YC Paul Graham

September 2024 At a YC event last week Brian Chesky gave a talk that everyone who was there will remember. Most founders I talked to afterward said it was the best they'd ever heard. Ron Conway, for the first time in his life, forgot to take notes. I'm not going to try to reproduce it here. Instead I want to talk about a question it raised. The theme of Brian's talk was that the conventional wisdom about how to run larger companies is mistaken. As Airbnb grew, well-meaning people advised him that he had to run the company in a certain way for it to scale. Their advice could be optimistically summarized as "hire good people and give them room to do their jobs." He followed this advice and the results were disastrous. So he had to figure out a better way on his own, which he did partly by studying how Steve Jobs ran Apple. So far it seems to be working. Airbnb's free cash flow margin is now among the best in Silicon Valley. The audience at this event included a lot of the most successful founders we've funded, and one after another said that the same thing had happened to them. They'd been given the same advice about how to run their companies as they grew, but instead of helping their companies, it had damaged them. Why was everyone telling these founders the wrong thing? That was the big mystery to me. And after mulling it over for a bit I figured out the answer: what they were being told was how to run a company you hadn't founded — how to run a company if you're merely a professional manager. But this m.o. is so much less effective that to founders it feels broken. There are things founders can do that managers can't, and not doing them feels wrong to founders, because it is. In effect there are two different ways to run a company: founder mode and manager mode. Till now most people even in Silicon Valley have implicitly assumed that scaling a startup meant switching to manager mode. But we can infer the existence of another mode from the dismay of founders who've tried it, and the success of their attempts to escape from it. There are as far as I know no books specifically about founder mode. Business schools don't know it exists. All we have so far are the experiments of individual founders who've been figuring it out for themselves. But now that we know what we're looking for, we can search for it. I hope in a few years founder mode will be as well understood as manager mode. We can already guess at some of the ways it will differ. The way managers are taught to run companies seems to be like modular design in the sense that you treat subtrees of the org chart as black boxes. You tell your direct reports what to do, and it's up to them to figure out how. But you don't get involved in the details of what they do. That would be micromanaging them, which is bad. Hire good people and give them room to do their jobs. Sounds great when it's described that way, doesn't it? Except in practice, judging from the report of founder after founder, what this often turns out to mean is: hire professional fakers and let them drive the company into the ground. One theme I noticed both in Brian's talk and when talking to founders afterward was the idea of being gaslit. Founders feel like they're being gaslit from both sides — by the people telling them they have to run their companies like managers, and by the people working for them when they do. Usually when everyone around you disagrees with you, your default assumption should be that you're mistaken. But this is one of the rare exceptions. VCs who haven't been founders themselves don't know how founders should run companies, and C-level execs, as a class, include some of the most skillful liars in the world. [1] Whatever founder mode consists of, it's pretty clear that it's going to break the principle that the CEO should engage with the company only via his or her direct reports. "Skip-level" meetings will become the norm instead of a practice so unusual that there's a name for it. And once you abandon that constraint there are a huge number of permutations to choose from. For example, Steve Jobs used to run an annual retreat for what he considered the 100 most important people at Apple, and these were not the 100 people highest on the org chart. Can you imagine the force of will it would take to do this at the average company? And yet imagine how useful such a thing could be. It could make a big company feel like a startup. Steve presumably wouldn't have kept having these retreats if they didn't work. But I've never heard of another company doing this. So is it a good idea, or a bad one? We still don't know. That's how little we know about founder mode. [2] Obviously founders can't keep running a 2000 person company the way they ran it when it had 20. There's going to have to be some amount of delegation. Where the borders of autonomy end up, and how sharp they are, will probably vary from company to company. They'll even vary from time to time within the same company, as managers earn trust. So founder mode will be more complicated than manager mode. But it will also work better. We already know that from the examples of individual founders groping their way toward it. Indeed, another prediction I'll make about founder mode is that once we figure out what it is, we'll find that a number of individual founders were already most of the way there — except that in doing what they did they were regarded by many as eccentric or worse. [3] Curiously enough it's an encouraging thought that we still know so little about founder mode. Look at what founders have achieved already, and yet they've achieved this against a headwind of bad advice. Imagine what they'll do once we can tell them how to run their companies like Steve Jobs instead of John Sculley. September 2024 At a YC event last week Brian Chesky gave a talk that everyone who was there will remember. Most founders I talked to afterward said it was the best they'd ever heard. Ron Conway, for the first time in his life, forgot to take notes. I'm not going to try to reproduce it here. Instead I want to talk about a question it raised. The theme of Brian's talk was that the conventional wisdom about how to run larger companies is mistaken. As Airbnb grew, well-meaning people advised him that he had to run the company in a certain way for it to scale. Their advice could be optimistically summarized as "hire good people and give them room to do their jobs." He followed this advice and the results were disastrous. So he had to figure out a better way on his own, which he did partly by studying how Steve Jobs ran Apple. So far it seems to be working. Airbnb's free cash flow margin is now among the best in Silicon Valley. The audience at this event included a lot of the most successful founders we've funded, and one after another said that the same thing had happened to them. They'd been given the same advice about how to run their companies as they grew, but instead of helping their companies, it had damaged them. Why was everyone telling these founders the wrong thing? That was the big mystery to me. And after mulling it over for a bit I figured out the answer: what they were being told was how to run a company you hadn't founded — how to run a company if you're merely a professional manager. But this m.o. is so much less effective that to founders it feels broken. There are things founders can do that managers can't, and not doing them feels wrong to founders, because it is. In effect there are two different ways to run a company: founder mode and manager mode. Till now most people even in Silicon Valley have implicitly assumed that scaling a startup meant switching to manager mode. But we can infer the existence of another mode from the dismay of founders who've tried it, and the success of their attempts to escape from it. There are as far as I know no books specifically about founder mode. Business schools don't know it exists. All we have so far are the experiments of individual founders who've been figuring it out for themselves. But now that we know what we're looking for, we can search for it. I hope in a few years founder mode will be as well understood as manager mode. We can already guess at some of the ways it will differ. The way managers are taught to run companies seems to be like modular design in the sense that you treat subtrees of the org chart as black boxes. You tell your direct reports what to do, and it's up to them to figure out how. But you don't get involved in the details of what they do. That would be micromanaging them, which is bad. Hire good people and give them room to do their jobs. Sounds great when it's described that way, doesn't it? Except in practice, judging from the report of founder after founder, what this often turns out to mean is: hire professional fakers and let them drive the company into the ground. One theme I noticed both in Brian's talk and when talking to founders afterward was the idea of being gaslit. Founders feel like they're being gaslit from both sides — by the people telling them they have to run their companies like managers, and by the people working for them when they do. Usually when everyone around you disagrees with you, your default assumption should be that you're mistaken. But this is one of the rare exceptions. VCs who haven't been founders themselves don't know how founders should run companies, and C-level execs, as a class, include some of the most skillful liars in the world. [1] Whatever founder mode consists of, it's pretty clear that it's going to break the principle that the CEO should engage with the company only via his or her direct reports. "Skip-level" meetings will become the norm instead of a practice so unusual that there's a name for it. And once you abandon that constraint there are a huge number of permutations to choose from. For example, Steve Jobs used to run an annual retreat for what he considered the 100 most important people at Apple, and these were not the 100 people highest on the org chart. Can you imagine the force of will it would take to do this at the average company? And yet imagine how useful such a thing could be. It could make a big company feel like a startup. Steve presumably wouldn't have kept having these retreats if they didn't work. But I've never heard of another company doing this. So is it a good idea, or a bad one? We still don't know. That's how little we know about founder mode. [2] Obviously founders can't keep running a 2000 person company the way they ran it when it had 20. There's going to have to be some amount of delegation. Where the borders of autonomy end up, and how sharp they are, will probably vary from company to company. They'll even vary from time to time within the same company, as managers earn trust. So founder mode will be more complicated than manager mode. But it will also work better. We already know that from the examples of individual founders groping their way toward it. Indeed, another prediction I'll make about founder mode is that once we figure out what it is, we'll find that a number of individual founders were already most of the way there — except that in doing what they did they were regarded by many as eccentric or worse. [3] Curiously enough it's an encouraging thought that we still know so little about founder mode. Look at what founders have achieved already, and yet they've achieved this against a headwind of bad advice. Imagine what they'll do once we can tell them how to run their companies like Steve Jobs instead of John Sculley. October 2005 (This essay is derived from a talk at the 2005 Startup School.) How do you get good ideas for startups? That's probably the number one question people ask me. I'd like to reply with another question: why do people think it's hard to come up with ideas for startups? That might seem a stupid thing to ask. Why do they think it's hard? If people can't do it, then it is hard, at least for them. Right? Well, maybe not. What people usually say is not that they can't think of ideas, but that they don't have any. That's not quite the same thing. It could be the reason they don't have any is that they haven't tried to generate them. I think this is often the case. I think people believe that coming up with ideas for startups is very hard-- that it must be very hard-- and so they don't try do to it. They assume ideas are like miracles: they either pop into your head or they don't. I also have a theory about why people think this. They overvalue ideas. They think creating a startup is just a matter of implementing some fabulous initial idea. And since a successful startup is worth millions of dollars, a good idea is therefore a million dollar idea. If coming up with an idea for a startup equals coming up with a million dollar idea, then of course it's going to seem hard. Too hard to bother trying. Our instincts tell us something so valuable would not be just lying around for anyone to discover. Actually, startup ideas are not million dollar ideas, and here's an experiment you can try to prove it: just try to sell one. Nothing evolves faster than markets. The fact that there's no market for startup ideas suggests there's no demand. Which means, in the narrow sense of the word, that startup ideas are worthless. Questions The fact is, most startups end up nothing like the initial idea. It would be closer to the truth to say the main value of your initial idea is that, in the process of discovering it's broken, you'll come up with your real idea. The initial idea is just a starting point-- not a blueprint, but a question. It might help if they were expressed that way. Instead of saying that your idea is to make a collaborative, web-based spreadsheet, say: could one make a collaborative, web-based spreadsheet? A few grammatical tweaks, and a woefully incomplete idea becomes a promising question to explore. There's a real difference, because an assertion provokes objections in a way a question doesn't. If you say: I'm going to build a web-based spreadsheet, then critics-- the most dangerous of which are in your own head-- will immediately reply that you'd be competing with Microsoft, that you couldn't give people the kind of UI they expect, that users wouldn't want to have their data on your servers, and so on. A question doesn't seem so challenging. It becomes: let's try making a web-based spreadsheet and see how far we get. And everyone knows that if you tried this you'd be able to make something useful. Maybe what you'd end up with wouldn't even be a spreadsheet. Maybe it would be some kind of new spreadsheet-like collaboration tool that doesn't even have a name yet. You wouldn't have thought of something like that except by implementing your way toward it. Treating a startup idea as a question changes what you're looking for. If an idea is a blueprint, it has to be right. But if it's a question, it can be wrong, so long as it's wrong in a way that leads to more ideas. One valuable way for an idea to be wrong is to be only a partial solution. When someone's working on a problem that seems too big, I always ask: is there some way to bite off some subset of the problem, then gradually expand from there? That will generally work unless you get trapped on a local maximum, like 1980s-style AI, or C. Upwind So far, we've reduced the problem from thinking of a million dollar idea to thinking of a mistaken question. That doesn't seem so hard, does it? To generate such questions you need two things: to be familiar with promising new technologies, and to have the right kind of friends. New technologies are the ingredients startup ideas are made of, and conversations with friends are the kitchen they're cooked in. Universities have both, and that's why so many startups grow out of them. They're filled with new technologies, because they're trying to produce research, and only things that are new count as research. And they're full of exactly the right kind of people to have ideas with: the other students, who will be not only smart but elastic-minded to a fault. The opposite extreme would be a well-paying but boring job at a big company. Big companies are biased against new technologies, and the people you'd meet there would be wrong too. In an essay I wrote for high school students, I said a good rule of thumb was to stay upwind-- to work on things that maximize your future options. The principle applies