The Everything Store – Jeff Bezos and the Age of Amazon by Brad Stone is a book detailing some factors that led to the rise of Amazon as one of the largest corporate success stories of all time. I opened it expecting to skim through some parts, but ended up reading it in full in one sitting, and enjoyed it thoroughly. It left me with a strong sense of what makes Amazon, well, Amazon. And the best answer to that question is without a doubt, Bezos himself.
Rather than a full book review, I’m going to share some quotes from The Everything Store that stood out to me. One fun thing to note is that when this book was published in 2013, Amazon was ‘only’ a $150B company, but today is worth over 1.5 trillion. It’s a wonderful book and is worth buying if you want to read stories about Jeff Bezos’ extreme confidence in himself and his company as they overcome challenges one after another full-speed ahead, from local stores to Barnes and Noble to Ebay to Walmart and beyond. Below are some quotes that I particularly liked, only from the first fourth of the book. I was quoting the book more than anticipated, so I stopped this post early but will leave it up as an advertisement for the book.
Bezos is an excruciatingly prudent communicator for his own company. He is sphinxlike with details of his plans, keeping thoughts and intentions private, and he’s an enigma in the Seattle business community and in the broader technology industry. He rarely speaks at conferences and gives media interviews infrequently.
There is so much stuff that has yet to be invented. There’s so much new that’s going to happen. People don’t have any idea yet how impactful the Internet is going to be and that this is still Day 1 in such a big way. Jeff Bezos
Amazon’s internal customs are deeply idiosyncratic. PowerPoint decks or slide presentations are never used in meetings. Instead, employees are required to write six-page narratives laying out their points in prose, because Bezos believes doing so fosters critical thinking. For each new product, they craft their documents in the style of a press release. The goal is to frame a proposed initiative in the way a customer might hear about it for the first time. Each meeting begins with everyone silently reading the document, and discussion commences afterward
“If you want to get to the truth about what makes us different, it’s this,” Bezos says, veering into a familiar Jeffism: “We are genuinely customer- centric, we are genuinely long-term oriented and we genuinely like to invent. Most companies are not those things. They are focused on the competitor, rather than the customer. They want to work on things that will pay dividends in two or three years, and if they don’t work in two or three years they will move on to something else. And they prefer to be close-followers rather than inventors, because it’s safer. So if you want to capture the truth about Amazon, that is why we are different. Very few companies have all of those three elements.”
Bezos interpolated from this that Web activity overall had gone up that year by a factor of roughly 2,300—a 230,000 percent increase. “Things just don’t grow that fast,” Bezos later said. “It’s highly unusual, and that started me thinking, What kind of business plan might make sense in the context of that growth?”
Jackie Bezos suggested to her son that he run his new company at night or on the weekends. “No, things are changing fast,” Bezos told her. “I need to move quickly.”
Internet records show that during that time, they registered the Web domains Awake.com, Browse.com, and Bookmall.com. Bezos also briefly considered Aard.com, from a Dutch word, as a way to stake a claim at the top of most listings of websites, which at the time were arranged alphabetically.
Bezos and his wife grew fond of another possibility: Relentless.com. Friends suggested that it sounded a bit sinister. But something about it must have captivated Bezos: he registered the URL in September 1994, and he kept it. Type Relentless.com into the Web today and it takes you to Amazon.
They set up shop in the converted garage of Bezos’s house, an enclosed space without insulation and with a large, black potbellied stove at its center. Bezos built the first two desks out of sixty-dollar blond-wood doors from Home Depot, an endeavor that later carried almost biblical significance at Amazon, like Noah building the ark.
During that time, the name Cadabra lived on, serving as a temporary placeholder. But in late October of 1994, Bezos pored through the A section of the dictionary and had an epiphany when he reached the word Amazon. Earth’s largest river; Earth’s largest bookstore.3 He walked into the garage one morning and informed his colleagues of the company’s new name. He gave the impression that he didn’t care to hear anyone’s opinion on it, and he registered the new URL on November 1, 1994. “This is not only the largest river in the world, it’s many times larger than the next biggest river. It blows all other rivers away,” Bezos said.
One early challenge was that the book distributors required retailers to order ten books at a time. Amazon didn’t yet have that kind of sales volume, and Bezos later enjoyed telling the story of how he got around it. “We found a loophole,” he said. “Their systems were programmed in such a way that you didn’t have to receive ten books, you only had to order ten books. So we found an obscure book about lichens that they had in their system but was out of stock. We began ordering the one book we wanted and nine copies of the lichen book. They would ship out the book we needed and a note that said, ‘Sorry, but we’re out of the lichen book.’
A week after the launch, Jerry Yang and David Filo, Stanford graduate students, wrote them an e-mail and asked if they would like to be featured on a site called Yahoo that listed cool things on the Web. At that time, Yahoo was one of the most highly trafficked sites on the Web and the default home page for many of the Internet’s earliest users.
In the meetings, Bezos presented what was, at best, an ambiguous picture of Amazon’s future. At the time, it had about $139,000 in assets, $69,000 of which was in cash. The company had lost $52,000 in 1994 and was on track to lose another $300,000 that year. Against that meager start, Bezos would tell investors he projected $74 million in sales by 2000 if things went moderately well, and $114 million in sales if they went much better than expected. (Actual net sales in 2000: $1.64 billion.)
Bezos later told the online journal of the Wharton School, “We got the normal comments from well-meaning people who basically didn’t believe the business plan; they just didn’t think it would work.”11 Among the concerns was this prediction: “If you’re successful, you’re going to need a warehouse the size of the Library of Congress,” one investor told him.
When his goals did slip out, they were improbably grandiose. Though the startup’s focus was clearly on books, Davis recalls Bezos saying he wanted to build “the next Sears,” a lasting company that was a major force in retail. Lovejoy, a kayaking enthusiast, remembers Bezos telling him that he envisioned a day when the site would sell not only books about kayaks but kayaks themselves, subscriptions to kayaking magazines, and reservations for kayaking trips—everything related to the sport. “I thought he was a little bit crazy,” says Lovejoy.
The IPO process was painful in another way: During the seven-week SEC-mandated “quiet period,” Bezos was not permitted to talk to the press. “I can’t believe we have to delay our business by seven years,” he complained, equating weeks to years because he believed that the Internet was evolving at such an accelerated rate. Staying out of the press soon became even more difficult. Three days before Amazon’s IPO, Barnes & Noble filed a lawsuit against Amazon in federal court alleging that Amazon was falsely advertising itself to be the Earth’s Largest Bookstore. Riggio was appropriately worried about Amazon, but with the lawsuit he ended up giving his smaller competitor more attention. Later that month, the Riggios unveiled their own website, and many seemed ready to see Amazon crushed. The CEO of Forrester Research, a widely followed technology research firm, issued a report in which he called the company “Amazon.Toast.”
It was a distilled version of the dissatisfaction felt by many early Amazon employees. With his convincing gospel, Bezos had persuaded them all to have faith, and they were richly rewarded as a result. Then the steely-eyed founder replaced them with a new and more experienced group of believers. Watching the company move on without them gave these employees a gnawing sensation, as if their child had left home and moved in with another family. But in the end, as Bezos made abundantly clear to Shel Kaphan,family. But in the end, as Bezos made abundantly clear to Shel Kaphan, Amazon had only one true parent.
“You seem like a really nice guy, so don’t take this the wrong way, but you really need to sell to Barnes and Noble and get out now,” one student bluntly informed Bezos. Brian Birtwistle, a student in the class, recalls that Bezos was humble and circumspect. “You may be right,” Amazon’s founder told the students. “But I think you might be underestimating the degree to which established brick-and-mortar business, or any company that might be used to doing things a certain way, will find it hard to be nimble or to focus attention on a new channel. I guess we’ll see.”
“There will be a proliferation of companies in this space and most will die. There will be only a few enduring brands, and we will be one of them.”
During that time, no one placed bigger, bolder bets on the Internet than Jeff Bezos. Bezos believed more than anyone that the Web would change the landscape for companies and customers, so he sprinted ahead without the least hesitation. “I think our company is undervalued” became another oft- repeated Jeffism. “The world just doesn’t understand what Amazon is going to be.”
As the company grew, Bezos offered another sign that his ambitions were larger than anyone had suspected. He started hiring more Walmart executives.
Around that time, Wright showed Bezos the blueprints for a new warehouse in Fernley, Nevada, thirty miles east of Reno. The founder’s eyes lit up. “This is beautiful, Jimmy,” Bezos said. Wright asked who he needed to show the plans to and what kind of return on investment he would have to demonstrate. “Don’t worry about that,” Bezos said. “Just get it built.” “Don’t I have to get approval to do this?” Wright asked. “You just did,” Bezos said. Over the next year, Wright went on a wild $300 million spending spree.
“Walmart did not even have Internet in the building back then,” says Kerry Morris, a product buyer who moved from Walmart to Amazon. “We weren’t online. We weren’t e-mailing. None of us even knew what he meant by online retail.”
The venture capitalists backing eBay asked around and heard that one did not work with Jeff Bezos; one worked for him.
Bezos went skiing in Aspen that winter with Cook and Doerr and finally told them what was coming. “He said, ‘We’re going to win, so you probably want to consider whether to stay on the eBay board,’ ” says Cook. “He thought it would be the only natural outcome.”
If you liked these quotes, consider reading the full copy (perhaps even buying it from Amazon), it’s definitely a nice read about an amazing company and individual.
Midday on July 15th, 2020, many high-profile Twitter accounts were compromised and began posting scams to entice users into sending them cryptocurrency (generally BTC, but also some others such as XRP for Ripple’s account). I’m not going to write about this in detail since everyone else already has, but for more information check out an article on the topic: Coindesk, TheVerge, TechCrunch, BBC, infinite others
What if instead of posting low-quality cryptocurrency scams, the attackers did something else?
Sure, they could have tried to use CEO accounts such as Musk and Bezos to make millions (or possibly billions) on the stock market by tweeting about earnings and purchasing large amounts of far-out-of-the-money near-expiration call options on the underlying stocks. But we have a lot of ways to catch people that try that, and many regulations and organizations that would make it more difficult (many more than just the SEC) to get away with (although as a side note, $TLSA’s stock+option trading volume is absurdly high, and it would be very difficult).
But, what if they had tried something else entirely, something not motivated by short-term financial gain?
