<?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 Jason Oviedo on Medium]]></title>
        <description><![CDATA[Stories by Jason Oviedo on Medium]]></description>
        <link>https://medium.com/@jasonoviedo?source=rss-1056a86afa3b------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Gn8QIEDu_RPShFXSWrFRCw.jpeg</url>
            <title>Stories by Jason Oviedo on Medium</title>
            <link>https://medium.com/@jasonoviedo?source=rss-1056a86afa3b------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 15:36:11 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@jasonoviedo/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[Mistakes]]></title>
            <link>https://medium.com/@jasonoviedo/mistakes-48e641d31b31?source=rss-1056a86afa3b------2</link>
            <guid isPermaLink="false">https://medium.com/p/48e641d31b31</guid>
            <category><![CDATA[mistakes]]></category>
            <dc:creator><![CDATA[Jason Oviedo]]></dc:creator>
            <pubDate>Tue, 09 Jun 2020 05:53:46 GMT</pubDate>
            <atom:updated>2020-06-09T05:53:46.133Z</atom:updated>
            <content:encoded><![CDATA[<p>You are usually advised to take risks, make mistakes, but it feels weird, “how can I invest myself into making more mistakes?”. How do you explain that to your family, your granny? “I insulted my boss, lost my job and haven’t paid the rent, but don’t worry, I’m just trying to add some more mistakes to my life”. You can’t design an experiment to force somebody to make mistakes.</p><p>You feel a mistake the moment you realize something you believed, was wrong. There are two parts to a mistake. The feel is quite strong, rooted by the physical pain when you tripped your little toe against the table’s leg. Rooted on every time you fell off the bike, every time the teacher asked you something you didn’t know, that ugly sensation in the underbelly.</p><p>Then comes the rationalization. The understanding of the facts, what happened? “There was gravel on the ground, I slipped and fell, it hurts and now my bike is broken”. There could be even more implications after the fact, the mistake starts to be absorbed. What did you do that led to the situation, was it you?</p><p>For it to be a mistake, you had to believe it was the best way, the right action, the right step. You were sure enough, hence you decided, it was you who took the action. Some of them feel bad. Some worse. Silly, grave, fatal, it is mainly in those moments, when we learn the most.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=48e641d31b31" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[We have been doing Zoom for robots]]></title>
            <link>https://medium.com/@jasonoviedo/we-have-been-doing-zoom-for-robots-30efddc39922?source=rss-1056a86afa3b------2</link>
            <guid isPermaLink="false">https://medium.com/p/30efddc39922</guid>
            <category><![CDATA[mobile-network]]></category>
            <category><![CDATA[4g]]></category>
            <category><![CDATA[etl]]></category>
            <category><![CDATA[bandwidth]]></category>
            <dc:creator><![CDATA[Jason Oviedo]]></dc:creator>
            <pubDate>Fri, 05 Jun 2020 01:23:18 GMT</pubDate>
            <atom:updated>2020-06-06T00:46:28.148Z</atom:updated>
            <content:encoded><![CDATA[<h4>Or how mobile networks work</h4><p>At Kiwibot we have been following a slightly different approach to the self driving problem, different from what many other (some already extinct) companies tried at first. We did not intend our robots to be fully autonomous, instead we developed a strategy of our own, namely, control the robots remotely and instead focus all our efforts on making it easier for remote supervisors to control the bots. It goes like this:</p><ol><li>Fully remote supervised robots</li><li>Build a robust ADA (Advanced Driving Assistance) system: Make it ever easier for supervisors, the way we call it, <em>reduce the amount of cognitive load required to supervise the robots</em>.</li><li>Scale. As the number of robots, hours and months increase, we collect the necessary data to improve our ADA, making it even easier for supervisors.</li><li>Improve and repeat.</li></ol><p>This we call, <strong><em>Parallel autonomy, </em></strong>in a sense, more similar to what Tesla did than any other strategy. It has worked for us, now looking to a fourth year of operations ahead.</p><p>The backbone of the entire system is of course, a robust and improving teleoperation platform that needs to be optimized for:</p><ul><li>A robot should be supervised from anywhere in the world.</li><li>Robots can be deployed anywhere in the world.</li><li>Cost: For trucks or self driving cars data transmission and hardware costs do not matter. In our business, having literally thousands of hours of uninterrupted video streaming every week means we are heavily constrained on the total cost of operation. Remember, our goal is to make every delivery cost less than a dollar total, <em>i. e.</em> virtually free. More on this in a future article.</li></ul><p>In order to build such a system, you need to really understand how networks, specifically mobile networks, work. This is a primer.</p><h3>Mobile networks 101</h3><p>Cellular networks use the electromagnetic spectrum in the frequency between 700MHz and 2100MHz, which correspond to radio waves. The frequencies are similar to those used in regular WI-FI (2400MHz) and home cordless telephones (900MHz). Being part of the electromagnetic spectrum, these waves are in essence the same as visible light (with a frequency on the order of 500THz), only light is just much more energetic than radio. Yet, it is quite helpful to imagine radio waves behaving similar to light. The cell antenna emits at a high power, a really bright light bulb of radio waves that illuminate all around, specially because most materials are translucent to radio, carbon fiber composites being one noteworthy exception. Metals and thick concrete tend to be quite opaque to radio too. And depending on the structure, metals in buildings can block all radio waves, although buildings with such structures are not as usual.</p><p>Cellular networks are organized in <em>Cells</em>. Actual antenna locations are not usually disclosed by carriers, but can be approximated.</p><p>Not all antennas use exactly the same frequencies. Carriers auction for the permission to use chunks of the spectrum, and divide them even further geographically, so adjacent antennas signal don’t interfere.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*BPfPP_p5S1Jn40hZ" /><figcaption>Figure 1. Verizon antenna locations in Berkeley, according to <a href="http://www.cellmapper.net/">www.cellmapper.net</a>. The bands (i.e. frequencies) that each use are shown.</figcaption></figure><p>The spectrum is further divided into bands, some of which are treated preferentially. First responders have band 14 allocated only to their needs. All other customers share the rest of the spectrum. Different carriers, after auction, have allotted chunks that only they can use, and some that are shared. In short, Cell networks modulate signal on frequency and repeat the same frequencies in antennas that are far away from each other.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*yevJ_h2EmmgWqaBF" /><figcaption>Figure 2. Frequency used by antennas close to each other is slightly different. The arrangement of antennas is chosen to maximise coverage.</figcaption></figure><p>Whenever you have signal, it means your cellphone is actively transmitting information back and forth to the tower, using a channel, a further subdivision of bands. When in high volume areas, the antennas are positioned in the nodes of the honeycomb like cells, and directional antennas are used.