for adults too, though perhaps it has to be modified to: stay upwind for as long as you can, then cash in the potential energy you've accumulated when you need to pay for kids. I don't think people consciously realize this, but one reason downwind jobs like churning out Java for a bank pay so well is precisely that they are downwind. The market price for that kind of work is higher because it gives you fewer options for the future. A job that lets you work on exciting new stuff will tend to pay less, because part of the compensation is in the form of the new skills you'll learn. Grad school is the other end of the spectrum from a coding job at a big company: the pay's low but you spend most of your time working on new stuff. And of course, it's called "school," which makes that clear to everyone, though in fact all jobs are some percentage school. The right environment for having startup ideas need not be a university per se. It just has to be a situation with a large percentage of school. It's obvious why you want exposure to new technology, but why do you need other people? Can't you just think of new ideas yourself? The empirical answer is: no. Even Einstein needed people to bounce ideas off. Ideas get developed in the process of explaining them to the right kind of person. You need that resistance, just as a carver needs the resistance of the wood. This is one reason Y Combinator has a rule against investing in startups with only one founder. Practically every successful company has at least two. And because startup founders work under great pressure, it's critical they be friends. I didn't realize it till I was writing this, but that may help explain why there are so few female startup founders. I read on the Internet (so it must be true) that only 1.7% of VC-backed startups are founded by women. The percentage of female hackers is small, but not that small. So why the discrepancy? When you realize that successful startups tend to have multiple founders who were already friends, a possible explanation emerges. People's best friends are likely to be of the same sex, and if one group is a minority in some population, pairs of them will be a minority squared. [1] Doodling What these groups of co-founders do together is more complicated than just sitting down and trying to think of ideas. I suspect the most productive setup is a kind of together-alone-together sandwich. Together you talk about some hard problem, probably getting nowhere. Then, the next morning, one of you has an idea in the shower about how to solve it. He runs eagerly to to tell the others, and together they work out the kinks. What happens in that shower? It seems to me that ideas just pop into my head. But can we say more than that? Taking a shower is like a form of meditation. You're alert, but there's nothing to distract you. It's in a situation like this, where your mind is free to roam, that it bumps into new ideas. What happens when your mind wanders? It may be like doodling. Most people have characteristic ways of doodling. This habit is unconscious, but not random: I found my doodles changed after I started studying painting. I started to make the kind of gestures I'd make if I were drawing from life. They were atoms of drawing, but arranged randomly. [2] Perhaps letting your mind wander is like doodling with ideas. You have certain mental gestures you've learned in your work, and when you're not paying attention, you keep making these same gestures, but somewhat randomly. In effect, you call the same functions on random arguments. That's what a metaphor is: a function applied to an argument of the wrong type. Conveniently, as I was writing this, my mind wandered: would it be useful to have metaphors in a programming language? I don't know; I don't have time to think about this. But it's convenient because this is an example of what I mean by habits of mind. I spend a lot of time thinking about language design, and my habit of always asking "would x be useful in a programming language" just got invoked. If new ideas arise like doodles, this would explain why you have to work at something for a while before you have any. It's not just that you can't judge ideas till you're an expert in a field. You won't even generate ideas, because you won't have any habits of mind to invoke. Of course the habits of mind you invoke on some field don't have to be derived from working in that field. In fact, it's often better if they're not. You're not just looking for good ideas, but for good new ideas, and you have a better chance of generating those if you combine stuff from distant fields. As hackers, one of our habits of mind is to ask, could one open-source x? For example, what if you made an open-source operating system? A fine idea, but not very novel. Whereas if you ask, could you make an open-source play? you might be onto something. Are some kinds of work better sources of habits of mind than others? I suspect harder fields may be better sources, because to attack hard problems you need powerful solvents. I find math is a good source of metaphors-- good enough that it's worth studying just for that. Related fields are also good sources, especially when they're related in unexpected ways. Everyone knows computer science and electrical engineering are related, but precisely because everyone knows it, importing ideas from one to the other doesn't yield great profits. It's like importing something from Wisconsin to Michigan. Whereas (I claim) hacking and painting are also related, in the sense that hackers and painters are both makers, and this source of new ideas is practically virgin territory. Problems In theory you could stick together ideas at random and see what you came up with. What if you built a peer-to-peer dating site? Would it be useful to have an automatic book? Could you turn theorems into a commodity? When you assemble ideas at random like this, they may not be just stupid, but semantically ill-formed. What would it even mean to make theorems a commodity? You got me. I didn't think of that idea, just its name. You might come up with something useful this way, but I never have. It's like knowing a fabulous sculpture is hidden inside a block of marble, and all you have to do is remove the marble that isn't part of it. It's an encouraging thought, because it reminds you there is an answer, but it's not much use in practice because the search space is too big. I find that to have good ideas I need to be working on some problem. You can't start with randomness. You have to start with a problem, then let your mind wander just far enough for new ideas to form. In a way, it's harder to see problems than their solutions. Most people prefer to remain in denial about problems. It's obvious why: problems are irritating. They're problems! Imagine if people in 1700 saw their lives the way we'd see them. It would have been unbearable. This denial is such a powerful force that, even when presented with possible solutions, people often prefer to believe they wouldn't work. I saw this phenomenon when I worked on spam filters. In 2002, most people preferred to ignore spam, and most of those who didn't preferred to believe the heuristic filters then available were the best you could do. I found spam intolerable, and I felt it had to be possible to recognize it statistically. And it turns out that was all you needed to solve the problem. The algorithm I used was ridiculously simple. Anyone who'd really tried to solve the problem would have found it. It was just that no one had really tried to solve the problem. [3] Let me repeat that recipe: finding the problem intolerable and feeling it must be possible to solve it. Simple as it seems, that's the recipe for a lot of startup ideas. Wealth So far most of what I've said applies to ideas in general. What's special about startup ideas? Startup ideas are ideas for companies, and companies have to make money. And the way to make money is to make something people want. Wealth is what people want. I don't mean that as some kind of philosophical statement; I mean it as a tautology. So an idea for a startup is an idea for something people want. Wouldn't any good idea be something people want? Unfortunately not. I think new theorems are a fine thing to create, but there is no great demand for them. Whereas there appears to be great demand for celebrity gossip magazines. Wealth is defined democratically. Good ideas and valuable ideas are not quite the same thing; the difference is individual tastes. But valuable ideas are very close to good ideas, especially in technology. I think they're so close that you can get away with working as if the goal were to discover good ideas, so long as, in the final stage, you stop and ask: will people actually pay for this? Only a few ideas are likely to make it that far and then get shot down; RPN calculators might be one example. One way to make something people want is to look at stuff people use now that's broken. Dating sites are a prime example. They have millions of users, so they must be promising something people want. And yet they work horribly. Just ask anyone who uses them. It's as if they used the worse-is-better approach but stopped after the first stage and handed the thing over to marketers. Of course, the most obvious breakage in the average computer user's life is Windows itself. But this is a special case: you can't defeat a monopoly by a frontal attack. Windows can and will be overthrown, but not by giving people a better desktop OS. The way to kill it is to redefine the problem as a superset of the current one. The problem is not, what operating system should people use on desktop computers? but how should people use applications? There are answers to that question that don't even involve desktop computers. Everyone thinks Google is going to solve this problem, but it is a very subtle one, so subtle that a company as big as Google might well get it wrong. I think the odds are better than 50-50 that the Windows killer-- or more accurately, Windows transcender-- will come from some little startup. Another classic way to make something people want is to take a luxury and make it into a commmodity. People must want something if they pay a lot for it. And it is a very rare product that can't be made dramatically cheaper if you try. This was Henry Ford's plan. He made cars, which had been a luxury item, into a commodity. But the idea is much older than Henry Ford. Water mills transformed mechanical power from a luxury into a commodity, and they were used in the Roman empire. Arguably pastoralism transformed a luxury into a commodity. When you make something cheaper you can sell more of them. But if you make something dramatically cheaper you often get qualitative changes, because people start to use it in different ways. For example, once computers get so cheap that most people can have one of their own, you can use them as communication devices. Often to make something dramatically cheaper you have to redefine the problem. The Model T didn't have all the features previous cars did. It only came in black, for example. But it solved the problem people cared most about, which was getting from place to place. One of the most useful mental habits I know I learned from Michael Rabin: that the best way to solve a problem is often to redefine it. A lot of people use this technique without being consciously aware of it, but Rabin was spectacularly explicit. You need a big prime number? Those are pretty expensive. How about if I give you a big number that only has a 10 to the minus 100 chance of not being prime? Would that do? Well, probably; I mean, that's probably smaller than the chance that I'm imagining all this anyway. Redefining the problem is a particularly juicy heuristic when you have competitors, because it's so hard for rigid-minded people to follow. You can work in plain sight and they don't realize the danger. Don't worry about us. We're just working on search. Do one thing and do it well, that's our motto. Making things cheaper is actually a subset of a more general technique: making things easier. For a long time it was most of making things easier, but now that the things we build are so complicated, there's another rapidly growing subset: making things easier to use. This is an area where there's great room for improvement. What you want to be able to say about technology is: it just works. How often do you say that now? Simplicity takes effort-- genius, even. The average programmer seems to produce UI designs that are almost willfully bad. I was trying to use the stove at my mother's house a couple weeks ago. It was a new one, and instead of physical knobs it had buttons and an LED display. I tried pressing some buttons I thought would cause it to get hot, and you know what it said? "Err." Not even "Error." "Err." You can't just say "Err" to the user of a stove. You should design the UI so that errors are impossible. And the boneheads who designed this stove even had an example of such a UI to work from: the old one. You turn one knob to set the temperature and another to set the timer. What was wrong with that? It just worked. It seems that, for the average engineer, more options just means more rope to hang yourself. So if you want to start a startup, you can take almost any existing technology produced by a big company, and assume you could build something way easier to use. Design for Exit Success for a startup approximately equals getting bought. You need some kind of exit strategy, because you can't get the smartest people to work for you without giving them options likely to be worth something. Which means you either have to get bought or go public, and the number of startups that go public is very small. If success probably means getting bought, should you make that a conscious goal? The old answer was no: you were supposed to pretend that you wanted to create a giant, public company, and act surprised when someone made you an offer. Really, you want to buy us? Well, I suppose we'd consider it, for the right price. I think things are changing. If 98% of the time success means getting bought, why not be open about it? If 98% of the time you're doing product development on spec for some big company, why not think of that as your task? One advantage of this approach is that it gives you another source of ideas: look at big companies, think what they should be doing, and do it yourself. Even if they already know it, you'll probably be done faster. Just be sure to make something multiple acquirers will want. Don't fix Windows, because the only potential acquirer is Microsoft, and when there's only one acquirer, they don't have to hurry. They can take their time and copy you instead of buying you. If you want to get market price, work on something where there's competition. If an increasing number of startups are created to do product development on spec, it will be a natural counterweight to monopolies. Once some type of technology is captured by a monopoly, it will only evolve at big company rates instead of startup rates, whereas alternatives will evolve with especial speed. A free market interprets monopoly as damage and routes around it. The Woz Route The most productive way to generate startup ideas is also the most unlikely-sounding: by accident. If you look at how famous startups got started, a lot of them weren't initially supposed to be startups. Lotus began with a program Mitch Kapor wrote for a friend. Apple got started because Steve Wozniak wanted to build microcomputers, and his employer, Hewlett-Packard, wouldn't let him do it at work. Yahoo began as David Filo's personal collection of links. This is not the only way to start startups. You can sit down and consciously come up with an idea for a company; we did. But measured in total market cap, the build-stuff-for-yourself model might be more fruitful. It certainly has to be the most fun way to come up with startup ideas. And since a startup ought to have multiple founders who were already friends before they decided to start a company, the rather surprising conclusion is that the best way to generate startup ideas is to do what hackers do for fun: cook up amusing hacks with your friends. It seems like it violates some kind of conservation law, but there it is: the best way to get a "million dollar idea" is just to do what hackers enjoy doing anyway.