What if the attackers wanted to cause chaos and violence, perhaps alongside putting certain political movements and goals forward? What if they had pre-written thousands of tweets about a topic, perhaps a fake and outrageous event occurring, paired with fake images and videos, perhaps even some higher-quality deepfakes? How many people could they get killed? Could they start a war?
You might think this sounds absurd at first glance. But remember, most of the world’s most influential people use Twitter, including the leaders of most national governments. Although a private corporation that plays by its own rules, Twitter is still the means with which many elected officials communicate with the public. Entire social movements have started and ended through the power of a single viral tweet, sometimes resulting in significant violence or many deaths. Social media platforms have been used by extremists of every type imaginable in the past, and this isn’t going to stop any time soon.
What if the next exploit affects much more than some Twitter accounts?
But, I want to go much further than talking about Twitter. What if instead of an exploit that allowed attackers to compromise Twitter accounts, it had been something much worse? What if they were able to compromise any web server, or any online Windows machine, or industrial control systems for utilities, power plants and military operations? None of these scenarios are by any means impossible. Enough software and hardware exists at enough layers of abstraction that there’s generally always 0-days lurking in critical systems, sometimes for years or decades, before they’re found. We know that 0days are found often by securityresearchers, private companies, governments, and others (sometimes rewarding up to $2,500,000), but also that they are less commonly exploited in obnoxious and harmful ways (generally being hoarded by government security agencies or reported in good faith).
We were unprepared for covid despite epidemics throughout all of history
It was said by many that the covid pandemic could have been predicted, in a sense (which is why it was not a true black swan event). Perhaps not the specifics of it such as the date, virus, and origin. But the general idea of “at some point in the future, something bad is going to happen like this, and we need to prepare for it.”
Another one of these “something really bad is going to happen in the future” categories involves cybersecurity, data privacy, and AI. Just one of these topics individually can be involved in a terrible catastrophe, and indeed have been before, but I think we’re coming close to a combination of all three that can lead to events much worse than we’re currently prepared for.
Security: Billions of humans live digital lives, including the most influential, famous, and dangerous. These people all have email accounts, phones, Twitter accounts, and more, all of which can be compromised, controlled, and manipulated by others.
Data Privacy: The amount of data that social media giants (among others) have on most people is massive, and in my opinion vastly underestimated both in quantity and power. The majority of human communication is now owned by private companies that store things forever. A large proportion of all human social connections, conversions, movements, opinions, and thoughts are stored in databases that not only will not forget, but that the user does not have any control or often even knowledge of.
AI: Advances in the area of content generation have been happening very quickly in the last few years. We now have GPT-3, which can write plenty ofthingsbetter than humans can. We have deepfakes, which can produce believable fake images and videos. We can do the same for voices and much more. Much of this isn’t yet perfect, but it’s clear that we’re improving quickly.
So, take the three above topics of security, data privacy, and AI, and combine them all. Bonus points if you throw in some political tension, which we’re certainly not lacking right now either.
We are not prepared for a true disaster involving technology
As a society, we’re woefully under-prepared for disasters in all of these areas.
We’re not prepared for critical infrastructure, both physical and digital, to be compromised or attacked by highly-funded and competent groups, maybe even state-ran.
Not prepared for the massive campaigns of disinformation, fake news, and propaganda that lie ahead. If you thought things were bad in the last few years, just wait, because we’re on the verge of accelerating it by 10x, and fact-checking is not a solution. China’s government seems to be working very hard both on the offensive and defensive here. Is anyone else truly competing?
Not prepared for how to deal with database leaks that will contain the life history of millions of people, including their ‘private’ conversations and deepest secrets, and items so egregious that they instantly spark violence. Plenty of data breaches have led to murders and suicides already. There are still many countries where you can face imprisonment or death for being gay, being atheist, being of a certain ethnicity, or speaking outagainst thegovernment (yes, we really don’t have it as bad here, huh!). Do you know what happens when these people have their private information carelessly leaked? It’s not pretty. And this is just for normal database leaks, let alone if a database leak had some information in it falsified (with the majority left intact, thus offering plausibility for the fake parts) to maximize its effect.
Not prepared for how to face that humanity is becoming increasingly controlled by viral algorithms that do not prioritize human values of happiness and love and truth, but rather nothing but outrage and in-group bias as the only bottom line. Most of us already feel powerless against this, but it may only just be beginning.
Not prepared for how anonymity is becoming a luxury only achievable by ultra-competent tech gurus, with most people having been forced to move their communication into more and more centralized ways over time, feeding all of the above issues. Not prepared for how one of the many reasons anonymity is getting much more difficult to obtain is because the easiest way to tell if someone is a bot or a human is to require verification of phone numbers, addresses, and more. And don’t let me forget to mention how many governments are eyeing up ways to ban end-to-end encryption.
I’m supposed to end on an optimistic note
How can we do a better job of addressing these problems?
Promote education on the importance of cybersecurity, especially at the government and corporate levels
Promote decentralized solutions instead of centralized social media platforms, allowing users to have control over their discourse, their platform, and their own data
Promote anonymity, even when it is difficult, and fight to ensure end-to-end encryption is a right for everyone forever
Promote better regulations around privacy and data security so that hoarding large amounts of personal data is less of an asset and more of a liability
Although a lot of this post might read as alarmist and pessimistic, I’m still (mostly) optimistic about these things in the long-long-term. The best part about terrible events like covid is that they make us stronger and better prepared for the next (similar) storm to hit us. Security used to be a second thought (or not a thought at all) for most companies, but we’ve improved significant in the last decade, and bug bounty programs and significant security spending are now common. I used to get looked at like I was insane for talking about how big of an issue the amount of tracking and data-collecting our society performs was a big problem, but even this is something that a lot of everyday people believe now as well. I just hope the stepping stones along the way to becoming prepared for the future aren’t so terrible that we don’t make it there in one piece.
It’s interesting how many people view fact-checking as a simple problem, where you just identify something that is not factual, then correct it. Problem solved! Misinformation has been defeated, the Internet is only full only of Truth, and now The People finally realize we were right all along!
Fact-checking is a very hard problem. A lot of people want to ignore this fact, because correcting people feels good, especially when they’re your enemies.
It takes a lot of virtuousness, of empathy, of vigor, and of rationality to decline the temptation to correct other people. We are driven insane by the fact that not only are other people wrong on the Internet, but they are wrong about basic facts! This is part of why websites like Twitter are so terrible. People cannot stand others being wrong on the Internet. They will gladly spend hours of their daily life willfully being miserable and angry just to have the chance to correct others, even if those they correct do not even change their beliefs at all, or even change them in the opposite direction.
Humor aside, this is literally what fact-checking is
Although some may think the reason why Mark Zuckerberg has come out against fact-checking politicians is so that he can reap profits and sow division while cackling maniacally, I think instead he has simply put a lot of thought into the problem, and not only realized how difficult it is, but also that it cannot work well long-term. He is much more concerned about the long-term future (decades) of Facebook than he is about some upsetting posts made by an upsetting person.
Difficulties with fact-checking
I. There’s no such thing as an unbiased fact-checker
Fact-checkers, whether humans, scripts, or AI, cannot be unbiased. Reality is always a state of incomplete information, and the ways that humans interpret statements vary from person to person. It’s possible for us to disagree on the veracity of a statement, but if we were to discuss things further with more specificity, actually agree on the state of reality. Many statements can not reasonably be interpreted as a boolean of true or false, and instead have subtle amounts of potential bias and nuance within them. Facts are constantly changing, and no single actor has perfect and unbiased information about all of them.
II. Even if there was, someone has to decide which content should be fact-checked
Even given impossibly-perfect moderation, someone has to cherry-pick the content that is to be moderated and checked in the first place, as the Internet has far too much content to police every thought and post manually. Most individuals in favor of fact-checking tend to focus on a very small subset of individuals or organizations that they think should be fact-checked, but this set itself is cherry-picked according to their preferences and attention. This is another process that inevitably introduces bias, potentially in many directions depending on the people and processes involved. Similar to how two completely opposing news networks may only report true information, but still promote entirely opposing narratives because they cherry-pick what is news and what is not news, fact-checking cannot avoid this selection problem. It’s very unlikely that any large organization can do a reasonable job at this.
III. Even given quality fact-checking, the results may not be what you seek
A lot of people do not trust certain fact-checkers, certain news networks, and especially certain social media companies. Even if you perform good fact-checking, there is little evidence that this will achieve your goal, which is not actually correcting text on the Internet, but correcting peoples’ beliefs in their own minds, which turns out to be pretty difficult. Fact-checking could sometimes have an effect similar to the Streisand effect, potentially even causing harm to one’s cause, although I can’t find recent studies on this specifically. Regardless, it should be well-known by now that many people will not instantly and flawlessly change their minds when presented with new opposing facts, especially when done so by their outgroup.
If an Internet platform undertakes significant fact-checking, it could even drive heavily-affected groups off of the platform and onto their own platform, where they would then be even more free to spread their own information in whichever way they want. Similarly, the amount of trust that is given to many companies could decrease significantly and cause greater problems further down the road that can then not be solved with fact-checking.
Long-term affects of big changes are impossible to predict, but it’s not too hard to think about a lot of unintended consequences not just for social media, but for governments, democracy, and humanity, further down the road. Most people don’t have to try too hard to imagine some pretty dystopian results and major failure modes when trillion dollar corporations and governments become arbiters of truth and information.
IV. alternative narratives are important for society
Fact-checking has been desired and attempted by those in power throughout history, often leading to disastrous results, even without considering the political extremes of events such as WWII, which entire books of tragedy are written on. Many may view the Copernician revolution as ancient history, being 500 years ago, but it was only 170 years ago when Ignaz Semmelweis’ controversial hypothesis that doctors should wash their hands and maintain cleanliness was mocked and ridiculed for decades, until it eventually became common practice and saved millions of lives.
History is full of countless examples of individuals that went against the grain of their encompassing culture in order to accomplish amazing things and drive progress. These people often armed themselves with what may have been originally considered to be misinformation by those in power during their zeitgeist. I’m glad that we didn’t have the centralization of communication and power we have now throughout history, because many rights that you take for granted were only gained thanks to the failures of past powers to control the narrative as strictly as they wanted to.
V. The narrative cannot be controlled
Just like everyone else, I too wish that everyone that was wrong on the Internet could be corrected. I wish my favorite narratives, facts, and causes were supported and known by more people. To not attempt to correct and control the narrative is a tough bullet to bite, but it’s something that a lot of thought and consideration must be put into, requiring very long-term thinking and awareness of history.