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*4iVOfJaMaexAyGNU" /><figcaption>Figure 3. Some of the bands (shared and exclusive) used by correriers.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*iUSJR9AIOF5r5Na2" /><figcaption>Figure 4. Frequencies and bands for LTE</figcaption></figure><h4>Signal</h4><p>You can think of signal power as the amount of “light” received from the antenna (RSRP).</p><blockquote><strong>RSRP</strong></blockquote><blockquote>&gt; -90 dBm =&gt; Excellent</blockquote><blockquote>-90 dBm to -105 dBm =&gt; Good</blockquote><blockquote>-106 dBm to -120 dBm =&gt; Fair</blockquote><blockquote>&lt; -120 dBm =&gt;Poor</blockquote><p>Depending on environmental conditions and other factors, the signal might be usable, or not, so it is typical to include a quality indicator (RSRQ) when reporting signal.</p><blockquote><strong>RSRQ</strong></blockquote><blockquote>&gt; -9 dB =&gt;Excellent</blockquote><blockquote>-9 dB to -12 dB =&gt;Good</blockquote><blockquote>&lt; -13 dB =&gt;Fair to poor</blockquote><p>There might be a lot of radio pollution on the location where you are (other sources emitting in the same frequencies), so a ratio of “how much signal” you can recover out of the background is calculated and reported (Signal to Noise Ratio).</p><blockquote><strong>SNR</strong></blockquote><blockquote>&gt; 10 =&gt; Excellent</blockquote><blockquote>6 to 10 =&gt; Good</blockquote><blockquote>0 to 5 =&gt; Fair</blockquote><blockquote>&lt; 0 =&gt; Poor</blockquote><p>Using these three values, the cell <em>receiver</em> (cellphone, mifi, router) calculates an indicator (RSSI), the actual amount of “bars” of signal you have.</p><blockquote><strong>RSSI</strong></blockquote><blockquote>&gt; -70 dBm =&gt; Excellent</blockquote><blockquote>-70 dBm to -85 dBm =&gt; Good</blockquote><blockquote>-86 dBm to -100 dBm =&gt;Fair</blockquote><blockquote>&lt; -100 dBm =&gt; Poor</blockquote><blockquote><em>Power</em>, <em>Quality</em> and <em>Indicator</em> are reported in dB, negative values, the closer to zero the better.</blockquote><p>It is important to note, even with excellent signal, the network conditions could be bad. Network could be saturated, or the traffic might be deprioritized, something that can happen specially often with carriers that don’t own their networks, but lease network capacity from ‘pure’ carriers (in Colombia you can find Avantel and Exito leasing network capacity, vs Movistar, Claro, and Tigo, who operate their own Networks). Or your data package might have just expired. We have encountered situations where the network just ‘gets stuck’, it is clear now that the problem goes well beyond signal.</p><h4>Transmitting back</h4><p>What we have covered so far, applies mostly for the cell tower transmitting to the receiver, but information needs to be sent back, for Kiwibot, this part is actually crucial, the robot is transmitting 10x what it receives. The tower usually transmits at 100W for omnidirectional antennas, and up to 500W for directional ones. A cell phone or router, by regulations can transmit at maximum 1W, the distance being the same in both directions, antennas at the tower can receive the information transmitted from the cell phone because they have a much higher sensitivity and are free from most background noise by design. The device and tower continuously monitor the communication quality, and actual transmission power used by the phone is selected dynamically, always under the threshold of 1W.</p><blockquote>Power ≈ 1/distance ^2</blockquote><p>Radio power decreases proportional to the inverse square of distance. Double the distance, power is reduced by four. But the inverse relationship is more interesting for us. The maximum effective distance depends on the transmission power. If initial power is too low, the phone would need to be really close to the tower for the tower to be able to pick the phone signal up. Under ideal circumstances, the phone will be transmitting at enough power (more battery usage clearly) for the signal to reach the antenna. Again, the system automatically monitors the quality of communications, the phone will try to increase power until it reaches the cap of 1W. Beyond that point, even if conditions couldn’t be better, there is a maximum distance at which communication is effective. We don’t control transmission power and we are capped by regulations. All of this does not consider the effects of the environment. There could be plenty of people using their cell phones around, creating background noise and further reducing range. Buildings, different materials, a bus parked on the street, all of it will have an effect, even if inside theoretical range, in practice, signal received by the tower can be heavily reduced after a certain distance.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/745/0*UwBXwxq_BhzYotWn" /><figcaption>Figure 5. Distance vs transmission power for two antenna gains. Lower is worse.</figcaption></figure><p>In our current version, we use carbon composites for the robot’s cabin (cost, resistance, and ease of manufacturing at this scale). The worst case scenario: a cell phone inside the cabin will effectively have its transmission power instantly reduced, because of the carbon opacity to radio. We don’t have yet a full characterization of the attenuation by the cabin’s composite, but we have observed it experimentally, there is a reduction. This might not be much of a problem if the robot happens to operate close to the antenna tower. But as soon as the robot starts moving away from the antenna, there will be a point where signal, and available bandwidth, will decrease. Certainly, the bigger the operations area, the higher the chances to go far from the antenna. That is why antennas need to be positioned outside the robot.</p><h4>Bandwidth</h4><p>When your phone successfully connects to the tower, you paid a data plan and still have some GBs to go, signal is good and you are not in the middle of a concert; when all of this happens (and the stars are aligned), you just established a connection. Think of a hose between your device and the other computer you are trying to connect to (a server, a supervisor’s console, your mum’s phone). How much water is this hose able to transmit per unit time?</p><p>That is your bandwidth, upstream, from your device to the other endpoint, and downstream the other way round. That is the only thing that matters. Transmitting data uses bandwidth. Literally, think of bandwidth as water flow. Your video call needs 250 liters per second (the robot aims to use 250 Kbps of bandwidth) upstream. If your hose cannot provide that, the connection will seem laggy, frizzy. The robot needs another 50 to 100 Kbps both upstream and downstream to communicate to the server and get control commands. We have done our best but we can still not guarantee that sometimes the bot will need more (for comparison, remote driving solutions for cars use around 1.5GB/hour, 3495Kbps, according to figures we have collected from several companies). The biggest, and only imperative, is to make more bandwidth available than is required. The moment this does not happen, packets will tend to get lost, requiring retransmission and worsening the problem, more bandwidth is required for retransmitting them in the same amount of time. Measuring bandwidth is not an easy problem, for one, it uses more bandwidth. The higher used bandwidth, the more data you used, the more you are paying at the end of the month.