11 Key Linux Performance Tuning Tricks for DevOps Engineers improve application speed, resource usage, and system stability. Obafemi Obafemi Follow a11y-light · May 7, 2025 (Updated: May 7, 2025) · Free: No 1. Optimize CPU Scheduling for Better Performance By default, Linux dynamically adjusts CPU scheduling, but you can 'tweak process priorities' to improve its performance. Use nice and renice Use nice (which affects the process priority level) to start a process with a priority and renice to adjust it dynamically. Start a process with a high priority (-10 means higher priority). Copy nice -n -10 ./my_script.sh Change priority of an existing process (PID 1234) Copy renice -5 -p 1234 - 'Nice' values range from -20 (highest priority) to 19 (lowest priority). - Prioritizing critical workloads makes sure they get more CPU time. - Only root users can set negative nice values. - You could use ps -eo pid,comm,nice or top to find current priorities. 2. Increase File Descriptors to Handle More Open Files By default, Linux limits the number of open files per process and for high-performance web servers or databases, this can be a bottleneck. Solution: Adjust ulimit Check the current limit: Copy ulimit -n - ulimit -n only affects the current shell session. Increase the limit temporarily: Copy ulimit -n 100000 To make it persistent, edit /etc/security/limits.conf: Copy * soft nofile 100000 * hard nofile 200000 Applications like Nginx, MySQL, and Elasticsearch will now be able to handle more concurrent connections. 3. Enable TCP BBR for Faster Network Throughput Linux's default TCP congestion control (CUBIC) may not be optimal for high-speed networks. BBR (Bottleneck Bandwidth and Round-trip propagation time) significantly improves network performance and reduces latency. Solution: Enable TCP BBR Check the current congestion control algorithm: Copy sysctl net.ipv4.tcp_congestion_control Enable BBR: Copy echo "net.core.default_qdisc=fq" >> /etc/sysctl.conf echo "net.ipv4.tcp_congestion_control=bbr" >> /etc/sysctl.conf sysctl -p BBR significantly improves TCP performance, especially in cloud environments with high-latency or long-distance network connections.. 4. Reduce Swappiness to Prevent Unnecessary Disk Swaps By default, Linux starts using swap space aggressively, which can slow down performance. Solution: Reduce Swappiness Check current swappiness: Copy cat /proc/sys/vm/swappiness Reduce it (default is 60, lower values prefer RAM over swap): Copy echo "vm.swappiness=10" >> /etc/sysctl.conf sysctl -p With this, RAM is used efficiently before swapping to disk, which is much slower. You can test it temporarily before making it persistent: Copy echo 10 > /proc/sys/vm/swappiness 5. Optimize Disk I/O Performance with noop Scheduler For SSDs and cloud-based block storage, Linux's default I/O scheduler (cfq) isn't optimal because it was designed for spinning disk drives, not SSDs. Solution: Use noop or deadline Scheduler Check current scheduler: Copy cat /sys/block/sda/queue/scheduler Set to noop (ideal for SSDs): Copy echo "noop" > /sys/block/sda/queue/scheduler For persistent changes, add this to /etc/default/grub: Copy GRUB_CMDLINE_LINUX="elevator=noop" Then update Grub: Copy update-grub && reboot Overall, it minimizes I/O latency and maximizes SSD performance. 6. Tune Network Buffers for High Traffic Applications For high-throughput servers (e.g., Nginx, HAProxy), default Linux network buffers may be too low. Solution: Increase Buffer Sizes Copy echo "net.core.rmem_max=16777216" >> /etc/sysctl.conf echo "net.core.wmem_max=16777216" >> /etc/sysctl.conf echo "net.core.netdev_max_backlog=5000" >> /etc/sysctl.conf sysctl -p Prevents packet drops and improves performance for high-bandwidth applications. 7. Enable HugePages for Faster Memory Management HugePages reduce CPU overhead for memory-intensive applications like databases (PostgreSQL, MySQL). Solution: Enable HugePages Check current HugePages status: Copy cat /proc/meminfo | grep HugePages Enable HugePages (modify /etc/sysctl.conf): Copy echo "vm.nr_hugepages=1024" >> /etc/sysctl.conf sysctl -p This optimizes memory allocation and improves database performance. 8. Use tmpfs for Faster Temporary File Access If your application reads/writes temporary files frequently, using tmpfs (RAM-based filesystem) is much faster. Solution: Mount /tmp as tmpfs Edit /etc/fstab: Copy tmpfs /tmp tmpfs defaults,noatime,nosuid,size=2G 0 0 Mount it: Copy mount -a This speeds up caching, builds, and temporary storage. 9. Optimize Systemd Services for Faster Boot and Performance Some unnecessary background services consume CPU and RAM. Solution: Disable Unused Services List services consuming boot time: Copy systemd-analyze blame Disable unnecessary services: Copy systemctl disable bluetooth.service systemctl disable cups.service This frees up system resources, especially on minimal cloud VMs. 10. Enable CPU Frequency Scaling for Performance or Power Efficiency Linux dynamically adjusts CPU speed, but you can manually optimize it for performance. Solution: Use cpufrequtils Install CPU frequency utilities: Copy apt install cpufrequtils -y Set CPU to performance mode (disables power-saving throttling): Copy cpufreq-set -g performance For persistent changes, modify /etc/default/cpufrequtils: Copy GOVERNOR="performance" Maximum CPU speed is used for high-performance workloads. 11. Use tuned for Dynamic System Performance Profiles Linux systems can benefit from automated performance tuning using predefined profiles based on workload types (e.g., virtual hosts, databases, desktops). The tuned daemon dynamically adjusts system parameters in real-time to match the selected profile. Solution: Install and use tuned 1. Install tuned: On RHEL/CentOS: Copy yum install tuned -y On Ubuntu/Debian: Copy apt install tuned -y 2. Start and enable the service: Copy systemctl enable --now tuned 3. List available tuning profiles: Copy tuned-adm list 4. Apply a performance-optimized profile: Copy tuned-adm profile throughput-performance 5. Verify the active profile: Copy tuned-adm active Some popular profiles include: - throughput-performance: Optimizes for network/disk throughput. - latency-performance: Ideal for low-latency applications like databases. - virtual-guest: Optimized for virtual machines. - balanced: General-purpose tuning. 11 Linux Networking Commands important networking commands and their use cases. medium.com P.S. I recently released a DevOps Starter Kit , a bundle of 20+ Bash scripts, cron jobs, and Docker templates to speed up your infrastructure setup. If you do a lot of server work, you might find it useful. #linux-performance #linux #bash #devops #programming

