-waiting-for-the-peter-drucker-of-machine-intelligence-9b674191b420?source=rss-86a8b1c606a8------2</link>
            <guid isPermaLink="false">https://medium.com/p/9b674191b420</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[venture-capital]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[James Cham]]></dc:creator>
            <pubDate>Mon, 25 Sep 2017 18:25:56 GMT</pubDate>
            <atom:updated>2017-10-09T18:38:41.072Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TBQ95z9Y0QhfXID6QBvaKg.png" /><figcaption>Illustration by <a href="https://medium.com/u/7933350e9eb6">Danny Yee</a></figcaption></figure><h3>We’re Waiting for the Peter Drucker of Machine Intelligence</h3><p>AI will change everything, yeah yeah. We’ve heard it, and it’s so true we no longer hear it.</p><p>Here’s the next big one: how leaders lead companies.</p><p>The old way was “leader as genius,” where a company’s job is to find a solution to a problem, and a leader decides on those solutions (mostly other people doing work), and introduces a process to solve the problem reliably (e.g., a workflow to make a car).</p><p>That way evolved to empower people “at the edge” to propose their own solutions (and sometimes decide on them), à la how Facebook is run vs. how Ford was run.</p><p>Peter Drucker was this way’s prophet, bringing humanity to the center of the organization.</p><p>Software accelerated this old way because, instead of people repeating these workflows, machines could automate the workflow’s dullardwork.</p><p>Software was merely the translation of people instructions into machine instructions. (Drucker even called computers <a href="https://www.highly.co/hl/4vOquoUrFqm5J0">total morons</a>, because they could not make decisions, only implement them.) Software could spread information, creating an “<a href="https://books.google.com/books?id=jhEDvba0I-UC&amp;lpg=PA129&amp;ots=oQtiob6NrH&amp;dq=drucker%20information%20based%20organization&amp;pg=PA129#v=onepage&amp;q=drucker%20information%20based%20organization&amp;f=false">information-based organization</a>” so people can make better decisions — but, in Drucker’s view the software never decides.</p><p><a href="https://machinelearnings.co/a-humans-guide-to-machine-learning-e179f43b67a0">Machine Intelligence</a> changes that. It enables software that remembers not only past data but also the results of past decisions — and distills them into a model. This is the secret concept for future AI management gurus to understand. Models go beyond automating workflows, they automate decisions.</p><p>In <a href="http://bloombergbeta.com/">our fund</a>, where we have been investing exclusively in the future of work from our first check, we see a new kind of company leader emerging. She does not see her role in just the old way — deciding and translating decisions into software and people workflows, or even as nurturing her people to do that. She sees her role in a new way — creating the conditions (in talent and data) to build the right models that can make better decisions than any of us, and — critically — understanding when not to trust models and when to update them. This creates an organization that can make decisions more rapidly (and improve more rapidly) than any pointy-haired boss could through tedious training or tenuous anecdotes.</p><p>The previous generation understood how to lead people and how to harness IT. The new generation will understand how to manage models, which are the flux capacitor of making software go beyond workflows to decisions.</p><p>If a top 1% CEO today understands how software applications get built and how that changes the way she manages, a top 1% CEO in the near future will understand how models get built and <a href="https://hbr.org/2017/07/how-ai-will-change-the-way-we-make-decisions">how that changes the way she builds her organization</a>.</p><p>A whole way of thinking about leadership, and about the economics of companies, with as many maxims and guidelines as the old way, is to come.</p><p>We’re waiting for the Peter Drucker of Machine Intelligence.</p><p>Join <strong>30,000+ people</strong> who read the weekly <a href="http://subscribe.machinelearnings.co/?utm_source=machine_learnings&amp;utm_medium=blog&amp;utm_campaign=guest_post&amp;utm_content=james_cham">🤖Machine Learnings🤖 </a>newsletter to understand how AI will impact the way they work and live.</p><p><em>Written with </em><a href="https://medium.com/u/9ed065b7bce9"><em>Roy Bahat</em></a><em>. For more on how Machine Intelligence is changing management, read (1) our collaborators at the University of Toronto, Ajay Agrawal, </em><a href="https://medium.com/u/bbc5265c9f51"><em>Joshua Gans</em></a><em>, and Avi Goldfarb on the </em><a href="http://bit.ly/2wKmu1c"><em>core economics of AI</em></a><em>; (2) MIT’s Andrew McAfee and Erik Brynjolfsson’s recent book </em><a href="http://bit.ly/2jWOucb"><em>Machine, Platform, Crowd: Harnessing Our Digital Future</em></a><em>; and (3) Stanford’s Susan Athey on </em><a href="http://bit.ly/2yAJWLI"><em>machine learning and economics</em></a><em>. Or watch the </em><a href="http://bit.ly/2xxrKUw"><em>recent NBER conference on the Economics of AI</em></a><em>.</em></p><p><em>One day, someone might say that Peter Drucker was the pre-Machine Intelligence version of these gifted thinkers.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9b674191b420" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[MI Landscape 2016]]></title>
            <link>https://medium.com/@jamescham/mi-landscape-2016-22a55d2a0982?source=rss-86a8b1c606a8------2</link>
            <guid isPermaLink="false">https://medium.com/p/22a55d2a0982</guid>
            <category><![CDATA[venture-capital]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[James Cham]]></dc:creator>
            <pubDate>Tue, 15 Nov 2016 20:15:22 GMT</pubDate>