</p><p>Our current goal is to guarantee that the robot operates well with under 1Mbps available bandwidth. It does not mean the robot will use all that bandwidth permanently, but it helps set a worst case scenario that we all need to aim for. At his rate:</p><blockquote>1Mbps = 1000000 bits / sec = 1000 Kbps = 0.119 M bytes / sec</blockquote><p>Note above difference between bits and Bytes, 1 Byte equals 8 bits. Internet speed is usually measured in bits per second, while total data consumption is expressed in Bytes. Our robot uses way less bandwidth when it is not moving, assuming it moves 70% of the time and each GB costs $4:</p><blockquote>1Mbps = 0.119MB *3600sec 0.3GB/H =$1.17/H</blockquote><p>Just data transmission, adds $1.17 to the robot operation cost, each hour. As our technology improves, we expect the robot to be able to operate at even lower bandwidth.</p><h4>IP</h4><p>The IP address is the main identificator to locate a device on the network. Think of it as your home address, only it changes from time to time. It changes in your home connection, it certainly changes on your cell phone. There will be IP changes, no matter what. A new IP is usually assigned by the network when the receiver switches from one antenna to the other. An address change means that connections will need to be reestablished (read: video lost for some seconds if the software knows how to handle it, black until restart otherwise). But rest assured, IP changes will occur. There are mechanisms to keep a fixed IP address per device, but it is not cheap. Software should know how to handle changes. It is a problem rather unexisting on cell phones, iOS for example handles the change under the hood, you don’t even notice.</p><p>In the current implementation, the IP change problem is handled in part by the supervisor refreshing the web page. It forces the system to try and restart the video connection. In the meantime, the control software will try to automatically recover when the web socket connection fails right after an address change.</p><h4>Latency</h4><p>We have long used latency as metric to quickly identify whenever we have problems. The process described up to this point happens mostly automatically, you don’t realize your cell phone is adjusting the transmission power, nor that WhatsApp audio call connection just restarted. Systems are designed to consider this.</p><p>Instead of reporting a handful of different metrics, many of which are not clear, even now, we decided for a catch-them-all metric. If all goes right, when a packet is sent from the robot to the server and a response is received back, the roundtrip latency gives you a quick sense of the quality of the network. But remember, it is a shortcut. Low latency means we are good, everything is alright. High latency might mean just anything, from a server implementation error, to you forgetting to switch used sim cards for fresh ones, to a zombie (covid) apocalypse where everything is failing down, all of which have happened in the past. But measuring latency is easy, and fast, it takes no extra bandwidth. It is a useful metric in several situations, and our implementation uses it to synchronize the supervisors console, the server, and the robot, so it is a helpful metric. Just not a helpful one when you want to debug.</p><h4>The real problem</h4><p>So, after it all, if the robot does not have enough bandwidth, it is not possible to transmit all of the information that is required, the system starts to clog and a queue of data starts to form. This happens on any transport system, not only data, the metro and public bus systems too. You exceed the available capacity, so two things will start to happen:</p><ol><li>Latency starts to increase</li><li>Not all data can be transmitted: video degrades (it looks worse and freezes and times), control commands are lost.</li></ol><p>It might be a failure at any point in the system just described, the problem is rather simple. Not enough bandwidth, the robot fails. We have tried to fight back most of the problems that could arise. That is why we have tried antennas, routers and different operators, we know that all of them, if improved, do help, but at the end of the day the problem persists. The only real solution remaining, is to have a secondary backup channel. Think your building has just one elevator, you want to add a second one. This might not solve all of the problems, as the second carrier might just have the same situations the first one has.</p><h3>Conclusion</h3><p>After this short review, we hope the reader will have a better understanding of all the components that need to be taken into account to build a platform for robot supervision. Video streaming is certainly hard, thousands of hours of uninterrupted video over mobile networks is harder, but doable. If you have reached this point and do want to continue the conversation, shoot ahead your questions in the comments. Thanks!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=30efddc39922" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Dark energy]]></title>
            <link>https://medium.com/@jasonoviedo/dark-energy-206a4c2bd43d?source=rss-1056a86afa3b------2</link>
            <guid isPermaLink="false">https://medium.com/p/206a4c2bd43d</guid>
            <category><![CDATA[short-story]]></category>
            <dc:creator><![CDATA[Jason Oviedo]]></dc:creator>
            <pubDate>Fri, 17 Apr 2020 07:02:17 GMT</pubDate>
            <atom:updated>2020-04-17T17:28:33.944Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ssfPgUT0_br3dSVXGqiyZw.jpeg" /><figcaption>This image belongs to someone out there</figcaption></figure><p>It was dark. As dark as it gets. Picture what will probably be the future of our universe. As the universe gets older, hydrogen fuses away, stars swell, and like old ladies start to die, some fading away into darkness, some, the youngest ones, wild and utterly bitchy exploding, bursting out their guts like the fireworks of heaven. It will take time. The longest that you could picture. Die and rebirth. Let go of the human scales, of the average sizes and average times that we experience on our average lives. Let go of the months and the kilometers. Of the years and the liters. Think of the sizes and times of the universe, live at that rate. See the universe evolving at the rate of the lightyears, the galaxies alive, their stars beating at the pulse of novas. Thinking at these scales, life is an abundant phenomenon on the universe. And it’s getting older by the minute.</p><p>But darker. When all the bitches are gone, no red giants, our sun a lot of time ago, when all those times had passed, our universe will probably gravitate towards an endless darkness, a uniformity so pervasive, so every way. Everything just the same. Void, empty and cold. Remnants of what it used to be old worlds too, just like the void, cold.</p><p>It was dark as that. And in the darkness, you have no wrist watch to watch the time go by, so the time stops. The concept of time itself loses all meaning, the time it takes to go to the supermarket is the same as a thousand eternities. So a thousand eternities had passed, in the dark, so much time that it was enough for random quantum fluctuations that occur even in the dark, bigger and smaller, some just a fraction of a breadth, after a thousand eternities one was big enough to explode, huge all at once into a gigantic brand new universe. The first fraction of a second, the first plank time after it occurred, a sphere of plank length radius contained all of the energy of this new universe, all at once. By some yet unknown rules, a fraction of a second later this baby universe jumply evolved into a healthy toddler growing at the speed of light. All energy, whatever the heck that means. I have never been able to picture it myself. Since its birth, for some time, our universe remained just a ball of energy. If you stop the simulation right around that moment, enlarge image, rotate, enlarge again, the universe is just a ball of energy. Like a Kamehameha before the screaming part.