As long as people are free, they will come up with their own narratives, causes, desires, and even facts and entire worldviews. It’s been said that we live in a post-truth world, but that has always been the case. It’s just more apparent now that you can see people from every other background and culture when you use the Internet. No single person, company, or government can control information flow and peoples’ beliefs to the extent that they wish, and any that grasp for such an unattainable level of control will find that it doesn’t work long-term. Even the countries with the strictest controls on information, reporting, and speech, historically, have never fared well after a long enough time.
Different people have different lives, different values, and even different facts that they live by. It’s possible for you to co-exist with them, but perhaps not best if you’re forced to live in the same room as them. But no matter how hard you try, it may be impossible to get them to live their life the way you want them to. I know it’s difficult, but sometimes the only option is to let others live in their own world, while you live in yours, still helping to make it the best you can. Spending your days being angry and miserable on social media will not accomplish what you want, no matter how right you are.
Although this post might not be particularly insightful and is almost too political for my taste, I hope at the least this may help some realize that fact-checking is more difficult than it appears at first glance, even if they still support it.
This document is an updated list of most of the supplements/drugs that I take daily, as well as notes on some other interesting substances. It contains information on exactly what I take, how much of it, how much it costs, and some information on the substance which should roughly explain my reasons for taking it.
May 7th 2024 update: I have replaced many of the pills in this post (but not all of them!), with pills from Blueprint, although I don’t consume any of the auxiliary products. I would strongly suggest reading my post and review of Blueprint as it is now an essential addendum to this post. It saves quite a bit of management time as well.
The first list contains supplements I take daily, with the second list containing supplements that I do not take daily but that nonetheless seem interesting, while the third list contains supplements that are interesting, but that seem less suitable for safe human consumption or speculation. As of 2023, many supplements on my list have been discontinued, and this is mentioned next to the dosage.
The focus of my supplementation is to find substances that are both very safe and also have a notable probability of improving health, lifespan, well-being, or productivity, with the ultimate goal being to significantly slow aging, even if it’s difficult to do at this time. I don’t take many nootropics as I don’t think there’s much room for intelligence improvements just from ingesting simple compounds (evolution has already put quite a bit of time into making us smart), with the exception of treating some deficiency or other issue, or improving productivity/concentration, which definitely possible (see: caffeine, modafinil, adderall, many others), but distinct from intelligence. It is worth noting this list is very specific to myself: if I had a perfectly optimal diet and lifestyle, I would likely take next to zero supplements. Like most mortals, my diet and lifestyle are definitely not perfect (and indeed, even knowing what a perfect diet would be for yourself can be intractable on its own), thus there’s always room for improvement.
This post is not an attempt to convince anyone of something
specific or to suggest anything specific, but I have decided to
publish it publicly in order to better keep myself accountable for my
reasoning, receive potential feedback, and to otherwise share some
potentially useful short summaries of information. Concordantly, I’m
not a doctor and this post contains no medical advice or suggestions.
Which supplements, if any, one should take, is a very personal
matter, as it is dependent upon many unique traits such as one’s
age, diet, lifestyle, genes, risk preference, finances, and more.
Notes on supplements
Section tl;dr: most supplement marketing tactics, websites, videos, brands, etc, are bullshit. Do your own research, ask questions, and be quantitative.
Although there are a lot of supplements that would be beneficial to many people, caution must be exercised both with research and purchasing. Supplements in the United States have very little regulation, with some sellers having poor quality control, fraudulent research, marketing, claims, and poor ingredient composition and sourcing. The supplement industry is worth billions of dollars and has many bad actors incentivized by profit over truth, so time and care must be exercised in order to find out what works best for you personally. Certainly, research can be found promising positive effects from thousands of various substances – but taking all of them would be impractical, expensive, and likely downright harmful.
It’s also important to pay attention to brands as well as to think logically about which supplements have quality differentials that are worth paying more for. For example, Vitamin D and Glycine are easy to synthesize, and it’s likely that cheaper versions of these supplements are just as good as more expensive versions. This may not be the case for a supplement like fish oil however, which is derived from complex living organisms that vary significantly on factors such as their environment, diet, quality controls, the types of fish used, and so on.
Concordantly, one of the strongest criteria I look for in most supplements is safety, which many times (not always!) comes alongside popularity and/or significant research affirming the safety of compounds. As many supplements offer marginal benefits at best, it would be irrational to purchase and consume them if they had a good chance of causing harm, as this would easily cause them to fail a basic cost/benefit or risk/reward analysis (there’s definitely some cool compounds that have very high coefficients in both the numerator and denominator of their risk/reward ratio too, so careful decision making is required).
Ideally one should attempt to find quantitative measures to
objectively evaluate if a substance is really helping them in the
desired manner. In some cases this is both easy and cheap to do, for
example with Vitamin D supplementation, which costs only a few cents
a day, does not need to be compared to a placebo, and can be tested
for in your blood for as little as $30. In other cases, proper
testing is difficult or impossible and may require significant effort
and time for very little benefit. Keeping one’s lifestyle, diet,
and other factors a perfect experimental constant is certainly
difficult, as is performing blind experiments on yourself, collecting
and analyzing data, and finding the proper quantitative desideratum
to test yourself on to begin with; testing if something specific has
definitively made you slightly smarter, happier, healthier, more
productive, or extended your lifespan, is certainly difficult if not
occasionally impossible to do in a scientifically rigorous manner
with a sample size of one.
Lastly, which supplements benefit an individual is a very personal matter. Vegans may want to take some supplements that are found in meat. Carnivores may want to take some supplements that are found in plants. Supplements that may benefit the elderly or those with common conditions such as hypercholesterolemia or diabetes often seem to be much less useful for otherwise healthy individuals. Indeed, for individuals that have many health conditions including the elderly, there’s significantly more that can be gained from supplementation, as there are many more problems that can be improved upon (although there are also be more risks as well). Supplements will effect someone differently depending on their weight, age, genetics, health, diet, and many other factors.
This means that it’s a bad idea to copy any individual’s routines completely, even if it’s a lot of work to do your own planning, research, and testing. It is also worth mentioning that the word ‘supplement’ is used here as a relatively generic word, simply meaning that the substance is only regulated as a food within the United States and thus requires no prescription (unless otherwise mentioned), but also offers few guarantees in terms of efficacy or consistency.
If you take 2 or more pills a day, the best pill organizer+dispenser by far is this (Image from Amazon reviews)
See also
If you find this page interesting, here are some similar pages from others that you may enjoy:
Information: Vitamin D3 [Examine, webmd, Wikipedia] (colecalciferol) is a vitamin made by the skin when exposed to sunlight. It’s a common deficiency and is very cheap to fix. The benefits of supplementation are generally found to be minor (it’s still a bit controversial if supplementation is beneficial at all, although I lean towards yes personally), but as I was notably deficient and it’s one of the cheapest supplements, it’s an easy choice for me. There is a lot of literature on Vitamin D, and many highly-powered studies including meta-analysis will often find only minor beneficial effects, but there are also quite a few studies that show notable benefits, including many related to covid as of late 2020. As noted above, this is also an easy supplement to receive a blood test for and ensure you’re taking the optimal dosage. I take half of my vitamin D earlier in the day without a meal, and the other half with my food, contrary to most other supplements. See also: Gwern on Vitamin D as well as on it harming sleep if taken at night
Name: Vitamin B6
Dosage: 1.4mg (included in BP)
Cost/Day: $0.04
Information: Vitamin B6 [Examine, webmd, Wikipedia]. This is included in other supplements I have and is otherwise not notable.
Name: Vitamin B12
Dosage: 125mcg (included in BP)
Cost/Day: $0.04
Information: Vitamin B12 [Examine, webmd, Wikipedia] (vitamin B3). This is included in other supplements I have and is otherwise not notable (although if you’re on metformin or a few others I used to try it’s better to add it, else unlikely to be deficient).
Name: Vitamin C
Dosage: 500mg
Cost/Day: $0.03
Information: Vitamin C [Examine, webmd, Wikipedia] (ascorbic acid) has a variety of effects, and being a vitamin, is an essential part of a human diet. I supplement vitamin C in order to fix a tested deficiency. I’m rarely deficient but it doesn’t hurt.
Name: Fish Oil
Dosage: 1-3g+ (depends on diet and estimated omega-6 intake)
Cost/day: $0.10 (1g)
Information: Fish oil [Examine, webmd, Wikipedia] (omega-3 EPA+DHA) is another common and cheap supplement. Although many studies find minor or sometimes no benefits, many others find a large amount of diverse improvements, even if they are minor. It’s likelythat the ratio of omega-6/omega-3 you consume is important, with most people consuming far too much omega-6 (which won’t hurt to reduce regardless) and not enough omega-3, so dosing of fish oil should be based on your diet, which is easily more than an order of magnitude more important to begin with.
Similarly, it’s probably good advice to 1) reduce fried food intake, 2) replace oils high in linoleic acid such as safflower and sunflower oils with oils that have much less such as coconut oil and olive oil (2021 edit: I am less sure about this than I was before, although I still lean towards it myself Deciding which oils/fats (and with what / prepared in which manner) are bad for you continues to be an extremely hard problem. See A Comprehensive Rebuttal to Seed Oil Sophistry for an example of a comprehensive potential counter-argument in the great seed oil/fat debate) and 3) increase my supplementation of high-quality omega 3s (fish oil) when I think I’ve had more omega 6s. For example, if I do decide to eat a lot of fried food, I take several fish oils, compared to only 1-2g normally. I also like to note that fish oil seems to be one of the supplements worth spending a bit more money on – quality is high-variance and of higher importance, and unlike other supplements which can trivially be synthesized, the production processes of fish oil vary greatly depending on the company and product. Also see this review on pubmed and this summary on Wikipedia
Name: Garlic
Dosage: 1-3g (included in BP now)
Cost/day: $0.02 (1g)
Information: Garlic [Examine, webmd, Wikipedia] is another popular and cheap supplement. There’s good evidence that it improveslipidprofiles, may help with some cancerrisks, and may have other very minor benefits (may activate AMPK too?). The most desirable compound in garlic is allicin, which is diluted in garlic that is microwaved, boiled, or aged. Dosage should be based on which type of garlic is being consumed. As many people enjoy the taste of garlic, it’s a good candidate to include in meals as well (which is probably optimal for most things, resulting in notably higher bioavailability on average).