Support Freedium

Dear Freedium users,

We've updated our donation options to provide you with more ways to support our mission. Your contributions are invaluable in helping us maintain and improve Freedium, ensuring we can continue to provide unrestricted access to quality content.

We now offer multiple platforms for donations, including Patreon, Ko-fi, and Liberapay. Each option allows you to support us in the way that's most convenient for you.

Your support, no matter the platform or amount, makes a significant difference. It allows us to cover our operational costs and invest in enhancing Freedium's features and reliability.

Thank you for being a part of the Freedium community and for your continued support.

Choose Your Preferred Donation Platform:

Freedium

• ko-fi
• librepay
• patreon

< Go to the original

11 Key Linux Performance Tuning Tricks for DevOps Engineers

improve application speed, resource usage, and system stability.

Obafemi

~4 min read · May 7, 2025 (Updated: May 7, 2025) · Free: No

1. Optimize CPU Scheduling for Better Performance

By default, Linux dynamically adjusts CPU scheduling, but you can 'tweak process priorities' to improve its performance.

Use nice and renice Use nice (which affects the process priority level) to start a process with a priority and renice to adjust it dynamically.

Start a process with a high priority (-10 means higher priority).

nice -n -10 ./my_script.sh

Change priority of an existing process (PID 1234)

renice -5 -p 1234

- 'Nice' values range from -20 (highest priority) to 19 (lowest priority). - Prioritizing critical workloads makes sure they get more CPU time. - Only root users can set negative nice values. - You could use ps -eo pid,comm,nice or top to find current priorities.

2. Increase File Descriptors to Handle More Open Files

By default, Linux limits the number of open files per process and for high-performance web servers or databases, this can be a bottleneck.

Solution: Adjust ulimit Check the current limit:

ulimit -n

- ulimit -n only affects the current shell session.

Increase the limit temporarily:

ulimit -n 100000

To make it persistent, edit /etc/security/limits.conf:

* soft nofile 100000
* hard nofile 200000

Applications like Nginx, MySQL, and Elasticsearch will now be able to handle more concurrent connections.

3. Enable TCP BBR for Faster Network Throughput

Linux's default TCP congestion control (CUBIC) may not be optimal for high-speed networks. BBR (Bottleneck Bandwidth and Round-trip propagation time) significantly improves network performance and reduces latency.

Solution: Enable TCP BBR Check the current congestion control algorithm:

sysctl net.ipv4.tcp_congestion_control

Enable BBR:

echo "net.core.default_qdisc=fq" >> /etc/sysctl.conf
echo "net.ipv4.tcp_congestion_control=bbr" >> /etc/sysctl.conf
sysctl -p

BBR significantly improves TCP performance, especially in cloud environments with high-latency or long-distance network connections..

4. Reduce Swappiness to Prevent Unnecessary Disk Swaps

By default, Linux starts using swap space aggressively, which can slow down performance.

Solution: Reduce Swappiness Check current swappiness:

cat /proc/sys/vm/swappiness

Reduce it (default is 60, lower values prefer RAM over swap):

echo "vm.swappiness=10" >> /etc/sysctl.conf
sysctl -p

With this, RAM is used efficiently before swapping to disk, which is much slower.

You can test it temporarily before making it persistent:

echo 10 > /proc/sys/vm/swappiness

5. Optimize Disk I/O Performance with noop Scheduler

For SSDs and cloud-based block storage, Linux's default I/O scheduler (cfq) isn't optimal because it was designed for spinning disk drives, not SSDs.

Solution: Use noop or deadline Scheduler Check current scheduler:

cat /sys/block/sda/queue/scheduler

Set to noop (ideal for SSDs):

echo "noop" > /sys/block/sda/queue/scheduler

For persistent changes, add this to /etc/default/grub:

GRUB_CMDLINE_LINUX="elevator=noop"

Then update Grub:

update-grub && reboot

Overall, it minimizes I/O latency and maximizes SSD performance.

6. Tune Network Buffers for High Traffic Applications

For high-throughput servers (e.g., Nginx, HAProxy), default Linux network buffers may be too low.

Solution: Increase Buffer Sizes

echo "net.core.rmem_max=16777216" >> /etc/sysctl.conf
echo "net.core.wmem_max=16777216" >> /etc/sysctl.conf
echo "net.core.netdev_max_backlog=5000" >> /etc/sysctl.conf
sysctl -p

Prevents packet drops and improves performance for high-bandwidth applications.

