Clever Geek Handbook

The path of automation: competition with AI, getting rid of the routine, unconditional base time



( c )







How to automate work and never work? Before answering this question, ask yourself about something else: how much time do you need for happiness? People who do not know how to take a weekend evening, do not need to automate many tasks.







Reducing the routine gives you a competitive advantage. Automation "extracts" the most valuable resource on the planet - time. Technical progress allows you to spend time on what you are interested in: on solving intellectual problems and, probably, more complex work, the implementation of which will provide more opportunities to improve the quality of life.







Further - about what makes life easier for us.







Front Fighting Routine









A common saying asserts that a hardworking person will be engaged in monotonous, exhausting work day by day all his life, and only a lazy person will think about how to do the same work much faster and put less effort into it. A year ago, Reddit shook the story of a Californian start- up tester: having automated his daily tasks, he had fun during work hours and did not actually work.







According to the anonymous author, he spent the first eight months at work to automate the conduct of software tests. The next six years, he actually did not work, but regularly received a salary.







There are many inconsistencies in this story - it is difficult to really believe that someone is able to automate their actions so much. A similar case, which was told on bash.im , at least can be reproduced in reality - there were craftsmen who posted their version of the code on GitHub to automate routine actions (from communicating with his wife and bosses to making coffee).







There are many examples of automating routine tasks that are easy to believe in - indeed, you can even repeat them yourself. The user of the site Stack Overflow complained of remorse because of the automation of the workflow associated with data entry. The developer, who suffered from excessive qualifications , wrote several scripts that began to do all the work for him. As a result, he began to give work only an hour a week, while the previous employee in this position worked fully.







Another developer was fired for outsourcing to China . An intelligent guy paid someone less than one-fifth of his six-figure salary. As a result, his daily routine was as follows:







9:00 - Arrival at work. Reads reddit a couple of hours. Watching videos with cats.

11:30 - Lunch.

13:00 - Shopping time on Ebay.

14:00 - It's time to post something on the social network.

16:30 - The end of the day. Sends mail.

5:00 pm - Going home.







However, the developer was not smart enough to cover his tracks. His company noticed IP addresses to connect to a VPN from China, while VPN was usually used so that developers could work from home. Further monitoring of activities brought to the computer "employee-dodger." Having recovered deleted files from his computer, he managed to get hundreds of invoices from a Chinese consulting firm.













Aaron Rogers was taken to the position of operator with the task of sitting in front of the screen panel and monitoring the performance of the system. If an error message occurred, Rodgers had to personally fix the problem or call the senior team member. When he did not look at the screens, he was responsible for archiving files, rebooting servers, creating backups and performing a truly terrible task, which consisted in comparing two extremely long data reports. For every 1000 lines, he found about five discrepancies, while the reports themselves could consist of tens of thousands of lines.







With only one course on coding in college, he decided to automate his routine tasks. The company had a set of files that had to be archived at the end of each working day, so instead of manual work, he used a simple PowerShell language. And that was just the beginning.







Two years later, Roger's scripts completely eliminated the need for operators of flesh and blood: the company was forced to upgrade it, providing an opportunity to engage in analytical work.







But the hero of this story was not even a programmer!







Such cases used to seem exceptional, but today there are few offices in which work cannot be automated. At the same time, there is a huge mass of workers who do not know how to overcome boredom. According to a Gallup poll (in the US, but also relevant for Russia), 70% of employees lose their motivation and simply “turn off” at work. Of this group, 18% have reached the point of not only not caring about their own work, but even reducing the productivity of others. Perhaps it is these people who fear competition with AI most of all, but it may also be that they would be happy to get rid of their work routine.







Unconditional base time





( c )







The question is really relevant, what should the management do with the employee who has automated everything around? Reduce the salary, fire, entrusting all the work to the scripts, or, conversely, increase, transfer to another front of work, where the skills of the employee will help more?







Automation gives you work without mistakes, reduces costs, improves productivity - everything that is required by any business. The result of automation is not just free time in which you can watch a series or play. Routine procedures are simplified, and the programmer has more time to simply think, form a deep understanding of the solution of the tasks.







Time gained as a result of automation is important to invest correctly: it is a resource that helps in any area (not only IT) to become more efficient for you and the whole business. Actually, if you believe in the invasion of robots and is simply useful if you do not plan to compete with a strong AI.







By analogy with unconditional basic income, one can imagine an economic concept in which time is the main value. You can be the richest man on earth, but if you spend all your time on routine tasks, can you call you really happy?







Each person at a conscious age does not fully control the unconditional base time — the minutes between awakening and that sweet moment when you close your eyes again. Thousands of things in the world take time: meetings, meetings, letters, Hiktimes. Automation can give everyone and everyone a few hours of compulsory and free time.