Name: Olive Leaf
Extract
Dosage: 500mg (discontinued, decided it was too low-value)
Cost/day: $0.02
Information: Olive Leaf Extract [Examine, Wikipedia] is a cheap and easy way to hopefully mimic the benefits of olive oil, as the leaves of the olive tree contain good amounts of relevant olive phenols such asoleuropein. It may still be better to consume olive oil instead, which is still a great thing to add to meals, but with such a low cost, this seems worth inclusion to me. I am not particularly excited about this supplement but have included it regardless.
Name: Magnesium Citrate (replaced with magnesium glycinate)
Dosage: 250mg (magnesium glycinate: 1500mg)
Source: $0.05
Information: Magnesium [Examine, webmd, Wikipedia, Gwern] deficiencies are moderately common (up to 45-60%) and easily fixed. Fixing a magnesium deficiency is cheap and seems to offer quite a few minor general benefits, and also sleep and anxiety improvements for some. Depending on your diet, supplementation may be unnecessary. Magnesium comes in a lot of different forms so close attention is needed when purchasing. I stick to citrate as it makes dosing easier, has good bio-availability, and is unlikely to cause digestion issues. The above Gwern link is a great resource on Magnesium as well. This is also another supplement that is easy to get before and after blood tests for to see if your intervention performed as desired.
If you want to have both magnesium and glycine, magnesium glycinate can be purchased which contains both, generally in a ratio of ~14% magnesium to ~86% glycine. This can be a great supplement to take before bed.
Name: Vitamin K2
MK-7
Dosage: 100mcg (sometimes included in BP)
Cost/day: $0.00
Information: Vitamin K [Examine, webmd, Wikipedia], like most vitamins, is primarily beneficial for those deficient in it, so it is best to examine your diet thoroughly and/or be tested. There are several forms of vitamin K, and also several forms of vitamin K2. Vitamin K2 MK-7 seemstobe one of the best forms to take in general, although K1 has decent evidence in favor of it as well, depending on one’s circumstances.
Name: Glucosamine
Sulfate
Dosage: 1-3g (often paired with MSM, back on as of Oct 2024)
Cost/day: $0.19
Information: Glucosamine [Examine, webmd, Wikipedia] is an amino sugar derived from shellfish that is commonly taken by the elderly to improve joint functionality and reduce pain. Glucosamine extends the lifespan of some mammals in studies, potentially in ways that are evolutionarily conserved, activating AMPK and therefore having slight similarity with metformin. Glucosamine may also induce autophagy via an mTOR-independent pathway, which may be the mechanism of action for its effects on lifespan. Due to its popularity as a supplement we can be relatively sure of its safety as well. Chondroiton is commonly included with glucosamine supplements, which appears very uninteresting for my own purposes, so I look for pure d-glucosamine/glucosamine sulfate, which is generally cheap.
Name: Lithium Orotate
Dosage: 5mg (included in BP)
Cost/day: $0.08
Information: Lithium [webmd, Wikipedia] is a metallic element that is often found in foods such as legumes, grains, vegetables, and in some places, drinking water. Lithium is generally present in most diets in notable quantities, and in slightly larger quantities in diets such as the Mediterranean diet. For purposes such as mine, it is supplemented at low doses, which is much different (~1/100th the dose) from the doses sometimes prescribed for some psychiatric disorders. Lithium reduces mortality, stabilizes mood, and promotes longevity, likely via multiple pathways, although the specific mechanisms of action are difficult to discern and more research is needed. As I was tested for lithium and had a very low concentration in my blood, I decided it was worth it to supplement it n low doses.
Name: Glycine
Dosage: ~10g+
Cost/day: $0.20
Information: Glycine [Examine, webmd, Wikipedia] is an amino acid that is often supplemented to improve sleep. Better sleep is formidable by itself, but some studies find that it increases lifespanin organismsvia methods that may be evolutionarily conserved. Although glycine is present in some foods and is also synthesized by your body, it may bethe case that glycine deficiencies are technically common in humans, as the amount that is able to readily be synthesized in-vivo is sub-optimal. This may be relatively asymptomatic from an individual perspective and only manifest itself via a slight probabilistic decrease in healthspan/lifespan, although users often notice quite a few improvements besides just better sleep. Given its heavy involvement in collagen, I wouldn’t be surprised to see skin benefits over the long-term. Glycine may improve insulin sensitivity and other similar metrics. There may be somelongevitybenefits of a diet low in methionine (meat, fish, eggs, etc) as well, which may be related to one’s effective glycine/methionine ratio. I still consume a lot of methionine from common sources such as chicken breast, so this is another potential way in which glycine could be beneficial. Glycine appears to be very safe, even in larger doses, and is relatively cheap, more so as a powder, as is the case of most substances. See also Induction of glutathione biosynthesis by glycine-based treatment mitigates atherosclerosis
I often take glycine in powder form, which makes it easy to consume arbitrary doses (including the ability to add it to drinks or meals if desired), and notably cheaper than buying large amounts of pills, which are generally 1g each at most (you can’t fit more in a pill). On days where I consume a lot of meat such as beef, I take significantly more glycine. This is partially an attempt to optimize my diet’s methionine/glycine ratio, but also intended to do a better job at mimicking what a more traditional consumption of animal meat might have been like, from an evolutionary perspective, which would have included much more glycine than most of us receive in the common cuts of meat that consumers generally use. As a side note, glycine does taste sweet and dissolves in water, so it’s a great addition to tea or coffee.
(Dec 2024 update) I also supplement collagen peptides. Collagen peptides are a more popular supplement which contain 1/3rd glycine along with other amino acids such as proline and hydroxyproline. Glycine seems to generally be the bottleneck for sufficient collagen synthesis, however it seems like collagen peptides in some cases don’t fully break down into the constituent amino acids and may thus offer other signaling benefits, e.g. “telling your body to produce more collagen”. As they’re obviously low risk and likely very good for you, I currently supplement both collagen peptides and glycine on a daily basis.
Regardless, given glycine’s near-flawlessly safe and simple profile, there should be zero harm in having a bit too much. My larger dosage was arrived at from a combination of the papers linked above (and linked to by those links), as well as some reasoning about my diet (high in methionine) and lifestyle. Unfortunately even with a blood plasma test of amino acid concentrations, it’s difficult to know if this is the optimal dose for human longevity, or if it is even helpful at all to begin with, but the cost/benefit analysis here still seems to lean heavily into the green. As a simple and common amino acid, it seems pretty difficult to hurt yourself with glycine, so even taking 10-50g a day shouldn’t be harmful.
Name: Allulose
Dosage: 0-10g+ (varies, uncommonly used as a sweetener with some meals)
Cost/day: $0-0.50
Information: Allulose is an amazing alternative to sugar with 90% fewer calories and the ability to decrease your blood sugar in response to high-carbohydrate meals. I wrote a full post on Allulose here
Name: Bacopa
Dosage: 445mg (discontinued, likely low value)
Cost/day: $0.09
Information: Bacopa [Examine, webmd, Wikipedia] is an herb that seems to offer reliable but likely very minor improvements to some areas of memory and generalcognition. Effects are likely difficult to notice without rigorous placebo-controlled self-testing, but it is relatively safe and cheap regardless. Digestive side-effects aren’t uncommon, as is the case with many herbal supplements. In the future I’d like to replace my bacopa with a placebo and attempt to look for differences in quantitative cognitive performance metrics such as my anki recollection, but performing this experiment well is difficult, both because the effect is very minor and because a proper experiment with n=1 is very difficult. I don’t think bacopa is likely to be a big deal, but I’ve added it for now. As of 2021 I sometimes don’t take this, as it may result in slightly poorer digestion, and the benefit was marginal at best, but I have left it on this list for now.
Name: Ashwagandha
Dosage: 470mg (occasionally discontinued due to potential digestive side-effects)
Cost/day: $0.15
Information: Ashwagandha [Examine, webmd, Wikipedia] is an herb that offers potential anxiety and lipid profile improvements. Some users report that it reduces anxiety and stress significantly, with some studies showing up to a 28% reductions in cortisol (in subjects with elevated levels). Lipid improvements can also be notable, with some studies showing a 10% reduction of total cholesterol, even in healthy subjects. As an uncommon herbal supplement, digestive side effects are a notable probability. Ashwagandha is likely worth trying if you feel that you have untreated anxiety, you never know when you’ll get lucky with how much of a benefit you receive from some things. Although not scientifically rigorous, it did appear like the periods during which I took ashwaganda resulted in a notably improved lipid profile, consistent with what many studies have shown. I’d like to test this on myself in an n=1 RCT both for lipids and for potential relaxation/anxiolytic effects, but haven’t gotten around to it.
Name: Fisetin
Dosage: 0-1500mg (varies, intermittently/rarely taken, but now included in BP at 100mg/day)
Cost/day: $0-3.80 (varies)
Information: Fisetin [Wikipedia] is a plant flavonol that is found in several vegetables and fruits, with the highest concentration being found in strawberries. Fisetin is a sirtuin-activating compound and has extended the lifespan of yeast, worms, flies, and mice. It has been shown to be a strong senolytic agent and may induce apoptosis and other effects via the PI3K/AKT/mTOR pathway. I do not take it every day and am quite uncertain about what the right regime for supplementation should be for it, but currently take ~1,500mg of it for 4 days continually once every few months. This likely has room for improvement and may change in the future. I’d like to write more about fisetin in order to justify this, but haven’t yet found the time. Here’s a single picture of a pretty mouse instead.
Astaxanthin has increased
the life span of C. elegans by 16-30%, with the authors stating
“These results suggest that AX protects the cell organelle
mitochondria and nucleus of the nematode, resulting in a lifespan
extension via an Ins/IGF-1 signaling pathway during normal aging, at
least in part”. While this is certainly interesting, expecting such
a lifespan increase in humans is far too optimistic from this case
alone.
However, Astaxanthin may
be able to activateFOXO3
in humans, an important
gene for human longevity which is present in many centenarians.
Some other well-known natural compounds such as resveratrol
and curcumin
also interface with FoxOs, although these substances are still
relatively speculative as far as anti-aging effects in humans go,
even if they do have many strong supporters.
There’s some other interesting potential effects of astaxanthin,
with some papers showing that it increases
neural stem cell proliferation and may be useful to help curb
dementia, and other papers showing that it can improve
skin health and appearance, leading it to become an ingredient in
some cosmetics.