</p><p>Growing and growing, the universe kept swelling until it reached a size where its density was right at the tip of not being enough, so some of the energy started to blip into matter, one here, other there, blip, blip, blip. One instant a few, the next, a million, one later enough for these newcomers to start hooking up feverishly, bonding in threesomes and plus-somes. The new universe kept growing.</p><p>Born from nothing, it blew away all possible natural laws there could be left in the cold. Rebel, created a new set of ones it liked. Selected and fine-tuned to its delight, playing with them while it was a toddler, like learning to use your hands. When some of its energy started to collapse to matter, it had already gravitated to a set of rules it knew would come in handy. By that time, the universe had already realized the rate of its expansion, and had thought it a natural thing of the universes, to just be created and start growing. The first matter collapse felt like a little tingle in the underbelly. Just tickles. Then stronger, but it thought the sensation to be normal things that happen. Growing some more.</p><p>Of course there was nobody else around, for around there was just black, so empty and cold that it really didn’t matter. As it got bigger, it was clear that there were a lot of changes going on inside. Its guts were ablaze with activity forming stars and shooting quasars. It started to realize that its rate of growth had not stopped, as the stretching all around was evident, normal things that happen to universes after all. Alone and stretching fast into some teen years, it used most of its time to reflect on the way itself worked, an indecipherable introspection only matched by the absolute irrelevance of the black outside, as all the universe had ever experienced was itself and nothing ever had come from the black, so the moment it heard it, like a rumor just, then outshone by a nova. But the universe had heard. It was something, something unknown coming from outside. There was an outside, the universe started to ponder, bewildered.</p><p>Some more stretching. A young healthy universe, a lot of its energy still intact, the one that became matter, hot, jumpy, collided and fused and formed planets where a lot of chemistry was emerging. Universes first are pure energy, it noted, and pure energy don’t like to hang around, it repels itself, like an antigravity thing that happens to balls of energy. On the contrary, the energy that collapses to matter attracts itself, otherwise why did it collapse in the first place? It had already realized that its rate of growth never really diminished, so it was actually growing faster than at the beginning. The feeble sensation of outside had repeated a couple of times, so it was also clear there was something out there, maybe other universes to hang out with. Maybe they were actually trying to communicate somehow. Maybe when bigger.</p><p>A lot of normal things happen to universes, the universe knew. Grow since birth, get colder. “With an infinity of time ahead, there will be a time”, it thought, “when I will find out the source”. There were other universes out there, it was sure, other consciences like its own, floating in the vast ocean of nothingness around, marbles in a carpet, far away to perceive each other, but with a chance of meeting out, maybe, before dissolving itself into the vast darkness again, so vast, so huge, so eternal that even its existence, a whole universe, was no more than a random fluctuation, a speck of life hoping to spark with somebody else close by.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=206a4c2bd43d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Some thoughts on (artificial) intelligence]]></title>
            <link>https://medium.com/@jasonoviedo/some-thoughts-on-artificial-intelligence-3e8450f74383?source=rss-1056a86afa3b------2</link>
            <guid isPermaLink="false">https://medium.com/p/3e8450f74383</guid>
            <category><![CDATA[evolution]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[consciousness]]></category>
            <dc:creator><![CDATA[Jason Oviedo]]></dc:creator>
            <pubDate>Wed, 18 Mar 2020 18:26:11 GMT</pubDate>
            <atom:updated>2021-07-22T17:11:18.818Z</atom:updated>
            <content:encoded><![CDATA[<p>Truth is that I´ve been obsessed with intelligence and ways to replicate it for years. My first business partner (and dear friend) Alejandro Navas has been as obsessed as myself with the topic, leading us to become really close the moment we met. We started to develop a theory back then trying to make a little sense of the understanding we had of intelligence in general. We thought there had to be a single process, a single cell, a single thing that the brain was doing in order to produce intelligence. At the time we called it the ‘Mindtom’ as in mind-atom. It has been years since that theory, but my obsession continues. Nowadays, in the time of AI, the questions of what exactly is intelligence, consciousness, what makes us humans different in general, are probably more important than ever. These are just some of the ideas I´ve been playing with in the last years. I encourage you to study the concepts I present here in order to better understand the whole thing.</p><h3>Evolution</h3><p>The very first thing we have to take into account is that the set of characteristics we associate with intelligence are the product of evolution. No matter what they are, no matter how little or big our understanding of them is, they evolved, selected for by the same processes that gave teeth to sharks, claws to tigers, and sonar to bats. In some moments of our long evolutionary history, environmental pressures favored more intelligent animals. In fact, when reviewing intelligence in animals, you will find a lot of examples of species that show ‘smart traits’. Without having a common definition of what intelligence exactly is, some might feel tempted to discredit this statement, yet I dare you to google ‘animal intelligence’ and go through the plethora of articles and videos you´ll find. If after a couple of hours you are still not convinced, go and visit a friend with a Border Collie.</p><p>Ravens, dogs, dolphins, elephants, octopuses, all of them show so interesting characteristics that it is really hard to argue there is something smart to them. Chimpanzees, gorillas, bonobos, primates in general, are easily described as being smart. Intelligence as a thing was not invented by humans. Our ancestors had a great platform to build upon and just developed from there. I´m not saying that there are not intelligent traits that are purely human, I just want you to picture a spectrum, from the dumbest things in nature to the smartest people, and imagine how, bit by bit, traits are added to different species as they evolve, a process that just accelerates as primates and bigger apes evolve and goes faster with the arrival of humans. It is a continuous process, never-ending, never stopping. It is probably still going on.</p><h3>Consciousness</h3><p>And yet, we are the only fully conscious animals. Or at least, that is what we like to think of ourselves. Problem is, just as intelligence in general, nobody has a definition of what consciousness is, no real way to test it on another being, either a person, an animal, or a machine. I don´t claim to have an answer, but an exercise.