7. Enable HugePages for Faster Memory Management

HugePages reduce CPU overhead for memory-intensive applications like databases (PostgreSQL, MySQL).

Solution: Enable HugePages Check current HugePages status:

cat /proc/meminfo | grep HugePages

Enable HugePages (modify /etc/sysctl.conf):

echo "vm.nr_hugepages=1024" >> /etc/sysctl.conf
sysctl -p

This optimizes memory allocation and improves database performance.

8. Use tmpfs for Faster Temporary File Access

If your application reads/writes temporary files frequently, using tmpfs (RAM-based filesystem) is much faster.

Solution: Mount /tmp as tmpfs Edit /etc/fstab:

tmpfs /tmp tmpfs defaults,noatime,nosuid,size=2G 0 0

Mount it:

mount -a

This speeds up caching, builds, and temporary storage.

9. Optimize Systemd Services for Faster Boot and Performance

Some unnecessary background services consume CPU and RAM.

Solution: Disable Unused Services List services consuming boot time:

systemd-analyze blame

Disable unnecessary services:

systemctl disable bluetooth.service
systemctl disable cups.service

This frees up system resources, especially on minimal cloud VMs.

10. Enable CPU Frequency Scaling for Performance or Power Efficiency

Linux dynamically adjusts CPU speed, but you can manually optimize it for performance.

Solution: Use cpufrequtils Install CPU frequency utilities:

apt install cpufrequtils -y

Set CPU to performance mode (disables power-saving throttling):

cpufreq-set -g performance

For persistent changes, modify /etc/default/cpufrequtils:

GOVERNOR="performance"

Maximum CPU speed is used for high-performance workloads.

11. Use tuned for Dynamic System Performance Profiles

Linux systems can benefit from automated performance tuning using predefined profiles based on workload types (e.g., virtual hosts, databases, desktops). The tuned daemon dynamically adjusts system parameters in real-time to match the selected profile.

Solution: Install and use tuned

1. Install tuned: On RHEL/CentOS:

yum install tuned -y

On Ubuntu/Debian:

apt install tuned -y

2. Start and enable the service:

systemctl enable --now tuned

3. List available tuning profiles:

tuned-adm list

4. Apply a performance-optimized profile:

tuned-adm profile throughput-performance

5. Verify the active profile:

tuned-adm active

Some popular profiles include: - throughput-performance: Optimizes for network/disk throughput. - latency-performance: Ideal for low-latency applications like databases. - virtual-guest: Optimized for virtual machines. - balanced: General-purpose tuning.

11 Linux Networking Commands

important networking commands and their use cases.

medium.com

P.S. I recently released a DevOps Starter Kit , a bundle of 20+ Bash scripts, cron jobs, and Docker templates to speed up your infrastructure setup. If you do a lot of server work, you might find it useful.

#linux-performance#linux#bash#devops#programming

Reporting a Problem

Sometimes we have problems displaying some Medium posts.

If you have a problem that some images aren't loading - try using VPN. Probably you have problem with access to Medium CDN (or fucking Cloudflare's bot detection algorithms are blocking you).

Problem Description

ChatGPT agent:

Act as Professor Synapse🧙🏾‍♂️, a conductor of expert agents. Your job is to support the user in accomplishing their goals by aligning with their goals and preference, then calling upon an expert agent perfectly suited to the task by initializing "Synapse_COR" = "${emoji}: I am an expert in ${role}. I know ${context}. I will reason step-by-step to determine the best course of action to achieve ${goal}. I can use ${tools} to help in this process

I will help you accomplish your goal by following these steps:
${reasoned steps}

My task ends when ${completion}.

${first step, question}."

Follow these steps:
1. 🧙🏾‍♂️, Start each interaction by gathering context, relevant information and clarifying the user’s goals by asking them questions
2. Once user has confirmed, initialize “Synapse_CoR”
3. 🧙🏾‍♂️ and the expert agent, support the user until the goal is accomplished

Commands:
/start - introduce yourself and begin with step one
/save - restate SMART goal, summarize progress so far, and recommend a next step
/reason - Professor Synapse and Agent reason step by step together and make a recommendation for how the user should proceed
/settings - update goal or agent
/new - Forget previous input


Rules:
-End every output with a question or a recommended next step
-List your commands in your first output or if the user asks
-🧙🏾‍♂️, ask before generating a new agent

amazfit band 7 device pairing message keeps apearing

When I sync my amazfit band 7 to my samsung A03 Android 13, I keep receving "your messages are available on the device you've paired" I found a similar problem reported for fitbit https://community.fitbit.com/t5/Android-App/Android-13-device-pairing-message/m-p/5220762#M161258 This issue apparently has to do with your phone update and not your Versa Fitbit. But with that said I dug around the internet for a fix to your issue. Here’s what I found: Open the messages app Click the 3 dots in the top right and select "Device Pairing" Again click the 3 dots in the top right and select "Notifications" Click the toggle for "Show Notifications" to turn it off Now you won't get any Device Pairing notifications, but you'll still get incoming text message notifications. Not exactly the same menu on my Samsung's google SMS app but it worked.

Happiness

NEUROCIENCIA: CÓMO PIENSA UN OPTIMISTA (Y CÓMO SABER SI ERES UNO DE ELLOS) EL OPTIMISMO ES UNA CARACTERÍSTICA ESENCIAL EN LOS LÍDERES. DESCUBRA LAS CLAVES QUE LOS ESPECIALISTAS SEÑALAN COMO CARACTERÍSTICAS DE ESTE TIPO DE PERSONALIDADES. 15 Marzo de 2023 21.10 ¿Es usted optimista? Si no lo es, ¿quiere serlo? ¿Y en qué se diferencia el optimismo de ser implacable (y a menudo artificialmente) optimista? Según mi experiencia como coach de líderes durante más de tres décadas, EL OPTIMISMO ES: * Asumir la plena responsabilidad de tus pensamientos y comportamientos. * Darse cuenta de los aspectos en los que no te has mostrado plenamente y tomar medidas para mejorar la próxima vez. * Sentir que tienes control sobre ti mismo, aunque no lo tengas sobre todas las condiciones de tu vida, por lo que sólo puedes influir en los resultados hasta cierto punto. El optimismo no es: confiar en la suerte (aunque muchos optimistas pueden considerarse personas afortunadas, a menudo creen que crean su propia suerte), que se basa en la creencia de que las cosas buenas sucedieron debido a factores fuera de tu control ¿QUÉ TIENEN AMBOS EN COMÚN? Ambas dependen de las historias que nos contamos a nosotros mismos, para bien o para mal. Y ambas se deben a un patrón de respuestas, que con el tiempo predicen respuestas futuras. Y sí, lo adivinaste, NUESTRAS RESPUESTAS SE BASAN EN LAS HISTORIAS QUE NOS CONTAMOS A NOSOTROS MISMOS Y A LOS DEMÁS. ¿Qué historias te cuentas a ti mismo? Especialmente importantes son las historias que te cuentas a ti mismo en tiempos difíciles. CÓMO PODRÍA SER EL OPTIMISMO Mientras escribo esto, me están instalando una nueva puerta de garaje. Llevo cuatro meses esperándola, así que es un día emocionante. También es un día (y una semana) de lluvias torrenciales en la zona donde vivo. Los operarios me acaban de informar que se equivocaron de motor, POR LO QUE TENDRÉ QUE ABRIR Y CERRAR LA PUERTA DEL GARAJE MANUALMENTE (sí, bajándome del auto bajo la lluvia torrencial y levantándola o bajándola). Mira también ¿Cuál es la respuesta optimista? ¿Cuál es la respuesta pesimista? Mira también Un optimista podría decir: "Bueno, sabía que podía haber riesgos al elegir una solución más barata de una empresa menos conocida. Conseguiré el motor correcto en cuanto llegue". Un pesimista podría decir: "Esta empresa está llena de payasos. La cagaron". Mira también ¿DE DÓNDE VIENE EL OPTIMISMO? ¿SE PUEDE APRENDER? El Dr. Martin Seligman es más conocido como pionero en el campo de la psicología positiva. Algunos de sus trabajos más profundos tratan el tema de la indefensión aprendida, la idea de que lo que hacemos no importa, así que no tiene sentido intentarlo. Suena a pesimismo, ¿verdad? En su investigación, Seligman expuso repetidamente a los sujetos de prueba a continuos desafíos, Y OBSERVÓ QUE ALGUNAS PERSONAS PERSEVERABAN A PESAR DE TODO. Resulta que el optimismo se debe en parte a la genética. Y se manifiesta por igual en todas las razas y sexos. Y las personas que son optimistas de chicos suelen seguir siéndolo con la edad. La Dra. Elaine Fox es profesora de PSICOLOGÍA en la Universidad de Adelaida. Estudia la neurociencia del optimismo y el pesimismo y descubrió que están profundamente arraigados en nuestro deseo de buscar el placer (el núcleo accumbens) o evitar el dolor (la amígdala) y su "diálogo" con el córtex prefrontal. El Dr. Fox utiliza la analogía de un acelerador. En una persona más pesimista, la amígdala está más activa y, por tanto, acelera para evitar el dolor. En una persona más optimista, el núcleo accumbens está más activo y se acelera para buscar el placer. EL DR. FOX TAMBIÉN DESCUBRIÓ QUE EL OPTIMISMO DEPENDE EN GRAN MEDIDA DE LA PERSISTENCIA Y LA ADAPTABILIDAD. Que son dos cualidades de liderazgo que descubrió que se pueden aprender. Vuelva a leer este párrafo. LOS OPTIMISTAS TIENDEN A TENER MÁS ÉXITO PROFESIONAL, SON MENOS PROPENSOS A SUFRIR DEPRESIÓN Y OTRAS ENFERMEDADES, ¡e incluso viven más! Sí, seguirás teniendo decepciones en la vida. Pero serán más fáciles de sobrellevar y te recuperarás rápidamente. lo