Astaxanthin appears to be very safe in humans and is a relatively popular dietary supplement, with a market estimated at over $500M USD annually, although the majority of this supply is used as a component in animal feed and cosmetics.
Name: Curcumin
Dosage: 0-500mg (often discontinued, included in BP)
Cost/day: 0-$0.17 (varies)
Curcumin [Examine,
webmd,
Wikipedia] is a
pigment found in tumeric.
Curcumin’s strongest benefit seems to be the reduction
in inflammation that it offers, although there appear to be some
other areas that may be improved as well such as lipid profiles,
mental health, potentially improved digestion, and reduced pain with
some conditions such as osteoarthritis.
It may exhibit a notable anti-tumor
effect via apoptosis. It seems relatively safe, although has low
bio-availability, so is often taken with substances to increases its
availability such as piperine, or taken in an otherwise proprietary
formulation that generally has some type of oil that improves
bio-availability instead. As inflammation is important in aging and
many other diseases, it’s something that is nice to be aware of.
I only sometimes take curcumin depending on my inflammation levels, generally measured via c-reactive protein.. When it is negligibly low, I stop taking it, and if I ever see it creep up in blood test results, I resume supplementation. Curcumin can be potentially tough on the liver, and in large doses has a greater potential to cause adverse affects. Some papersshowquiteafew potential drug interactions that can occur by taking curcumin, especially in larger doses, and via a variety of mechanisms, including its affect on platelets and potential interactions with enzymes such as CYP3A4, potentially affecting the metabolism of a large amount of drugs.
Name: Berberine
Dosage: 1.2g (discontinued, replaced with metformin which was then deprecated)
Cost/day: $0.28
Information: Berberine (Examine, webmd, Wikipedia] is an extract from various plants. It appears to be a pretty strong natural mimetic of metformin, a popular drug for diabetes with many alluring potential anti-aging properties. It often improves lipid profiles and blood glucose, and thus may have many of the long-term benefits that metformin may have. Concordantly, the possibility for digestive side-effects is relatively high, and it’s sometimes taken several times a day in smaller doses as a result. Examine suggestions that it also inhibits enzymes such as CYP2D6 to some extent, which could lead to undesirable interactions with some drugs. It’s likely better to be on metformin than berberine, as drugs are kept to a significantly higher regulatory standard than supplements are and we have much more data on users of metformin.
Name: Caffeine
Dosage: 50-200mg
Cost/day: $0.10
(much higher If drinks are considered)
Information: Caffeine [Examine, webmd, Wikipedia] is something you likely don’t need an introduction to. I try to keep my dosage relatively low to avoid issues with tolerance, using a combination of coffee, tea, or caffeine pills, depending on the amount desired and my mood. When taking 100mg or more of caffeine, I generally have 100mg of L-theanine as well.
Dosage: 0-200mg, (not taken often, 100mg if taken generally)
Cost/day: $0.20
Information: L-theanine [Examine, webmd, Wikipedia] is an amino acid that is present in tea leaves which is often combined with caffeine for supposedly synergistic effectsoncognitionandmood, improving the upsides of caffeine while helping to ameliorate some of the potential downsides. I generally only take it if I’m having more caffeine than average on a given day, since I keep my caffeine intake pretty low.
Name: Melatonin
Dosage: ~300mcg or less (not taken consistently)
Cost/day: $0.04
Information: Melatonin
[Examine,
webmd,
Wikipedia,
Gwern] is a hormone
secreted by the pineal gland with an important role in regulating
your sleep cycle. Melatonin production can be suppressed in many
individuals that are otherwise healthy, for example by exposure
to blue light from computer screens before bed (which solutions
like the program f.lux and blue-light blocking glasses attempt to
solve). The generally accepted benefits of melatonin are a reduction
in the time to fall asleep, although some individuals claim that it
reduces their need for sleep as well (often by 15-60 minutes). For
those with sleep conditions such
as insomnia or jet
lag (or just being older in many cases), melatonin can be a much
greater aid in improving sleep and quality of life.
One meta-analysis
(K=10, N=653), found melatonin supplementation may have helped
significantly reduce some instances of cancer mortality (R = 0.66
after 1yr). Some
studies also find improvements
in gastroprotection, healing
and reducing the rate of stomach ulcers.
Melatonin has
increased the
lifespan of some mice by 18%, primarily given as a supplement
later in life in an attempt to give older mice more effective pineal
gland functionality (directly giving older mice the pineal glands of
younger mice was also performed, which also was very beneficial).
Melatonin levels similarly decline
with age in humans (as most important things do), and
supplementation may be increasingly
beneficial as one ages.
The proper dose of
melatonin to take varies between individuals and many melatonin pills
for sale are dosed too high (5-10mg), so approximate
self-experimentation can be used such as starting with 0.5mg and
increasing your dosage until benefits are noticed. The above link to
Gwern’s website on Melatonin points to a good in-depth analysis
that is worth reading as well.
A few other things you may want to be aware of if you’re interested in melatonin is that you can buy delayed-release and extended-release versions easily otc. Additionally, you can use a beta-blocker like propranolol to inhibit melatonin synthesis to assist in circadian rhythm regulation.
Name: Spermidine
Dosage: ~1-10+mg (various sources, now included in BP, previous source was wheat germ)
Dosing for spermidine is difficult. It’s obviously very safe, but 1mg is likely not enough for the level of effect that we want. The average daily nutritional intake of spermidine varies from 7 to 25mg, and we can see how much one might want to consume for blood levels of spermidine to increase by 39%: perhaps 10mg per day (calculated by multiplying the 66g of natto consumed per day by its approximate spermidine content of 150mg/kg to yield 10mg per day). Although we don’t have plasma concentrations of spermine and spermidine in humans in relation to mortality, this is available in several mice studies. I need to spend more time on this, but I think one might want to supplement as much as 5-20mg of spermadine per day, assuming that it’s not present in their diet in notable quantities already (which is quite possible, as some Mediterranean, Japanese, and other diets contain notable quantities of it).
I currently consume spermidine via wheat germ, which seems to have around 243mg/kg of spermidine in it. If I wanted 10mg a day, this would result in having to consume 41g of wheat germ per day, which although feasible, is a bit tedious, potentially unsavory depending on the method of consumption, and would also result in an additional 164 calories consumed per day. There are some spermidine supplements on amazon, but I am not sure that I trust any of them very much (with the most recent one having the most obvious fake reviews I have ever seen on a supplement), and many of them are simply wheat germ inside of a capsule, which is not only likely to be an insufficient dosage, but also much more expensive. It may be worth mentioning for some readers that spermidine is also present in human sperm, but not in enough quantities to warrant consumption unless you consume copious amounts of it (~0.1mg per ejaculation, assuming 3.5mL and 31ug/Ml).
Information: Metformin [webmd, Wikipedia] is a prescription drug for diabetes and is one of the most popular drugs taken by those interested in longevity, often taken for this purpose by individuals without diabetes. Metformin is said to mimic some of the potential benefits of caloric restriction. It increases the lifespan of mice, increasing AMPK activity and antioxidant protection, resulting in reductions in both oxidative damage accumulation and chronic inflammation. Lifespans of other organisms such as silkworms and nematodes are also increased. There exists a vast literature on Metformin with respect to its mechanisms of improving longevity apart from just this; it’s currently the most popular drug taken to combat aging.
Due to the prevalence of diabetes, Metformin has over 80 million users (the vast majority taking it for diabetes), which gives us wonderful data on its safety, with its side effects rarely including anything besides minor gastrointestinal issues. Metformin is also cheap, costing only $5-$25 a month in the United States. For the above reasons and many others, Metformin appears to be one of the best candidates for an anti-aging drug, leading it to become one of the only drugs making clinical progress in this area with trials such as TAME (Targeting Aging with Metformin). Metformin deserves a larger write-up than I’ve given it here, so you’re encouraged to perform your own research on it (just as you should for anything written about on this page).
For long-term Metformin usage, be sure that you are not hypoglycemic, as well as that your levels of vitamin B6 and B12 are in acceptable ranges, as deficiencies in these are slightly associated with Metformin usage. Metformin may also diminish some health improvements from exercise, and although more research is needed, this factor should be considered for non-diabetics considering Metformin usage. See also: Gwern on metformin
Name: Acarbose
Dosage: varies greatly, infrequently taken at the start of meals with high carbohydrates
Cost/day: $0-$1
Information: Acarbose is a simple diabetic drug which inhibits alpha glucosidase, causing your glucose to spike less than it normally would when ingesting carbohydrates. It is very safe and common, especially in countries such as the United States. Various studies onacarbose in mice have consistently shown it to expend lifespan, sometimes as high as 22% in males, generally much less in females. The probability this applies to humans is, in my opinion, moderately likely, although it is unlikely to be nearly as strong of an effect. Although mice have a lot of similarity with humans (more than many would expect!), their digestive system and diet are more dissimilar than most other categories. With that said, this drug is very safe and provably reduces the glucose spikes in your blood that occur when ingesting large amounts of carbohydrates, which in general seems to be a good thing. It therefore has a lot more potential when taken at the start of eating a large pizza rather than a normal meal (unless pizza is your normal meal, in which case it’s hard to blame you, but you should probably eat other things as well).
Name: 17-α-estradiol / estradiol valerate
Dosage: 0 mg / varies
Cost/day: $0-1
Description: 17-α-estradiol (a non-feminizing form of estrogen) significantly extends male mice lifespan, and this may apply to humans as well. This section turned long so I turned it into its own post. I currently micro-dose estrogen and am experimenting with some other potential solutions here myself; I’d like to write more on this and on related HPG interactions and estrogens/androgens in general. Past links included under this.
I’ve messed around with quite a few hormonal profiles but would like a lot more research in this area, especially on SERMs. If you have quality information in this area please message me!
Some foods that are plentiful in taurine include chicken (378mg/100g), tuna (332mg/100g), and slightly less but still large amounts in crab, shrimp, lamb, beef, eggs, and cheese. An equivalent dosage in humans for the longevity results found in mice may be somewhere in the range of 3-8g. This may seem like a large amount for a supplement, but if you were to eat 1lb of chicken that alone would contain ~1.7g. Acquiring notable amounts of taurine without supplementation can be difficult as a vegetarian and near-impossible as a vegan.
It’s also found in many energy drinks, sometimes in quantities as large as 0.5-1g, and there’s a bit of evidence that it may pair well to reduce the jittery aspect of caffeine in a way similar to l-theanine. It seems like 3g/day can be reasonably consumed safely, and I’d suspect that this carries at least up to 5=10g/day, if not much higher. I’m only moderately bullish (at most) on taurine improving human lifespan, especially if one begins with the assumption of a good diet, but I find it really interesting nonetheless.