</p><p>Picture yourself as a computer. You are just circuitry that can process data coming from whatever sensors you have attached. A powerful computer nonetheless. You are put into a small robot in the center of a room, some really interesting algorithms running in your ‘head’. The very first moment you are turned on, you are given the task of ‘identifying the environment’. This little robot can move around, slopy as it might be. It has a couple of cameras and ranging sensors. I´m not going to detail the algorithms required for the task: pattern matching, relationship building, etc, just try to picture it for a second. This small robot, you, will start by analyzing the image feed. Some color patterns on it move independently of others, you associate them to be an entity. Some entities move, some others don´t. After analyzing for long enough (hours? Years?), the objects will be clearly identified and a common pattern amongst them starts to emerge. They are separated from the rest, each has some ‘individuality’. The concept of ‘entity’ starts to take shape, as the characteristics of these objects are combined and abstracted.</p><p>Soon enough, our algorithms will detect that these entities interact. So the task we were assigned can be restated as ‘identify the objects and their interactions’. Our little robot goes on, analyzing the objects in the environment, the way they interact and the concept of ‘interaction’ itself starts to take shape. It might take a long time, but given enough, our little robot will end up with a model of its world.</p><p>We are then presented with other robots, doing the same process of modeling the world, so a new concept appears. Some of the objects in this environment can interact with it by analyzing and modeling the world. As we observe the others, we start to understand and model this process they are performing as a new category of interaction. Our world so far is composed of objects, interactions and these funny other special objects that are modeling the world.</p><p>Finally, we are presented with a mirror. A strange device that reflects the world. We can see some of the objects we already now in the mirror. We can see some of the other robots in the mirror, we can correlate them in our model and realize they are just a reflection. And then we model ourselves. The little robot identifies itself in the mirror, models itself as just another funny special object modeling the world. A small leap here is required for our robot to model the fact that it <strong>is</strong> part of this world we are describing, doing the same the other robots are doing.</p><p>It models itself, it models its process of modeling. It models itself modeling. It models itself modeling itself. It models itself modeling itself modeling itself. In its model of the world, there are now objects. Interactions. A special kind of object that can model the world. And itself, in the world, modeling it, modeling itself and modeling itself modeling itself.</p><h3>Predictions</h3><p>When Jeff Hawkins was done inventing Palm some 20 years ago, he moved to study intelligence and, in his book ‘On Intelligence’, he presented the idea of intelligence as a memory-based, future predicting phenomenon. He stated the sole purpose of intelligence is to be able to predict what is next. There are some animals that behave purely reactive. Fish and amphibians with Mauthner cells will trigger a fast escape response the moment stimuli are detected. Clearly a last-line-of-defense survival response, it is reactive. The animal needs for a predator to touch it before escaping. Hawkins’s argument goes, intelligence evolved as a means to be able to predict what is going to happen, before it happens.</p><p>Predicting the future, even near future, seconds or milliseconds in advance, will give the animal with this ability a great advantage. If you are a fish, your chances of survival are greater. If you are a predator, you will be better fed. If you can predict the world around you to great detail, long into the future, you become the dominant species on the planet.</p><p>Our smartest algorithms allow us to predict human behavior, weather patterns and people on the streets so self-driving vehicles know when to stop. Humans are really good at predicting the future. The best part is that the skills that allow for future prediction are really useful for a lot of other intelligent abilities. Think of planning. Future prediction, as it happens, allows for future shaping. This idea, I´ll probably expand on, on another article.</p><h3>AI</h3><p>I still have some thoughts regarding what the Mindtom might be, but I am less and less convinced it is just one singular thing that creates intelligence. Don´t take my word verbatim though, it is still too early to have a final say. What is clear is that we have embarked on the task, for the past 7+ decades, of decoding, understanding, modeling, and replicating, the traits that have been clearly identified as being part of intelligence. And we have been good at that, even if not yet 100% successful. From the very first automated processing machines of long ago, any edge on computing power, data processing, algorithms, etc, gives the owner a business advantage. Market pressures imply that any new development is deployed as fast as possible, and that trend is only accelerating.</p><p>Market pressures are selecting for smarter solutions in the same way nature selected for smarter animals. In a similar way the spectrum of intelligence was traversed by evolving species, it will be traversed by evolving systems, not necessarily in the same order.</p><p>I´m not sure when general artificial intelligence will be achieved, but I have little doubt it will. I´m certain though that it will not be one momentous discovery, but a stream of improvements being performed worldwide, industry-wide, continuous, constant. It has been like that for decades and it will continue. Rest assured, intelligence is coming on a scale that currently is only limited by budget (both money and energy). Superhuman intelligence is coming. Algorithms and systems already control most of our world, from what we eat (think data-based marketing), see, listen and do. And we are coevolving with it. We are already trapped in a cycle, entwined with our smart machines in a way quite reminiscent of Isaac Asimov´s Gaia. And we won´t leave the cycle aside, but just spiral in it, faster and faster. Whether the spiral goes down to doom or up to ascension is to be seen, we don´t have full control anymore. Market does. And market is king.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3e8450f74383" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Kiwibot Framework for Technology Hires]]></title>
            <link>https://medium.com/kiwicampus/kiwibot-framework-for-technology-hires-6b010d18a2df?source=rss-1056a86afa3b------2</link>
            <guid isPermaLink="false">https://medium.com/p/6b010d18a2df</guid>
            <dc:creator><![CDATA[Jason Oviedo]]></dc:creator>
            <pubDate>Thu, 04 Jul 2019 02:33:13 GMT</pubDate>