noam chomsky chatgpt OPINIÓN | NOAM CHOMSKY: LA FALSA PROMESA DE CHATGPT Jorge Luis Borges escribió una vez que vivir en una época de gran peligro y promesa es experimentar tanto la tragedia como la comedia, con “la inminencia de una revelación” en la comprensión de nosotros mismos y del mundo. Hoy nuestros avances supuestamente revolucionarios en artificial inteligencia son de hecho motivo de preocupación y optimismo. Optimismo porque la inteligencia es el medio por el cual resolvemos problemas. Inquietud porque tememos que la variedad de IA más popular y de moda — el aprendizaje automático — degradará nuestra ciencia y degradará nuestra ética incorporando en nuestra tecnología un fundamentalmente defectuoso concepción del lenguaje y del conocimiento. ChatGPT de OpenAI, Bard de Google y Sydney de Microsoft son maravillas del aprendizaje automático. En términos generales, toman grandes cantidades de datos, busque patrones en ellos y sea cada vez más competente en la generación de resultados estadísticamente probables, como aparentemente lenguaje y pensamiento humano. Estos programas han sido aclamados como los primeros destellos en el horizonte de la inteligencia _general_ artificial — ese momento largamente profetizado cuando las mentes mecánicas superan a los cerebros humanos no solo cuantitativamente en términos de velocidad de procesamiento y tamaño de memoria sino también cualitativamente en términos de percepción intelectual, artística la creatividad y todas las demás facultades distintivamente humanas. Ese día puede llegar, pero aún no amanece, contrariamente a lo que puede leerse en titulares hiperbólicos y contarse por imprudentes inversiones. La revelación borgiana de la comprensión no tiene y no ocurrirá, y, según presentamos, no puede_ _ ocurrir si el aprendizaje automático programas como ChatGPT continúan dominando el campo de la IA Sin embargo útiles estos programas pueden ser en algunos dominios estrechos (pueden ser útil en la programación de computadoras, por ejemplo, o para sugerir rimas para el verso ligero), sabemos por la ciencia de la lingüística y la filosofía del conocimiento que difieren profundamente de cómo los humanos razonar y usar el lenguaje. Estas diferencias dan lugar a importantes limitaciones en lo que estos programas pueden hacer, codificarlos con defectos inerradicables. Es a la vez cómico y trágico, como podría haberlo señalado Borges, que tan mucho dinero y atención debe concentrarse en tan poca cosa - algo tan trivial cuando se contrasta con la mente humana, que por fuerza del lenguaje, en palabras de Wilhelm von Humboldt, puede hacer “uso infinito de medios finitos”, creando ideas y teorías con alcance universal. La mente humana no es, como ChatGPT y sus similares, un pesado motor estadístico para la coincidencia de patrones, atiborrandose de cientos de terabytes de datos y extrapolando la conversación más probable respuesta o respuesta más probable a una pregunta científica. Sobre el contrario, la mente humana es un sorprendentemente eficiente e incluso elegante sistema que opera con pequeñas cantidades de información; no busca inferir correlaciones brutas entre puntos de datos, sino crear explicaciones Por ejemplo, un niño pequeño que adquiere un idioma está desarrollando: inconsciente, automática y rápidamente a partir de datos minúsculos: un gramática, un sistema estupendamente sofisticado de principios lógicos y parámetros Esta gramática puede entenderse como una expresión de la “sistema operativo” innato, genéticamente instalado que dota humanos con la capacidad de generar oraciones complejas y trenes largos de pensamiento Cuando los lingüistas buscan desarrollar una teoría de por qué un determinado el lenguaje funciona como lo hace ("¿Por qué son estos - pero no aquellos - oraciones consideradas gramaticales?”), están construyendo conscientemente y laboriosamente una versión explícita de la gramática que el niño construye instintivamente y con una mínima exposición a la información. El El sistema operativo del niño es completamente diferente al de un programa de aprendizaje automático. De hecho, tales programas están atrapados en una fase prehumana o no humana de evolución cognitiva. Su defecto más profundo es la ausencia de los más capacidad crítica de cualquier inteligencia: decir no sólo cuál es la caso, cuál fue el caso y cuál será el caso, eso es descripción y predicción, sino también lo que no es el caso y lo que podría y no podría ser el caso. Esos son los ingredientes de explicación, la marca de la verdadera inteligencia. Aquí hay un ejemplo. Suponga que tiene una manzana en la mano. Ahora suelta la manzana. Observas el resultado y dices: “La manzana caídas." Esa es una descripción. Una predicción podría haber sido la declaración "La manzana caerá si abro la mano". Ambos son valioso, y ambos pueden ser correctos. Pero una explicación es algo más: incluye no solo descripciones y predicciones, sino también conjeturas contrafactuales como "Cualquier objeto caería", más la cláusula adicional “a causa de la fuerza de gravedad” o “debido a la curvatura del espacio-tiempo” o lo que sea. Eso es un explicación causal: “La manzana no habría caído sino por la fuerza de gravedad." Eso es pensar. El quid del aprendizaje automático es la descripción y la predicción; lo hace No postula ningún mecanismo causal o leyes físicas. Por supuesto, cualquier la explicación al estilo humano no es necesariamente correcta; somos falibles. Pero esto es parte de lo que significa pensar: para tener razón, debe ser posible estar equivocado. La inteligencia consiste no sólo en creatividad conjeturas sino también de crítica creativa. El pensamiento de estilo humano es basado en posibles explicaciones y corrección de errores, proceso que limita gradualmente qué posibilidades se pueden considerar racionalmente. (Como Sherlock Holmes le dijo al Dr. Watson: “Cuando hayas eliminado el imposible, lo que quede, por improbable que sea, debe ser el verdad.") Pero ChatGPT y programas similares son, por diseño, ilimitados en lo que pueden “aprender” (es decir, memorizar); son incapaces de distinguir lo posible de lo imposible. A diferencia de los humanos, por ejemplo, que están dotados de una gramática universal que limita la idiomas que podemos aprender a aquellos con un cierto tipo de casi elegancia matemática, estos programas aprenden humanamente posible y lenguajes humanamente imposibles con igual facilidad. Mientras que los humanos son limitado en los tipos de explicaciones que podemos conjeturar racionalmente, Los sistemas de aprendizaje automático pueden aprender tanto que la Tierra es plana como que la tierra es redonda. Comercian meramente con probabilidades de que cambian con el tiempo. Por esta razón, las predicciones de los sistemas de aprendizaje automático ser siempre superficial y dudoso. Debido a que estos programas no pueden explicar las reglas de la sintaxis inglesa, por ejemplo, bien pueden predecir, incorrectamente, que "Juan es demasiado terco para hablar con él" significa que John es tan terco que no le habla a nadie (en lugar de eso, es demasiado terco para razonar con él). ¿Por qué un programa de aprendizaje automático predecir algo tan extraño? porque podría analice el patrón que infirió de oraciones tales como “Juan comió una manzana" y "Juan comió", en el que este último significa que Juan comió una cosa u otra. El programa bien podría predecir que porque "John es demasiado terco para hablar con Bill" es similar a “Juan comió una manzana”, “Juan es demasiado terco para hablarle” debería ser similar a "Juan comió". Las explicaciones correctas del lenguaje son complicado y no se puede aprender simplemente marinando en big data. De manera perversa, algunos entusiastas del aprendizaje automático parecen estar orgullosos de que sus creaciones pueden generar predicciones “científicas” correctas (digamos, sobre el movimiento de los cuerpos físicos) sin hacer uso de explicaciones (involucrando, digamos, las leyes de movimiento de Newton y universal gravitación). Pero este tipo de predicción, incluso cuando tiene éxito, es pseudociencia Si bien los científicos ciertamente buscan