Name: Semaglutide (now discontinued)
Dosage: 0.15-0.3mg (subcutaneous injection, once per week. may use in the future again)
Cost/day: $1-2 ($30-50/month, but reduced food intake can actually make this profitable to take!)
Description: Semaglutide [Wikipedia] is one of the most important inventions of the century. It is the weight loss drug known as Ozempic, Wegovy, or Rybelsus. There is so much literature on this drug (and it now has millions of prescriptions in the US) that I’m not going to go over it here, but the short version is that it not only works, but it works well, and seems to often have few or no negative side effects, and sometimes even positive side effects, from reduced cardiac events to many reports of improved willpower.
Obesity leads to millions of deaths, and now there is a cure for it which can actually be applied at-scale, even in the presence of relatively unrestrained markets with highly-addictive and caloric food. Semaglutide should receive a Nobel Prize.
Although it is known for being expensive when bought through FDA-approved mediums, it is possible to purchase it safely for significantly less as it is not a controlled substance. As I’m neither a doctor nor a lawyer, you should speak with your doctor if you’re interested in trying it out. If you have injectable (liquid) semaglutide, it should be refrigerated.
2024 update: I’ve discontinued semaglutide for two reasons, the first being that I no longer desire to lose more fat, and the second being that it reduces my hedonic baseline by a bit (that is, all of life is slightly less pleasurable). I may hop on it or try microdosing it in the future.
Name: Rapamycin
Dosage: 4-12mg (schedule and doses vary, taken at most once a week, many other factors)
Cody/day: $1-4
Information: Rapamycin is perhaps the most exciting substance for me in longevity right now. Rapamycin notably extends the lifespan of most organisms we have given it to thus far, but lacks proper research in humans aside from its use as an immunosuppressant. It’s a very popular drug to research in the area of longevity, and deserves a longer write-up than I’ve given it here. It’s also potentially quite dangerous and we have little data in humans (aside from those we give it to for organ transplants), so please don’t take it yourself (Jan 2021 update: mTOR Inhibitors Associated with Higher Cardiovascular Adverse Events ‐ A Large Population Database Analysis). Dosage for rapamycin is a bit tricky. There’s a lot of speculation involved, but it seems like many people converge onto something either like 6-8mg/week, or 8-15mg/2 weeks, perhaps with some breaks between.
2024 Update: Rapamycin blood tests are reasonably available – I currently shoot for around 25-45 ng/mL ~2 hours after initial ingestion as a reasonable baseline.
Out of all of the longevity agents I am interested in and/or take myself, it is likely to be the one that I have the highest hope for in humans. We have a decent understanding of the mechanism (compared to many other things, at least), it works very consistently and strongly in several other organisms, and the mechanism of action is strongly evolutionarily conserved. As for safety concerns, it seems like if taken in a low dosage and infrequently enough, the safety profile improves significantly and it may be a net-plus in many areas (this may be related to mtorc1 vs mtorc2 activation depending on the dosage and timing (it does have a pretty long half-life!), which also makes it seem like it can be taken without actually suppressing one’s immune system or causing some other undesirable effect categories).
Although I do know of many others that take rapamycin, I still don’t suggest it to anyone myself, firstly because I don’t offer medical advice of that nature regardless of my cost/benefit analysis (are there risks of potentially bad unknown side-effects with long-term usage? sure, but the risk of *not* taking longevity agents is also pretty large, and results in a much earlier likely death), and secondly because it is still likely to be higher risk than a lot of other simple things that I do often suggest to others, like glycine supplementation, which I see as close to zero risk. I’d hope that anyone that takes it themselves has blood panels done (if not much more) to ensure they’re not doing easily-observable harm to themselves as well. I’d like this section to be more comprehensive, but I’ll follow with some relevant pubmed papers for now:
Selegiline/L-deprenyl
is a MAO-B (and sometimes MAO-A) inhibitor sometimes used to help
treat Parkinson’s or depression which may be able to improve
lifespan in humans.
Dosage: 10mg/day for 5-10 days in a row, once or twice per year
Cost/day: $40
Epitalon is really interesting. A few concerns in humans and I don’t buy that the telomerase activation is the reason why it may have potential (if it does), but there’s a chance it passes EV calculations depending on how you do the math. I’ve tried it a few times and I’m uncertain if I’ll continue or not. More info here to be posted later.
Additional supplements I do not currently take
The above section concludes the actively-maintained list.
Under this I include a lot of links/notes from previous research, which is not comprehensive nor updated.
This section contains a list of supplements that I think might be worth taking, but that I currently don’t use. Substances in this section seem to be relatively safe, and I’m generally only not taking them because I have more doubts about their usefulness to me specifically.
Aspirin
Aspirin [Wikipedia]
is used for more than just treating temporary pain or fevers. As an
NSAID,
it reduces both acute and long-term inflammation, and may
also affect oxidant production, cytokine responses, and block
glycooxidation reactions. Consuming a low dose of aspirin daily
appears to lower the risk of CVD in higher-risk groups (generally
older
individuals with a relevant medical history), although appears to
have little
effect in otherwise
healthy individuals. The
risksof a
few cancers may be lowered
slightly by long-term
continualuse
of aspirin, although this is generally a minor effect, and
doesn’t seem to be the case for all types of cancer. Someorganizationssuggest
daily aspirin use in small doses for those in certain risk groups,
generally those that have already experienced a heart attack or
stroke.
Among aspirin’s more common adverse effects is an increased risk
of gastrointestinal bleeding, which is one of the reasons it’s not
suggested by most organizations for otherwise healthy individuals
with low CVD risk. Aspirin has increased
the average lifespan (although not the maximum lifespan) of mice
in some studies, but this is unlikely
to be the case in humans unless significantly more needs to be
taken, which would increase the probability of adverse effects
notably.
To summarize, it’s very likely that continual aspirin usage
reduces the risk of some types of cancer and moderately likely that
it can reduce the risk of CVD in some higher-risk groups. Although
side-effects are negligible for most individuals, it is difficult to
tell if aspirin is worth taking for healthy and young individuals.
It’s likely much more beneficial for the elderly or middle-aged, as
they’re at a much higher risk of cancers as well as CVD. As a
result of this, I don’t take aspirin regularly.
Cocoa Extract
Cocoa [Examine,
webmd,
Wikipedia]
is well-known as a major component of chocolate. Although the sugar
added to most modern chocolate definitely does not benefit one’s
health, cocoa itself has many bioactive substances with potential
benefits. Among the most notable is (-)-epicatechin, which can offer
improvements
in blood flow and a corresponding reduction
in blood pressure for many individuals. As usual, the most
notable improvements in blood pressure and cholesterol occurred in
individuals with pre-existing elevated levels. Some age-related
markers improve in mice when supplemented with (-)-epicatechin,
although no direct increase in lifespan has yet been noted.
Supplementation with some form of cocoa (supplemented or consumed as ultra-dark chocolate) may be beneficial for some individuals, although consuming too much sugar with cocoa would likely offset any positive effects. Quality cocoa extract is more expensive than many of the other supplements listed on this page, coming in at $1-2 day for a proper dosage.
Also, I’d love to purchase this and test it on myself for awhile to see if the effects can easily be measured.
CoQ10
CoQ10 [Examine,
webmd,
Wikipedia]
(Coenzyme Q10 / ubiquinone) is a substance found in meat and fish
that is primarily present in mitochondria and aids ATP production.
Although supplementation is likely safe, it’s difficult to find
convincing evidence that CoQ10 supplementation would be effective for
longevity. It may improve
lipid peroxidation, blood
flow and offer minor improvements in other areas, but in my
opinion doesn’t appear to stand out from most supplements, both
experimentally and theoretically.
Creatine
Creatine [Examine,
webmd,
Wikipedia] is an
organic compound used in the recycling of ATP
in humans. It can be found in notable amounts of muscle meat and can
also be synthesized
in humans via glycine, arginine, and methionine. Creatine is a very
popular supplement for athletes with strong evidence that it notably
increases
power output and lean
mass, with some evidence that it can offer minor improvements in
related areas such as recovery, fatigue, and some biomarkers that are
positively associated with quality anaerobic exercise. It’s very
safe, has little potential for any side effects, and is relatively
cheap. The only reason I don’t take creatine right now is that I’m
not doing many activities to build muscle, although I’ll likely
start taking it soon, even if only alongside basic resistance
training, calisthenics, or even cardio.
Quercetin
Quercetin [Examine,
webmd,
Wikipedia] is a
flavanoid found in fruits and vegetables. As usual, eating the right
fruits and vegetables is good for you on its own, and may make
supplementation less beneficial, or completely irrelevant. I likely
get enough of this from my diet, although there may be benefits to
infrequent high-dose supplementation.
L-carnitine
L-carnitine [Examine,
webmd,
Wikipedia] is
an ammonium compound found in notable quantities in meat such as
beef. Supplementation sometimes appears to offer some decent results,
but I’ve determined that I like already get a sufficient amount
from my diet.
PQQ
PQQ [Examine,
Wikipedia]
is a redox cofactor found in human breast milk and some foods such as
kiwis. PQQ alters
indicators of inflammation and mitochondrial-related metabolism.
It’s likely very safe, the main reason I’m not currently taking
it is there’s very little evidence showing that it notably benefits
already-healthy young humans, and it costs a bit higher than most
supplements on this page.
Resveratrol
Resveratrol [Examine, webmd, Wikipedia] cannot go without being mentioned, as the extract from grapes that inspired the ‘red wine is great for you’ craze many years ago, it has been a constant source of speculative benefits and is still a very popular supplement in longevity communities. Although it hasn’t quite lived up to its initial hype, there’s still a lot of research on how it may be beneficial for longevity in one way or another. I’m personally not very into resveratrol and don’t see it as that interesting by itself. A summary is currently excluded here and you’re encouraged to read the above links if interested, but to be rather blunt, I think resveratrol is very likely approximately worthless, and is just yet another case study in now media hype in no way correlates with actual efficacy.
Sulforaphane
Sulforaphane [Examine, webmd, Wikipedia] is a compound found in vegetables such as broccoli and cabbages, with the best sources of it being broccoli sprouts and cauliflower sprouts. I’ve taken sulforaphane previously, but it will be difficult to know if it had a notable effect on me or not. I’m currently focusing more on my diet and have decided against taking sulforaphane. I’ve excluded a research summary in favor of the above links.