Stanislav Sazhin, the founder of the Doctor at Work startup, recently published a note “ Success depends on how much time you DO NOT work ”, which protects the “doing nothing” position to increase efficiency in work. Stanislav thinks 10 hours every day, without being distracted by routine tasks. Just thinking. While walking, driving a car, reading the sources that allow you to think. Such a schedule makes it possible to make decisions more effectively, but it is difficult to apply it for an ordinary hired employee.







On the back of the distribution of working time - Ilon Musk, who lives 100-hour working weeks. Mask practically does not take breaks at work, preferring to have a snack during meetings, at the same time answering important letters, devotes very little time to his family and practically does not rest.







Stanislav Sazhin and Ilon Mask use different approaches, but they have one goal - to increase efficiency. We can use their advice or “get” time ( which you can now spend on kittens ) using various services and applications.







New tools









Anyone can optimize tasks using code, even a non-programmer. Here are just a few tools you can start with:









The mechanics of automating the work of programmers and testers depend on specific tasks, but even here progress does not stand still. Recently, researchers at the Massachusetts Institute of Technology have developed a system that allows you to automatically reuse the code.







The programmer selects a fragment of the source code in one program and an insertion point in the second program. CodeCarbonCopy will automatically make changes that integrate the code into the new context. The system superimposes the data presentation algorithm of the donor program to the algorithm of the recipient program in such a way as to import the code without problems.







Using CodeCarbonCopy, the same code can be used again and again. This is another step towards automating the software development cycle. Perhaps humanity has already written most of all the necessary software - it now remains to assemble code from the ocean of existing programs for any tasks.







For now, the experimental system works well with rigidly organized file formats and image editors that store representations of data in arrays, which are essentially rows of memory blocks of the same size. In the future, CodeCarbonCopy will have to work at all with any software.







Automation can seriously facilitate the work and highly skilled engineers. For example, the FBAR system makes it unnecessary for a person to participate if a hard drive crashes or a hardware error occurs on the server. FBAR automates software recovery and debugging. Such samopisnye systems are being introduced in many companies.







Operation of intelligent algorithms









While some are afraid that the AI ​​will soon take away their work (at least turn it into gray mucus ), others use any computer capabilities to make their lives easier.







There is no need to go far for examples: Prisma, Artisto, Vinci and other applications of the “neural networks boom era” appeared after scientists published in public access the results of studies on convolutional neural networks.







The next breakthrough may be expected in the field of text generation, but a decade ago, computer-generated articles were used by people for their own enrichment. Economist Phil Parker has carefully tuned the process of automated book writing . The problem arose in his preparation of his own academic publication - the book required thoroughness, time and money.







An unexpected decision - to entrust the work to computer algorithms - led him to the wonderful world of high-quality copy-paste. He "created" hundreds of thousands of books - from a guide for doctors on Klinefelter syndrome to countless sets of crosswords.







The work is simple: computers compile information gathered from open sources into books of various genres of 100–200 pages. The whole process for one copy takes 20 minutes. Then the e-book is put up for sale, it can also be ordered on the principle of " print on demand ."







Many of Parker's books have impressive circulations - tens and even hundreds of copies. Medical libraries bought almost everything he produced.







The quality of such literature leaves much to be desired, and anyone with access to the Internet will quickly find the answers to his question, but the revolution in the creation of texts has already happened.







Today, robots write sports scores, weather reports, insurance reports, new product presentations, economic news - and they do it in such a way that you don’t notice the catch.







The Wordsmith system, created by Automated Insights, writes over a billion (!) Annually. For example, she is responsible for some publications in Forbes . Automated Insights take data from press releases of companies and official reports, compare them with publications for the previous period and then give out a news item based on the information received. Of course, Wordsmith does not replace a journalist, but it saves him from the routine, boring work with the same information.







An interesting opinion was expressed by Ginny Romette, IBM CEO. According to her, the development of automation will lead to the fact that people no longer have to program. In IBM, everything is reduced to Watson, who will one day be able to look at the data, understand them and reason over them, but speculations about the almighty AI, which will take away all the work, belong to the field of hypothetical speculation. Yes, in the distant future this is indeed possible, but how many people are willing to give up programming or pick up their child from educational IT courses?







From a practical point of view, we are not interested in what a strong AI can do one day. It is more useful to know which automation tools will help right now - let them be non-intelligent, not based on neural networks and not using all the possibilities of bigdates. Why wait for the future if everyone can become their own AI? We have at least one quality for this - we are talking about human intelligence, who once figured out how to do work, and now thinks about how to reduce its quantity to zero.







More articles:


All Articles