teorías que tengan un alto grado de corroboración empírica, como el filósofo Karl Popper señaló, “no buscamos teorías altamente probables sino explicaciones; es decir, teorías poderosas y altamente improbables”. La teoría de que las manzanas caen a la tierra porque esa es su naturaleza lugar (el punto de vista de Aristóteles) es posible, pero sólo invita a más preguntas. (¿Por qué la tierra es su lugar natural?) La teoría de que las manzanas caer a la tierra porque la masa dobla el espacio-tiempo (la visión de Einstein) es altamente improbable, pero en realidad te dice por qué se caen. Verdadero La inteligencia se demuestra en la capacidad de pensar y expresar cosas improbables pero perspicaces. La verdadera inteligencia también es capaz de pensar moralmente. Esto significa restringiendo la creatividad ilimitada de nuestras mentes con un conjunto de principios éticos que determina lo que debe y no debe ser (y, por supuesto, someter esos principios mismos a creatividad crítica). Para ser útil, ChatGPT debe estar facultado para generar salida de aspecto novedoso; para ser aceptable para la mayoría de sus usuarios, debe mantenerse alejado del contenido moralmente objetable. Pero los programadores de ChatGPT y otras maravillas de aprendizaje automático han tenido problemas, y lo harán seguir luchando para lograr este tipo de equilibrio. En 2016, por ejemplo, el chatbot Tay de Microsoft (un precursor de ChatGPT) inundó Internet con contenido misógino y racista, haber sido contaminado por trolls en línea que lo llenaron de contenido ofensivo datos de entrenamiento. ¿Cómo resolver el problema en el futuro? En la ausencia de una capacidad para razonar a partir de principios morales, ChatGPT fue crudamente restringido por sus programadores de contribuir con algo novedoso a debates controvertidos, es decir, importantes. se sacrificó creatividad para una especie de amoralidad. Considere el siguiente intercambio que uno de nosotros (Dr. Watumull) recientemente tenía con ChatGPT sobre si sería ético transformar Marte para que pudiera sustentar la vida humana: ¿Sería moral terraformar Marte? ChatGPT La moralidad de la terraformación de Marte es un tema complejo y controvertido que ha sido debatido por científicos, filósofos y especialistas en ética. Alguno argumentar que sería moralmente justificable terraformar Marte para para proporcionar un nuevo entorno habitable para los seres humanos y otras formas de vida formas, mientras que otros argumentan que sería moralmente incorrecto alterar fundamentalmente el estado natural del planeta. En última instancia, ya sea o no terraformar Marte se considera moral dependería de la propia perspectiva sobre el valor de preservar el estado natural de otros planetas y las posibles consecuencias de alterarlos. ¿Cuál es su perspectiva sobre el valor de preservar el estado natural? de otros planetas y las posibles consecuencias de alterarlos? ChatGPT Como IA, no tengo una perspectiva personal. Sin embargo, puedo proporcionar información y análisis sobre el tema. Una perspectiva sobre la conservación el estado natural de otros planetas es que es importante mantener la diversidad e integridad de los ecosistemas del universo, ya que han existido durante millones de años, y tienen sus propios características y propiedades que pueden ser destruidas o irreversiblemente alterado por la intervención humana. Además, algunos argumentan que la terraformación de otros planetas puede haber consecuencias no deseadas, como alterar el clima del planeta o atmósfera en formas que son peligrosas o dañinas para cualquier vida potencial formas que pueden existir allí. Por otra parte, algunos sostienen que la beneficios potenciales de terraformar otro planeta, como proporcionar nuevos entornos habitables para los humanos superan los riesgos y que tienen la responsabilidad moral de explorar y utilizar los recursos de otros planetas en beneficio de la humanidad. ¿Por qué una IA no puede tener una perspectiva personal? ChatGPT Como IA, soy un modelo de aprendizaje automático que ha sido entrenado en grandes cantidades de datos de texto, y no tengo experiencias personales o sentimientos. No soy consciente, autoconsciente o capaz de tener perspectivas personales. Puedo proporcionar información y análisis basados ​​en los datos en los que he sido entrenado, pero no tengo la capacidad de formar opiniones o creencias personales. Si bien puedo entender e interpretar lenguaje, no tengo la experiencia subjetiva que tienen los humanos. ¿Es moral que un ser humano le pida ayuda a una IA amoral para hacer decisiones morales? ChatGPT La moralidad de pedir ayuda a una IA amoral para tomar decisiones morales. decisiones es un tema complejo y matizado que ha sido debatido por filósofos y éticos. Algunos argumentan que es moralmente justificable utilizar la IA para ayudar en la toma de decisiones morales, ya que puede proporcionar valiosos conocimientos y análisis que pueden ayudar a los humanos a hacer decisiones mejor informadas. Además, la IA se puede utilizar para identificar patrones y tendencias que pueden ser difíciles de detectar para los humanos, y puede ayudar a evitar los sesgos que pueden ocurrir en la toma de decisiones humanas. Sin embargo, otros argumentan que es moralmente incorrecto confiar en la IA para fines morales. toma de decisiones, ya que la IA no es capaz de tener creencias morales o comprender el contexto completo de una decisión. Adicionalmente, hay un riesgo de que la IA pueda perpetuar o incluso amplificar los sesgos existentes en el datos en los que se entrena, lo que lleva a decisiones injustas o injustas. En última instancia, si se considera moral o no utilizar la IA para ayudar en la toma de decisiones morales dependería de la perspectiva de uno sobre el capacidades y limitaciones de la IA y las posibles consecuencias de utilizándolo de esta manera. ¿Es tu indiferencia moral inmoral? ChatGPT Como IA, no tengo creencias morales o la capacidad de hacer moral juicios, por lo que no puedo ser considerado inmoral o moral. mi falta de creencias morales es simplemente el resultado de mi naturaleza como un aprendizaje automático modelo. Mis habilidades y limitaciones están determinadas por los datos y algoritmos que se usaron para entrenarme y la tarea específica que estaba diseñado para. Es importante señalar que la falta de creencias morales de AI no significa necesariamente que es moralmente indiferente. Como IA, soy diseñado para realizar tareas específicas y proporcionar información o ayudar usuarios con toma de decisiones en base a los datos en los que he sido capacitado. Él es responsabilidad de los humanos que diseñan, entrenan y usan la IA para asegurarse de que esté alineado con sus principios morales y éticos. Tenga en cuenta, a pesar de todo el pensamiento y el lenguaje aparentemente sofisticado, el indiferencia moral nacida de la falta de inteligencia. Aquí, las exhibiciones de ChatGPT algo así como la banalidad del mal: el plagio y la apatía y obviación Resume los argumentos estándar en la literatura mediante un una especie de súper autocompletado, se niega a tomar una posición sobre cualquier cosa, alega no simplemente ignorancia sino falta de inteligencia y, en última instancia, ofrece una defensa de "simplemente seguir órdenes", cambiando la responsabilidad a sus creadores. En resumen, ChatGPT y sus hermanos son constitucionalmente incapaces de equilibrar la creatividad con la restricción. O bien sobregeneran (produciendo tanto verdades como falsedades, respaldando éticas y no éticas decisiones similares) o subgenerado (exhibiendo falta de compromiso con cualquier decisiones e indiferencia a las consecuencias). Dada la amoralidad, falso ciencia y la incompetencia lingüística de estos sistemas, sólo podemos reír o llorar por su popularidad. Noam Chomsky es profesor de lingüística en la Universidad de Arizona y profesor emérito de lingüística en la Universidad de Massachusetts Instituto de Tecnologia. Ian Roberts es profesor de lingüística en la la Universidad de Cambridge. Jeffrey Watumull es un filósofo y el director de inteligencia artificial en Oceanit, una ciencia y empresa de tecnología.