Some other supplements I am currently considering
Nicotinamide Mononucleotide (NMN) (Wikipedia): todo
Trimethylglycine [Exmine, Wikipedia] is a betaine amino-acid derivative found in some plants. It is notable for reliably reducing homocysteine levels in healthy subjects, sometimes by as much as 10%, and as much as 10-40% in unhealthy individuals. It appears that it might have a slightly negative effect in increasing, or preventing to some extent a decrease in, LDL, which is why I’m currently not taking it. It’s a nice molecule to be aware of and might deserve a spot in my stack at a later point, but as usual it would be nice to have more research available.
A lot of supplements have been excluded from this list, including many which are very interesting. Individuals who follow nootropic or longevity communities will definitely be curious why their favorite substance may have been excluded from this page, to which my answer is mostly that there’s too many substances for me to include all of them, so I did quite a bit of picking personal favorites. Even so, there’s likely many substances I’d like to include, but which I haven’t yet heard about or done enough research on. Feel free to message me on Twitter if you have any great suggestions here.
More interesting and potentially unsafe
substances
This section contains some brief notes and links on substances that appear to be a lot more ‘experimental’ than the above sections, but have some interesting potential. In some cases it’s impossible to find proper tests of safety, or even basic toxicity, in humans. Regardless, they’re all interesting chemicals, sometimes increasing the lifespan of organisms such as mice by large amounts. A lot of compounds have been excluded from this list as there are too many for me to list currently. The most interesting item of this list is currently rapamycin, by a large margin. Also see list of potential CRMs.
Allantoin
Allantoin
is a compound present in some cosmetics, toothpaste, shampoo,
lotions, and more, which has improved lifespan in C. elegans in
multiple studies.
Rejuvant®, a potential life-extending compound formulation with alpha-ketoglutarate and vitamins, conferred an average 8 year reduction in biological aging, after an average of 7 months of use, in the TruAge DNA methylation test: https://www.aging-us.com/article/203736/text
Astragalus Membranaceus
astragalus
membranaceus contains a compound called TA-65 that may
activate telomerase, extending the lengths of the shortest
telomeres in humans. This compound is lacking in notable research,
and much of what exists is clearly for-profit.
C60 is an interesting fullerene that has extended lifespan in some animals notable, but has little data on human consumption and safety. In an original study that gained quite a bit of attention, it was reported to ‘almost double’ the lifespan of rats. Now in 2021 the two most recent studies I see show that it did not extend lifespan, and that it only extended it by around 7%. There is less information on mechanism of action than we would want, but it is generally suggested to be related to free radicals.
There’s apparently quite a few people that have been taking this themselves, buying it from less-than-reputable Internet sources and hopefully not letting it be contaminated with light, as when exposed to light it degrades and becomes very dangerous to consume. This is certainly not something I plan to touch myself in with the current state of our knowledge on it, but it does seem like a very interesting chemical nonetheless.
Benzofuranylpropylaminopentane
Benzofuranylpropylaminopentane
is an unusual and understudied drug, in some ways similar to
selegiline noted above. It has prolonged lifespan to a minor extent,
such as 4% in mice.
GDF11 is also really interesting. Similar to Epitalon it’s obvious that the wikipedia article was authored by someone into longevity, but this is another thing I’m going to keep an eye on.
If you enjoyed this post or have corrections feel free to say hi on Twitter or any method on my about page
Google’s ReCAPTCHA is often the first tool that many webmasters reach for when confronted with the need to stop spam and automated malicious traffic from harming their services. In this post I explain several reasons why ReCAPTCHA is usually not the best solution to use for this purpose, as it is often unnecessary, inconveniences users, and subjects users to intensive tracking and fingerprinting that they are not able to opt-out of. Several alternative solutions to ReCAPTCHA for various threat models are presented as well as best practices for implementing captchas in general.
The face of evil
ReCAPTCHA is harmful
ReCAPTCHA is yet another free-of-charge product offered benevolently by Google for any webmaster to implement within their own services. How does ReCAPTCHA differentiate legitimate human users from bots? ReCAPTCHA relies extensively on user fingerprinting, putting emphasis on the question of “Which human is this user?” rather than the ordinary “Is this user human?”. It’s worth noting how much easier it is to successfully solve ReCAPTCHAs when the user is logged into their Google account, thus allowing Google to associate their actions with their real identity. A similar effect is often reported for users of non-Google browsers, who notice ReCAPTCHAs take more time to complete in Firefox over Chrome. This is in-line with many other anti-competitive techniques that Google has used over the years to help grow their market share.
Although determining
exactly how ReCAPTCHA works is very difficult, with Google not only
heavily obfuscating its JavaScript, but also implementing an entire
VM in JavaScript with its own bytecode language, there have still
been many attempts to reverse-engineer some of the client-side code,
as well as to theorize about how the server-side logic operates.
Initial attempts
at reverse-engineering ReCAPTCHA show copious amounts of information
belong collected, including but not limited to: plugins, user agent,
IP address, screen resolution, execution times, timezone, language,
click/keyboard/touch information within the frame of the captcha,
test results of many browser-specific functions and CSS evaluation,
information about canvas element rendering, and cookies, including
those affiliated with your Google account that were placed within the
last 6 months.
There is a good reason why ReCAPTCHA uses the google.com domain instead of one specific to ReCAPTCHA. This allows Google to receive any cookies that they have already set for you, effectively bypassing restrictions on setting third party cookies and allowing traffic correlation with all of Google’s other services, which most users use. ReCAPTCHA collects enough information that it could reliably de-anonymize many users that simply wish to prove that they are Not A Robot. As JavaScript is now required to even view a ReCAPTCHA, even a user running software such as TBB (Tor Browser Bundle) may find themselves giving away more information than they intend to, for example if they have resized their browser window (which is discouraged for exactly this reason).
Some unlucky ReCAPTCHA questions seem down right impossible for some users
Correspondingly, webmasters that use Google’s ReCAPTCHA on their websites must link to both Google’s Privacy and Terms pages (included in the form by default in a small 8px style that makes them appear unclickable). Although Google used to have its own privacy and terms pages for ReCAPTCHA, these links are no longer specific to ReCAPTCHA, but rather are the privacy and terms pages for all users of Google services in general, regardless of which service is being used, or if the user has (or even wants) a Google account to begin with. Therefore accepting these terms (implicitly, by attempting to prove you are Not A Robot) grants Google permission to do everything that they do to their regular users of their services to you, and little information is available as to what specifically is done (GDPR is likely to be unhelpful here, given ReCAPTCHA’s spam-stopping purpose). Not only are the unhelpful links in the ReCAPTCHA box never opened by users, but there is also no Google logo or visual reference to indicate that ReCAPTCHA is a Google service, so many users have zero indication that they have just consented to all of Google’s tracking just because they tried to leave feedback or create a ticket on your website. If you thought you could use the Internet without using Google’s services, try using the Internet without filling out a single ReCAPTCHA, which for some users is required to pay their bills, file their taxes, and sometimes even use Government websites (if you somehow manage this, next try never sending email to Gmail/Gsuite addresses or using Google APIs for a more exciting challenge). Good luck.
It is worth mentioning that caring about user privacy to this extent is likely to be outside of the scope of concern for most websites. Many websites are already so tightly coupled to Google’s services (commonly including Google analytics, Google ads, Google APIs, Google tag manager, Google static resources, Google OAuth, Google Compute Engine, and many others) that the addition of a Google captcha appears negligible. With that said, different websites have different values and different users, and many do not want to require users to agree to Google’s tracking and labor for basic usage. The level of centralization that ReCAPTCHA forces is not good for anyone except Google.
Apart from the privacy implications of ReCAPTCHA usage, the actual captcha is very tedious for many classes of users, sometimes becoming so difficult that users find themselves unable to to complete the captcha at all. Users hate ReCAPTCHA. TheyreallyhateReCAPTCHA. ReCAPTCHAissohated that some websites have a profit model of charging users $20 annually to disable ReCAPTCHA, which thousands of users pay for. If this sounds like a great new business model to you and now you want to add ReCAPTCHAs to every page of your website to attempt to maximize profit, I will find you. And I will force you to complete a ReCAPTCHA every time you want food or water until you die from malnutrition after the first week. I have read countless posts from users that became so frustrated with a service that used excessive ReCAPTCHAs that they swore to never use the offending website again. These are often intelligent users with no disabilities who are simply tired of being treated like dirt and wasting their time. Be kind to your users and help minimize the amount of ReCAPTCHAS that they have to solve just to be allowed to use the Internet.
ReCAPTCHAs become significantly more difficult if the user attempts to ‘opt-out’ of Google’s services and tracking by using software that hinders it, such as VPNs, TBB, and many anti-tracking browser addons and modifications. To demonstrate what is meant by ‘very tedious’, below is a real-time recording of myself solving a single ReCAPTCHA using TBB:
Spambots are known to give up when forced to be patient
I got lucky and only needed to complete two challenges. Sometimes there are ten or more. Watching the above video, you might think to yourself “I knew the tor network was slow, but I didn’t know it was that slow!”. You would be correct to take note of this discrepancy. If we open up the web developer console, we can see that the HTTP requests for new captcha images only take a few hundred milliseconds. Despite this, Google’s heavily-obfuscated JavaScript intentionally delays the appearance of the new images by several seconds every time, which I’m sure has something to do with the fact that bots give up when forced to wait, probably. This is not a nice way to treat users that don’t want to perform unpaid labor and be fingerprinted by Google. Keep in mind that the above video demonstrates one of the worst possible cases of ReCAPTCHA UX (which some userscripts may improve), and that the average user has a significantly quicker experience, providing that they are not attempting to thwart any of Google’s tracking and don’t make many mistakes.
In addition to this tediousness, the actual labor that the user is performing is directly used to benefit Google. Worry not however, as Google is eager to brag about the selfless humanitarianism that you’re engaging in by choosing ReCAPTCHA, stating the following on their main ReCAPTCHA page:
“Hundreds of millions of captchas are solved by people every day. ReCAPTCHA makes positive use of this human effort by channeling the time spent solving captchas into digitizing text, annotating images, building machine learning datasets.”
This
is certainly a very rosy way of convincing
you to feel good about forcing
your users to engage in unpaid labor that directly benefits the
world’s most powerful surveillance corporation. ReCAPTCHA
is free for a reason.