            <atom:updated>2019-07-04T03:09:28.888Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AkymhVlLDOvMYQjM1fHubg.png" /><figcaption>Kiwi Campus team — Medellín, Colombia. Nov 2018.</figcaption></figure><p>As obvious as it is, I’ll say it out loud. Hiring technical people is no easy task. Even for a company like ours, in which, for its very nature attracts a ton of great people, attracting, selecting and actually hiring the <strong><em>right</em></strong> people is a hard problem. Keeping true to our startup DNA, we have always been ready for a good dose of experimentation. Hiring, of course, is not outside the realm of trying and improving over your mistakes. What I present next is our current philosophy for hiring, together with some insights and learnings. This document is intended for whoever is currently in the process of hiring, either technical recruiters, CTO’s and technical leaders. All comments are more than welcome 😃.</p><h3>Attract talent</h3><p>Technical people are first and foremost motivated by learning. The two most important items that will attract the right candidates will be <em>what</em> you as a company are doing, and <em>who</em> the teammates are going to be.</p><h4>What we are doing</h4><p>The <em>what</em> question is a little tricky. A lot of writers and thought leaders will tell you that you have to first speak about the <strong><em>why</em></strong><em>.</em> I have found that when you describe really well what you are doing, it actually works better, mainly because the right <strong><em>what</em></strong> will convey a lot of right <strong><em>why’s</em></strong><em>.</em></p><p>When I speak about what we do, I like to first establish that we are not a tech company, but a delivery company. That is our business. Delivery is what we want to accomplish. The technology in our case is not a goal in itself, but a medium to get where we want. Having this clarity allows you to set the expectations right from the beginning, creating the right alignment even before candidates are hired. But of course, it does not end like that, because you have to show your ambition, your goals. Our case again: We are a delivery company, but an ambitious, innovative one. We want to make all traditional delivery companies obsolete, reducing cost by an order of magnitude. We want to make delivery completely free for everyone. Fast, reliable, scalable. We want to create a world where humans and machines interact alongside daily, out in the streets, the most unstructured environment possible. We want to enable new business models and defeat the likes of Amazon, Doordash and Postmates in their own game.</p><p>The bar is set really high from the start. Technology is not a goal but the only way we can achieve such high goals. Let your candidates imagine the kind of technology they will have to create in order to solve the problems posted but the <em>what.</em></p><h4>The team</h4><p>There is no better way of learning than having the most amazing team. The dynamics of teachings, learnings, and exploration that occur amongst a great team are incredible. Showcase the strengths of the team from the very beginning. If you feel proud of your team, make sure your candidates know it!</p><h3>Understanding Candidates</h3><p>People’s motivations at work are quite different. You should define your filters to try to understand what makes your candidates move. The main motivators I’ve found so far are:</p><ol><li>Learning: Quite a big deal for engineers and technical people. This one is a MUST for ANY candidate. If he/she is not interested in learning, we should rule them out right away. But do not get fooled, not all positions offer the same learning opportunities for everybody. If you, above all else, want to learn AI, getting into a Mobile Dev position will become frustrating even if you are on an AI heavy company. You should be open and honest about the learning possibilities, understanding exactly what is the candidates are really interested in.</li><li>Achievement: Often the main motivator for salespeople. The feeling of a closed deal or a successfully completed project should be really important for candidates applying to positions in charge of guiding the execution for the company. This positions should be fulfilled by goal-oriented people. It is less necessary for technical positions, but a good measure of this motivator complements a candidate, independently of the position they are applying for.</li><li>Money/perks/life quality: Quite important for people, we want our employees to be relaxed about the more mundane aspects of life. But if candidates are way too focused on this part, it might be a red flag right away. We will compensate appropriately people, but this never should be the main motivator.</li><li>Love for the art: Of course it varies amongst people, but some of them just want to do their craft for craft’s sake. They love what they do and don’t care about much else. Not the context, not the company nor anything beyond the art itself. If this is the main motivator for a person, they become purists. it is clear we want people who love what they do, but working with purists can become tricky in no time. If this the main motivator for a candidate, he or she should be assessed quite carefully as the problems they bring might outweigh the gains in no time.</li></ol><p><strong>The main take away is: Understand what candidates want deep in their hearts, see if it matches what you have to offer.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qrCgBhsx4rYYmJa0GmG66A.png" /><figcaption>David Cardozo — AI &amp; Robotics Lead at Kiwi Campus</figcaption></figure><h3>Candidate dimensions</h3><p>By identifying individual performance dimensions and testing them individually, we can have a better match between candidate and position, making sure we get what we need at the end.</p><h4>Logic and abstract reasoning</h4><p>In general the brighter the better. Right now we are applying tests that serve to identify a candidate’s ability level. Our programming test is designed to apply algorithms that require a good enough level of logic and abstract reasoning. The questions located at the end also help measure this.</p><h4>Learning Ability</h4><p>Not sure yet how to directly test for this in a quick interview. One way that we are trying to do so is by sending our questions an hour and a half in advance. This way, we are also testing for candidate’s resourcefulness.</p><h4>Resourcefulness</h4><p>We ask directly: “How do you handle situations when you don’t know what to do?”. We expect answers that describe an attitude of:</p><ul><li>Problems can be solved, you just have to look for the answer</li><li>The candidate first tries to solve it himself, really fast</li><li>Candidate does not stop easily</li><li>Asking for help is ok but if it is the only resource the candidate speaks about, it is a red flag.</li></ul><h4>Discipline</h4><p>Discipline is always good, as the old Japanese saying reads: “Sooner or later, discipline defeats intelligence”. Intelligence plus discipline are the perfect combination.</p><h4>Knowledge</h4><p>Specific tests will depend on the specific position</p><h4>Experience</h4><p>We have found that experience is important, but do not over-rely on that. So far, the best people we have worked with are a good balance of enough experience (not necessarily the most) and the other dimensions here mentioned.</p><h4>Soft skills</h4><p>Candidates that cannot speak to people are a rather poor selection.</p><h3>Proposed Hiring Framework</h3><p>This is yet an initial approach to a formalized framework, a guide for our hiring efforts. So far it is intended for technical recruits mainly, but might be further extended. Again, just an initial approach. This document might guide the selection process for several levels though (from intern to senior)</p><ol><li>What do you need? I mean, what you really, really need across several dimensions.</li><li>Define your minimum across dimensions</li><li>Not necessarily the best people for all circumstances? I mean, seniority level is important. It will depend on the intended activities to be executed by the role. Developers are quite expensive. Amazing people will have a great impact but some roles might just not need that much. That being said, you should always try to hire the best people for your budget.