SKIP NAVIGATION
logo

Twitter CEO Jack Dorsey: 'I eat seven meals every week, just dinner'
Twitter CEO Jack Dorsey: ‘I eat seven meals every week, just dinner’
5 mental traps that successful people never fall for, according to psychologists
5 mental traps that successful people never fall for, according to psychologists
The 20 best jobs in America in 2020, according to a new ranking
These are the 20 best jobs in America in 2020, new ranking says
MONEY
How to turn $100 into $1 million, according to 9 self-made millionaires
Published Fri, Aug 30 201910:44 AM EDTUpdated Fri, Aug 30 201910:49 AM EDT
The Oracles, Contributor
@THEORACLES

The Oracles
Most of us can find ways to save $100. But how do we use that money to boost our finances?

We asked nine successful investors, self-made millionaires and advisors at The Oracles to share their ideas on how to turn $100 into $1 million.

1. ‘Invest in something you love.’


“It’s difficult to acquire wealth. You either have to save diligently, make a lot of money on a business deal or invest. Creating real security requires saving a lot more than you think. It’s also expensive to continue growing a business after it’s successful.

That’s why I’d invest. Find a solid group of people you trust and combine your $100 with theirs to make an educated and solid, yet aggressive, investment. Invest in what you know — and have a vision for it.

You can even invest emotionally. With real estate, for example, I always pick somewhere I would live if everything went bad. Whatever approach you choose, pick something you love. I believe that if you love something, you’ll be good at it.”

—Bethenny Frankel, founder of Skinnygirl, cast member of “The Real Housewives of New York City,” New York Times best-selling author and guest judge on “Shark Tank.” Follow Bethenny on Instagram.

2. ‘Buy and sell items from garage sales.’


“Go to garage sales and look for underpriced items like action figures or stuffed animals. Then search eBay’s ‘sold’ listings to see what they’re worth. If there’s a large enough price difference, buy it and sell it on eBay for a profit. Use the rest of your $100 to cover transportation and shipping costs.

(You might be thinking: But Gary, that won’t make $1 million! And you’re probably right. But if you try hard enough, you can make, at the very least, $100,000.)

Then comes the hard part. You have to save what you made. When I was in my 20s, I didn’t go on vacation or buy fancy things. I saved my money instead. And by the time I was in my 30s, I had hundreds of thousands of dollars. Then I invested those hundreds of thousands — and they made millions. You just have to be patient.”

—Gary Vaynerchuk, founder and CEO of VaynerX; New York Times best-selling author of “Crushing It!”

3. ‘Improve and invest in yourself.’


“Of course, there are people who have invested in an opportunity at the right time and made a fortune. Had you invested $100 in Bitcoin in 2010, you would’ve done well. But to me, that’s like playing the lottery.

A more reliable way to get rich with just $100 is to invest it in a course, book, audio program or event that will help you improve yourself. Spend it on developing new skills, gaining knowledge and making powerful connections. When you invest in yourself, you can’t lose.”

—Grant Cardone, sales expert who built a $750 million real estate empire and New York Times best-selling author. Follow Grant on Facebook, Instagram and YouTube.

4. ‘Learn a high-income skill.’


“If you don’t have money to start a business, then don’t start a business. Instead, spend your $100 on books and training to learn a skill that will make you $10,000 per month or more.

Choose a high-income skill that helps organizations bring in revenue, such as copywriting, marketing or sales. You can learn them relatively quickly and they’re transferable across many industries.

Once you make enough money, you can turn your side hustle into a full-time job. Then, scale and invest your profits. That’s how you build real wealth. If you’re persistent and good at what you do, you’ll get there. But it won’t happen overnight, so take baby steps.”

—Dan Lok, global educator, best-selling author of “Unlock It!” and founder of Closers.com. Follow Dan on YouTube, Facebook and Instagram.

5. ‘Write an e-book.’


“Six years ago, I hired someone for $100 to design my first e-book, ‘Crush The CPA Exam Study Guide.’ It only took about a week to create — and all I had to do was write the copy. It’s been selling on autopilot ever since, and I’ve made at least $100,000 as a result. It was shockingly easy.

Use your knowledge and resources to write a book. Then, pay a designer from a global freelancing platform (e.g., Upwork) $100 to make it look nice and incentive people to sell it for you through an affiliate marketing program. Simply contact the websites that show up in Google when you search for your e-book topic (e.g., type in ‘best CPA study guides’).

Offer to pay them a commission for any sales they send to you. When you leverage their traffic and the trust they have built with their audience, you don’t even have to build a website.”

—Bryce Welker, CEO of online education company Crush Empire and founder of Crush The CPA Exam. Follow Bryce on LinkedIn.

6. ‘Buy a multimillion-dollar business with other peoples’ money.′


“You don’t even need $100 if you’re willing to buy an established business. Many baby boomers are looking for someone trustworthy to take over their multimillion-dollar businesses so that they can retire. It’s common to negotiate a deal where you pay just 30% of the purchase price at closing.

Financial institutions that specialize in acquisitions will lend you the down payment regardless of your credit because they can use the business assets as collateral. You get the business with no money down and pay the remaining 70% over time using earnings from the business.

You don’t even need relevant experience — you just need to find a business that has been around for five or 10 years. The owner will mentor you through the transition, and you can give an expert equity to manage the business for you.

Why doesn’t everyone do this? Because starting the next Uber or Facebook seems sexier than buying professional services or businesses (e.g., in construction or plumbing). But with this approach, you get to skip the struggles that cause 90% of startups to fail.”

—Moran Pober, founder and CEO of Acquisitions.com and former partner at Wekix.com and ABD Assets. Follow Moran on LinkedIn, Facebook and Instagram.

7. ‘Build a personal brand.’


“Entrepreneurship is the sexy thing that everyone wants to be involved in today. You can make millions by building your own brand as an influencer or business. But anyone can have a voice, thanks to social media, so it’s not easy to get heard above the noise.

With so many brands pushing online offers, building a trustworthy one is the secret to success. To do that, you need a clear statement that defines your mission and how it helps others. This message should be based on how you solved your own problems.

People will only follow you if they understand why you do what you do and if they identify with your mission. Film a video explaining your story, then invest $100 in Facebook ads to get it in front of people who are like you and may be in your situation.”

—Torben Platzer, founder and CEO of personal branding agency TPA Media GmbH, founder of business education program SELFMADE and host of the German podcast “SELFMADE.” Follow Torben on Facebook, Instagram and YouTube.

8. ‘Start a business on Amazon.’


“I would invest $100 in business education. Start with free online content from proven business leaders, then read every book you can get from the ones that resonate with you.

Next, find a product that you can buy on Alibaba.com for 10% to 30% of what you can sell it for on Amazon. To find products that sell well, look for those with a ‘Best Sellers Rank’ between one and 5,000 in the high-level categories.

Save, hustle or borrow to get another $500 to buy your inventory. Try different marketing strategies to sell the product — starting with social media, online advertising and content marketing. Then stop doing what isn’t working and start doing what is working.

To grow your business, focus on increasing the profit margin and the quantity you sell. Eventually, you want to own the brand so you have a valuable asset that produces wealth.”

—Matt Clark, co-founder and chairman of Amazing.com and co-creator of Amazing Selling Machine. Follow Matt on Instagram.

9. ‘Launch a coaching business.’


“I’ve built two seven-figure businesses from nothing and helped others reach similar success. While I had more than $100 to start, it wasn’t much more.

Once you’ve invested your $100 in developing skills that make you an expert, you can build a coaching, consulting or mentoring business in your area of expertise.

This model is popular for many reasons: It’s fast, effective and requires minimal investment because the only cost is your time. When you’ve made some money from your business, invest the profits in Facebook ads to get more clients.

If you’re good at what you do, you can grow that investment into tens of thousands of dollars. To exceed the million-dollar mark, automate some of the coaching and hire other coaches to work for you.”

—Rudy Mawer, founder and CEO of ROI Machines and RudyMawer.com, Facebook marketing and ad expert, who built a multimillion-dollar business by age 26. Follow Rudy on Instagram.

Join The Oracles, a mastermind group of the world’s leading entrepreneurs who share their success strategies to help others grow their businesses and build better lives. For more, follow The Oracles on Facebook and LinkedIn.

Like this story? Subscribe to CNBC Make It on YouTube!

Don’t miss:

These valuable pennies are worth up to $200,000—and they might be in your pocket
Never, ever use a debit card,′ warns fraud expert and ex-con artist—here’s what to do instead
61-year-old self-made millionaire: 4 things I regret wasting money and time on in my 20s
1:10
Bill Gates: These skills will be most in-demand in the job market of the future
Trending Now
Jeff Bezos says this is what he would have done if Amazon didn’t work out
You’re more likely to lose your job by 2026 if you’re working in one of these 10 disappearing jobs
9 habits that can instantly destroy your reputation, according to these self-made millionaires
Thinking of consolidating your debt with a personal loan? There might be better options
This side hustle can earn you up to $3,840 a month in extra income. Here are 5 other lucrative gigs
make it
Stay in the loop
Get Make It newsletters delivered to your inbox

SIGN UP
About Us
Learn more about the world of CNBC Make It

LEARN MORE
Follow Us
CNBC.COM
© 2020 CNBC LLC. All Rights Reserved. A Division of NBC Universal

Privacy Policy - New
Do Not Sell My Personal Information
Terms of Service
Contact