            <atom:updated>2016-11-15T20:38:13.287Z</atom:updated>
            <content:encoded><![CDATA[<p>Presentation given at AI World, November 8 2016 with <a href="https://medium.com/u/509cf8b26804">Shivon Zilis</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QkVceF5LoXHmhjEw9Yntkg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fQ9_Sqtinrusa5CTpY4swA.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ypgHUN8WpILW1J0ZwTM6xw.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XN98se4MSlqEff5DVGr8rg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EAL5FBa0RBVxrKcynLAd2w.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nU6Nau7rLh-QszvQzoc4ag.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2-tkeVU5rGuK5pp0TyzuAg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MpJjnqPLQGUNtMW8uJ5fUQ.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-bkenMeFo5f4gBhXnC1anw.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*780QMqM6TVsAPeFqfSY9ew.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*H7X7IO1gVnJdac-hkYw8Zg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kO4x-SCtB-jQOWjj9XCOmA.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EcvJKwoLaOON7nlDP7eAug.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AM_fJ7FLp7mDHndIZA1d9w.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zJWApXl2u8pfKUmnn2wdig.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RrqrNcoXX8n6t1RjqS4oig.jpeg" /><figcaption>How <a href="http://www.textio.com/">Textio</a> works.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*A7X3eF-f0DlSO7K1m2WenQ.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LojkRcFZniyYFo3ag7qWGA.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XTTwQ-csmJzurE1Y2k4BgQ.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HNdAX3Wcw1jk2kme809G7g.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BxEpB1gxjEs3So3thK3f4Q.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*B-2RfJvaVs29IPHSnspp9Q.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SknFy4nG5VeUZXyL4jynJQ.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DDjEtNv92vzf68Vtol8zdw.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uc2b8U5_UMnk9Q8YKGilUg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fJJuorP1C_UZbxBxzG5gOQ.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*CXOtKkdMIk8oxIKtVNV0fQ.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WnW3TU0IaKMmkGHLaK-GNQ.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ka_Ypp8d9llhmB81RKM3Ng.jpeg" /><figcaption>Draft of Ajay, Joshua and Avi’s paper is avaialable at <a href="http://bit.ly/economics-of-AI">http://bit.ly/economics-of-AI</a></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Oo1SC4zZUBF5DxhDSmEybg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ILpPATppJ0RO5p-cHCFpyg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6zSTG0v2-1icVsvmMqpvDA.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MfLOvJO-dZ2GNYap8phn0w.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PeG_q98D42E1LBwTa77A5Q.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GFKVKmA97rQ4jnVdFj6ZNQ.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iWVnqdEzG7hAEiAEyPtNZA.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Oo-NcgxeKWUQGwqjUh4kCg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QkVceF5LoXHmhjEw9Yntkg.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=22a55d2a0982" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Could AI be the answer to today’s biggest business theory dilemma?]]></title>
            <link>https://medium.com/@jamescham/could-ai-be-the-answer-to-todays-biggest-business-theory-dilemma-90293fbbdc75?source=rss-86a8b1c606a8------2</link>
            <guid isPermaLink="false">https://medium.com/p/90293fbbdc75</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[James Cham]]></dc:creator>
            <pubDate>Tue, 01 Nov 2016 23:50:17 GMT</pubDate>
            <atom:updated>2016-11-01T23:50:17.629Z</atom:updated>
            <content:encoded><![CDATA[<p>Business philosophers have hit a dead end. Most know that maximizing profit for shareholders ultimately results in anti-social behavior but have struggled to provide a better alternative. That’s because much of current business theory rises from 1950’s-era operations research ideas of maximizing the value of equations and it is too difficult to come up with an objective description of how a corporation should act.</p><p>Since the rise of Kaplan, we’ve been keenly aware that the logical end of maximizing shareholder behavior — ethics-pushing and unsavory — has only been limited by the fact that employees are often unwilling to do the wrong thing because they’re bound by human intuition and morality. But business professors have argued that’s too imprecise and difficult to describe.</p><p>Enter the once arcane field of AI ethics. Instead of objective description, imagine something that can be programmed. And instead of limiting the actions of soulless corporation, imagine preventing killer robots. Stuart Russell (who wrote the seminal textbook on AI) has been thinking about how to limit and manage the ethics of robots. His student’s framework, inverse reinforcement learning, boils down to balancing many different objectives by learning from actual behavior — think do as I do rather than as I say.</p><p>This is still in the fringe of AI research as programs haven’t been flexible enough to think this way. But it has been growing quickly as AIs become more powerful. Don’t be surprised if it doesn’t have profound implications for traditional corporations soon.</p><p><a href="https://en.wikipedia.org/wiki/Apprenticeship_learning">https://en.wikipedia.org/wiki/Apprenticeship_learning</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=90293fbbdc75" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>