</li><li>Sourcing of candidates: LinkedIn plus referrals have been our best allies so far.</li><li>Cultural fit. Will they fit in our culture? The way we work is suitable for them? It is important to test for this and set the expectations right from the very beginning. I’ve found it indispensable to be honest and clear from the onset. Many candidates won’t actually back down if you say the bads in your company. No company is perfect, yours probably is not either. By the way, transparent, clear honesty is a value we have put a lot of effort towards. We have not yet got there, but I’ll talk more about culture in a different post.</li><li><strong>Important</strong>: What we have to offer is something they want? This one is tricky. For example, we do a lot of AI, and people at the company can learn a lot from their pairs. That said, not all positions will offer a chance to directly work on AI. What I said about motivators above applies here.</li><li>Test each dimension individually as much as possible.</li></ol><h4>Our process</h4><p>What we are doing right now (and subject to change at any moment) is:</p><ol><li>Reach out</li><li>Over the phone cultural fit interview. Some basic questions trying to identify candidates with the correct intelligence, discipline, and character. We want nice people working with us. It is indispensable for us to be resourceful.</li><li>Video call technical interview. This one I’m doing myself. I test in depth for intelligence, abstract reasoning, and general tech knowledge.</li><li>Supplementary interviews or programming test depending on previous performance.</li></ol><h3>Final thoughts</h3><p>Hiring is yet to be fully invented and we keep on improving our process. All comments will always be welcomed.</p><p>Between, did I mentioned we are hiring? I would post the links here, but where is the resourcefulness to that 😉?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6b010d18a2df" width="1" height="1" alt=""><hr><p><a href="https://medium.com/kiwicampus/kiwibot-framework-for-technology-hires-6b010d18a2df">Kiwibot Framework for Technology Hires</a> was originally published in <a href="https://medium.com/kiwicampus">Kiwibot / Food delivery robot</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[No, this is not accident, it's by design]]></title>
            <link>https://medium.com/@jasonoviedo/no-this-is-not-accident-its-by-design-29177a469434?source=rss-1056a86afa3b------2</link>
            <guid isPermaLink="false">https://medium.com/p/29177a469434</guid>
            <category><![CDATA[delivery]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[gameplan]]></category>
            <dc:creator><![CDATA[Jason Oviedo]]></dc:creator>
            <pubDate>Tue, 04 Dec 2018 05:37:06 GMT</pubDate>
            <atom:updated>2018-12-04T05:42:31.902Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8b7ngvORMVdhdJ46atrOmQ.jpeg" /></figure><p>Since we started KiwiBot, we often hear that we are making a mistake, that trying to do two things at a time is wrong. A Rappi or DoorDash like delivery app, autonomous delivery robots, both at the same time. The worry people feel when thinking that we are splitting focus is quite understandable. Yet they are in fact wrong. We are not building a DoorDash AND an autonomous robot delivery company at the same time. You are mistaking strategy with purpose. Yes, we have a food delivery app, which happens to be growing really fast. Yes, we use AI and are developing our own robots and algorithms. But it seems we failed to tell the story the right way.</p><p>Right now, e-commerce is growing double digits, while on-demand grows even faster. But it is a nascent market. DoorDash, UberEats and the likes will burn billions in the next years developing the market, just making customers used to order on-demand, online. They don’t care how much they pay for the delivery itself because its cost is almost marginal compared to their CAC.</p><p>But sooner or later they’ll be ready to focus on lowering costs down. The company that is able to perform hundreds of thousands of orders a day, each for less than $1, will be then in a position so privileged it will be worth billions. That company could possibly offer their services to whoever. They could own a niche market (huge still). They could render the big cash-burning companies irrelevant. In this respect then, we don’t plan on competing with DoorDash. We will either be acquired, sell robots-as-a-service or plainly make DoorDash irrelevant by offering what their customers want, delivery, 10X cheaper. The opportunity is so big in part because, having delivery so cheap will enable a myriad of new opportunities, new markets and new business models.</p><p>So the big question is: how to get to that point? Again, hundreds of thousands of orders a day, cheap. The app is the strategy we use to get the orders we need to develop the technology, to create our processes and polish our techniques. Much of what we do only makes sense while having enough orders. The goal itself is to get to enough orders. We won&#39;t wait for any third party to send our way the orders we need. No time for that. We actually tried and it was more of a road block.</p><p>Campuses is the strategy we have used to keep focused, ease adoption, test faster, learn faster. No need no get permits from the city (although we venture into the city daily). Berkeley is the perfect war zone to see the interaction between machines and people. Some people hate machines and robots. But most customers just love what we do. And it was precisely that love, the acceptance we saw, what motivated us to launch our app and get the orders ourselves. We will never regret that decision.</p><p>And what about automation? Risky business certainly, but not rocket science. Everybody is building self-driving, mapping algorithms and hardware. The competition is so fierce that we don&#39;t think there is any competitive advantage in having the technology first any more. If we fail, there are already 10 companies offering their navigation-as-a-service (RB2.io is looking great by the way). There are thousands of GitHub repos with open tools. ROS2 will include ready to use packages. Whatever we (or Starship, or Robbie, or whoever) build will be replicated by some high school kid 6 months later. Furthermore, you can always fall back to remote control. But we won&#39;t fail, our team is just amazing, talented and hungry. By having engineering in Colombia, we go as fast, cheaper, happier :) . By the way, do not ever underestimate the momentum and hunger for success of Colombians.</p><p>The hardware part is certainly harder, so we have focused more on that. But the end game, what nobody is even dreaming of doing, is doing this at scale. Today. That is something far from copy-pastable. It&#39;s hard, I can tell you that. Very. That is our secret sauce. I&#39;m quite open about it because it is so hard to replicate. No trade secrets here, just a lot of dedication, discipline, processes. That is what we are building. Just the biggest robot delivery network. The cheapest, and at that, we are years ahead of anybody else. We can already do 200 orders a day at 1/3 the usual cost. We have a brand new fleet (100 robots), our fourth iteration, that we manufactured in China so we can scale our operation 3x or more. Funny thing is, neither of us had the tiniest idea of robotics, self-driving or manufacturing a year ago. But this is just the beginning. One year down the road we will have 2 new robot iterations, 1/4 delivery cost, 10x the number of orders. No way any regular kid replicates that.</p><p>So, this is our game plan, and for it to work, we need the best, risk taking, big thinking, proactive partners. If after reading this you want to learn more, we are happy to start a conversation.</p><p>With love from <a href="https://www.kiwicampus.com/">Kiwi</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*4fGRTzSZAkQi0965wDFBjQ.jpeg" /><figcaption>We are still trying to figure out who did this :)</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=29177a469434" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[La paz en la era de la revolución espacial]]></title>
            <link>https://medium.com/@jasonoviedo/la-paz-en-la-era-de-la-revoluci%C3%B3n-espacial-a04d5586f55d?source=rss-1056a86afa3b------2</link>
            <guid isPermaLink="false">https://medium.com/p/a04d5586f55d</guid>
            <category><![CDATA[paz]]></category>
            <category><![CDATA[colombia]]></category>
            <category><![CDATA[farc]]></category>
            <category><![CDATA[stream]]></category>
            <category><![CDATA[spacex]]></category>
            <dc:creator><![CDATA[Jason Oviedo]]></dc:creator>
            <pubDate>Wed, 28 Sep 2016 07:07:20 GMT</pubDate>
            <atom:updated>2016-09-28T07:07:20.768Z</atom:updated>
            <content:encoded><![CDATA[<p>Podría ser tal vez una de las semanas más emocionantes que haya visto. Luego de 50 años estamos a pocos días de la oportunidad histórica más grande que hayamos tenido como país. Colombia se encuentra justo a punto de firmar un acuerdo de paz con las FARC, a pesar de ser el grupo militar más odiado y repudiado por todos. El acuerdo de paz dista de ser perfecto. Es lo que hay. Pero lo más importante, no es que represente el final de la guerra. Para mi, representa la oportunidad de un nuevo comienzo.</p><p>Por supuesto, habrá retos. Por supuesto, no será, en caso de ser aprobado en el plebiscito, más que el inicio de un camino largo y complicado. Pero es una oportunidad de cambio, de reiniciar el pc. Lo interesante desde mi perspectiva personal, es que desde hace unas semanas estoy en San Francisco, California, así que he tenido la oportunidad de presenciar estos eventos desde un ángulo distinto. Mientras trabajo incansablemente por llevar la <em>start up</em> a la que pertenezco (<a href="http://www.kiwicampus.com/">Kiwi Campus</a>) a nuevos mercados, ocurre todo el tema de la paz. Así que, sí, emocionante.</p><p>Y como si fuera poco, Elon Musk y SpaceX: gran anuncio. Para quienes no conocen, <a href="https://en.wikipedia.org/wiki/Elon_Musk">Elon Musk</a> es un emprendedor de gran trayectoria. Nacido en Sudáfrica, emigró sólo hacia Norte America muy joven y desde entonces no ha hecho más que lanzar una revolución tecnológica tras otra. Hoy su fortuna se estima en 11 mil de millones de dólares, todo gracias a haber revolucionado los pagos por internet, la industria automovilística, energía solar y ahora, con <a href="https://en.wikipedia.org/wiki/SpaceX">SpaceX</a>, la tecnología espacial.</p><p>Hoy, SpaceX <a href="https://www.youtube.com/watch?v=A1YxNYiyALg">acaba de anunciar</a> su <a href="https://www.youtube.com/watch?v=0qo78R_yYFA">plan maestro</a> para llevar humanos a Marte. Una colonia. Humanos. En Marte! Todo esto parece un sueño.</p><p>¿Y que tiene esto que ver? Pues que es la oportunidad perfecta para ver el caso Colombiano desde una perspectiva distinta. Si nos detenemos un momento a pensar, cada vez que miramos la historia reciente del país, lo principal que se puede apreciar es guerra, narcotráfico, corrupción. Mientras el resto del planeta ha ido avanzando en tecnología, economía y sociedad, nosotros nos hemos mantenido atascados luchando guerras de hace 50 años y resolviendo problemas del pasado.</p><p>No estoy seguro si la ironía es evidente, así que me permito repetirlo: Mientras nosotros nos matábamos en los 60s y 70s, el mundo estaba pensando en revoluciones sociales. Mientras nosotros solo teníamos en la mente darnos bala durante los 80s y 90s, el mundo estaba pensando en la revolución del pc. Mientras en Bogotá no sabíamos si montarnos en el “bus” del sistema integrado, por acá estaban pensando en carros eléctricos. Al mismo tiempo que el gobierno no tiene la más remota idea de que hacer con Uber, por acá están desarrollando carros que se manejan solos (sí señores taxistas, carros que se manejan solos. A ver como les va cuando les toque esa). Mientras nosotros seguimos aún pensando si somos capaces de dejar atrás el pasado, la guerra y los odios de antaño, el mundo está pensando en energías renovables e inteligencia artificial. Y ahora, viajar a Marte.</p><p>Durante las últimas décadas, no hemos sido más que espectadores de las revoluciones globales. Tristemente el único protagonismo que hemos tenido es al hablar de narcotráfico. Brasil se convirtió en una potencia económica. Israel se transformo en la nación de las start ups. Chile, aún a pesar de su dictadura se convirtió en una de las economías más fuertes del continente. Nosotros? Masacres y corrupción. Y hasta burros bomba.</p><p>Tuvimos también nuestro momento (CIVETS? alguien?). Pero no. Nos empeñamos como sociedad en mirar hacia atrás. El gobierno prefirió invertir billones en la industria petrolera, mientras que el ecosistema emprendedor se asfixia sin recursos. Mientras Holanda y Alemania migraban hacía energía eólica, nosotros talábamos más selva y sacábamos más carbón. Se firmaron tratados de libre comercio, pero a nadie se le ocurrió pensar en infraestructura, tecnología o esa cosa medio importante, ¿cómo le llaman? Ah si, competitividad.</p><p>Lo verdaderamente triste es que aún en este preciso instante, todavía tenemos el descaro como pueblo de preguntarnos si vale la pena o no dejar la guerra atrás. Absurdo. Hoy, a puertas de las mayores revoluciones tecnológicas y sociales que el mundo haya presenciado, seguimos poniéndole atención a políticos de otra era, encasillados en perseguir sólo el bien personal.</p><p>Sin embargo, ésta es una oportunidad de cambio. Pero como toda oportunidad, es nuestro deber tomarla, aprovecharla al máximo. No basta con firmar la paz. No basta con decir SÍ en el plebiscito. Es necesario dejar de lado los rencores y los prejuicios. Dejar atrás los complejos anacrónicos, dejar de sentirnos un pueblo incapaz e inferior. Es necesario cambiar nuestra forma de pensar y darnos cuenta que es posible ser parte del mundo, que es posible hacer negocios sin estafar a la DIAN ni consiguiendo terrenos en zonas francas, o tumbándose la plata de los contratos. Éste es el momento en que debemos decir NO a todo lo se convierta en lastre. La invitación, para todos mis compatriotas, pero en particular para las nuevas generaciones, es que nos demos cuenta que podemos hacer parte del mundo, de los negocios, de la cultura y el deporte internacional. Pero todo esto requiere un pequeño sacrificio: cambiar el chip, nuestra forma de ver el mundo. Y claramente no estoy proponiendo que empecemos sólo a imaginar las estrellas teniendo tantos problemas terrenales. Pero claramente la forma de pensar tradicional no sirve.</p><p>Así que reitero mi invitación. Dejemos el pasado atrás, y empecemos a mirar hacia el futuro. Démonos el beneficio de la duda, que tal vez somos capaces de crear un futuro distinto y recorrer un camino diferente al que nos trajo hasta acá. Dejemos de matarnos, empecemos a construir un país mejor. Y por que no, démonos la oportunidad de soñar con las estrellas y ser parte del mundo que se viene.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a04d5586f55d" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>