Lastly, ReCAPTCHA is popular. Very popular. While this brings some advantages, it also means that there’s significant efforts to break ReCAPTCHA, and those efforts all potentially affect your website, with your captcha implementation being perfectly identical to a million others. As a result of this, there have been manypaperspublishedthatbreakReCAPTCHA over the years, generally with Google making modifications to improve their captcha afterwards. There have also been paid-forservices that use human labor to solve captchas on behalf of a paying client for less than a cent each. For a modern and user-friendly example of bypassing ReCAPTCHA, see Buster. Buster is a modern browser extension (Firefox+Chrome+Opera) which helps you to solve difficult captchas by completing reCAPTCHA audio challenges for you by using speech recognition.
Captchas are not always necessary
Before implementing a captcha, it’s worth considering if one is
necessary to begin with. To help with evaluating this proposition,
consider if your threat model is concerned over customized or
uncustomized spam. Uncustomized spam is pervasive across many
Internet protocols, and you will encounter it quickly after enabling
HTTP, SSH, or many other protocols on a server. It is generally
unintelligent, cheap to execute, and easy to block, even without
captchas. Customized spam, however, is spam that has been written to
specifically affect a given company, service, website, or user. As
customized spam is created by an actor that is able to tailor it to
your service, it is more dangerous than uncustomized spam, and more
effort is required to effectively limit it.
Many developers vastly over-estimate the likelihood
of customized spam. As a
competent programmer, it is easy to imagine how effortlessly
someone could decimate your service with spam if they were
sufficiently dedicated. One
could imagine a malicious actor writing a simple script that could
spam or DoS your website by just using Curl and bash. Even if you
have a captcha, you can imagine them using OCR or machine learning to
automatically bypass it, then
using proxies and VPNs to automatically bypass
your IP rate-limiting. While
in this imaginative trance, you’ve forgotten that 99% of users have
no clue how to do any of this, and
do not even know what Curl or HTTP are.
Furthermore, your service
likely offers
very little prospective rewards to would-be competent attackers.
Just because someone could
spend hours (or
minutes) writing
a program to spam your website does not mean
that
someone
will.
Your
personal blog about the latest vegan bacon is not a high-priority
target for anyone. Adding a
ReCAPTCHA to your Contact Me page is just
a great way to get
no one to talk to you. I’ve
ran several websites with millions of pageviews that have received
zero customized abuse and have spoken to other webmasters with
similar experiences. Jeff
Atwood of codinghorror.com
once wrote
similarly:
The comment form of my blog is protected by what I refer to as “naive captcha”, where the captcha term is the same every single time. This has to be the most ineffective captcha of all time, and yet it stops 99.9% of comment spam.
This is not a suggestion to do nothing, ignore basic security, and be unprepared for attacks, but rather to realistically consider your threat model and apply only what is necessary.
ReCAPTCHA
alternatives for uncustomized spam
For uncustomized spam, a full
captcha implementation is
rarely necessary. This section lists some simple and
effective tricks that stop most uncustomized spam from impacting
your website.
Hidden form elements
Uncustomized spam is not intelligent enough to know when it should
or should not fill out a form element. For example, adding a form
element with a name of ‘url’ and hiding it with CSS allows you to
reject any request that is made with it filled, which spambots are
eager to do. To maintain accessibility be sure to add a label to this
element so that users who use screen readers do not fill it out.
Other good hidden form element names include ‘website’,
‘firstname’, ‘lastname’, ‘email’, and ‘name’, unless
they are already being used legitimately.
Static questions
Uncustomized spambots are also so unintelligent that they do not correctly answer simple questions such as “What is 2+3?”, or “what is the name of this website?”. These questions effectively stop almost all uncustomized spam. Common software stacks such as WordPress and Drupal have freeplugins that will allow you to quickly create questions like these.
Community-specific questions
If your website is community-centric such as a forum or blog, you
can ask a community-specific question that prospective members of
your community should know the answer to. This is a simple and great
way to prevent users from joining your community that you believe
shouldn’t be participating, either because they lack basic relevant
knowledge, or because they are unable or unwilling to learn it. As an
example, a community of mathematicians might ask the user to name a
simple formula or solve an equation, given an image of it.
Effective at keeping out the arithmophobic
For
another example, a community of niche media connoisseurs might ask
the user to identify a certain character that they deem to be
important to their shared culture.
The quality of our community members is of the utmost importance
JavaScript
Did I mention uncustomized spambots are unintelligent? Basic JavaScript is not executed or parsed by most uncustomized spambots, so using it to calculate the value of a form element is also effective. JavaScript can also be used to submit the form itself, set a CSRF token properly, or perform many other simple tasks. If your site has many users with JavaScript disabled, it is better to offer an alternative solution as well.
Third Party Services
From WordPress plugins like Akismet, spam-detection APIs like StopForumSpam, and APIs that evaluate users or IPs such as abuseIPDB, there are a lot of free (and paid) third party services to aid you in stopping spam in ways that are not visible to most of your users.
ReCAPTCHA
alternatives for customized spam
If you operate at sufficient scale and/or if automated usage of your website is inherently lucrative enough, customized abuse will eventually happen for one reason or another. Remember that a captcha is just a tool to help verify that a given user is a human. It is not the only tool to help with this, and it is not the right tool for every use case. No solution is perfect and can stop a sufficiently-resourced attacker from abusing your service. This section lists some alternatives to ReCAPTCHA in roughly increasing order of complexity.
Popular CMS
solutions generally have at least one simple captcha plugin that is
suitable for basic purposes. Here are some examples for WordPress,
Drupal,
and generic PHP.
Secureimage PHP captcha
Drupal match+slide captcha
Custom JavaScript functionality
Just as basic
JavaScript stops
most uncustomized spam, more
advanced scripting
can stop a lot of customized spam as well. For example, some websites
require you to slide a jQuery
slider element in order to successfully submit
a form. There are examples of
this for wordpress,
jQuery
(jQuery
UI slider, Bootstrap
slider),
Prestashop,
Node,
and more, although
these examples may not be suitable for production use
and I have not tested them.
Slide to unlock
Just
including
true JavaScript evaluation as
a requirement will raise the
bar for attackers, and can be done without the user having to perform
any actions. If you choose to
write a lot of custom front-end code to evaluate users, be sure to do
extensive user testing on every type of device and log failures so
that they can be analyzed to further remove false positives.
Capy Puzzle CAPTCHA
Capy offers a simple puzzle captcha that requires the user to drag a puzzle piece into an empty slot.
All of the fun of finishing a jigsaw puzzle with none of the effort
SolveMedia
SolveMedia
offers a captcha and corresponding plugins for a variety of popular
software stacks, including vBulletin, WordPress, MediaWiki, Dupal,
Joomla, and more. The captcha can scale its difficulty based on the
threat score of a user.
he’ll come when you least expect it
If for some reason you feel the need to profit off of your captcha implementation, fear not, as there’s also a version fit for the capitalistdystopia of the near future:
Please drink verification can
Geetest
Geetest appears to use some fingerprinting, but otherwise works similarly to most puzzle captchas. Notable for being used on Binance, one of the world’s largest cryptocurrency exchanges.
hCaptcha
Lastly, as of 2020, even Cloudflare has switched away from ReCAPTCHA, instead using hCaptcha, an alternative that seems just as difficult to bypass as ReCAPTCHA, but that respects user privacy, and potentially even pays some clients for their users’ labor in data labeling.
hCaptcha’s website, showing example captcha challenges
This list is nowhere near exhaustive and many similar captchas have been excluded from it. If you are a software engineer, you likely think many of these captchas could be solved by software, which although correct, misses the point. Although in theory a captcha should be a perfect turing test, in practice, they are only used to make attacking your service more difficult so that spam is no longer cost-effective. Even a perfect captcha provides no guarantee of stopping all abuse. Nonetheless, you may be surprised at how few attackers are willing to execute JavaScript or perform OCR to automatically attack your website unless you run an extremely popular service.
Captcha best practices
If you have decided that you do
need a captcha, consider if
it’s truly necessary to implement it in all of the locations where
you want to throttle automation. Showing
users fewer captchas not only
provides
a better UX, but also improves KPIs
like conversions and user
retention.
Use
rate-limiting where possible
As the purpose of a captcha is to confirm that the end-user is a
human, a user should generally only have to correctly solve a captcha
once. If there is an action that you would like to throttle to ensure
it is not performed too often by a user, consider using rate-limiting
as an alternative (or in combination with) a captcha.
Use reasonable thresholds for captcha
presentation
Set reasonable thresholds for actions that you want to limit with captchas. Rather than presenting a user with a captcha after a single failed login attempt, allow several attempts. Brute-forcing secure passwords in this manner is not feasible to begin with, and if credentials from a database leak are being automatically cross-validated with your service, a post-login-failure captcha won’t even help.
Stop showing captchas to users that are just trying to read content. If your blog asks me to complete a captcha just to read a single post because I’m using a super-scary VPN as a result of your CDN’s “premium military-grade bot protection” feature, I’m going to close the tab. There are sometimes cases where captchas are more reasonable for read-only actions such as stopping active application-level DDoS attacks. Your blog is not one of these cases.
Do not require repeated captcha solves
If a captcha is part of a form that may fail validation and is reloaded upon failure, do not force the user to solve another captcha if they correctly solved the first one. This prevents users having to frustratingly solve captchas several times in a row as they fix their input (for example, adhering to your revolutionary password policy that requires at least 1 non-printable character, 1 Egyptian hieroglyph, and 1 iOS-only emoji).
Intelligently use other sources of validation
Consider if you have reasonably validated that a user is likely to be a human during previous interactions with them. If a user has a confirmed email address and phone number or proper two-factor factor authentication, it may be unnecessary to show them a captcha. Similarly, if a user has been a paying customer for several months without issues and is attempting to make a new purchase with their existing billing information, it is also a bad time to make them fill out a captcha. I mention this only because I’ve had to do it before.
The future of verification
It’s important to note that a sufficiently-resourced attacker
can bypass any mechanisms you have in place to
some extent. When a service has a billion users (Facebook and
Twitter) or otherwise provides significant incentives for abuse
(anything related to cryptocurrency), difficult trade-offs must be
made when attempting to verify users.
Faced with this, some services that operate at very large scales not only use ReCAPTCHA, but also perform phone and/or email verification and employ a significant amount of custom automation-detecting heuristics. Twitter is a good example of this, as new users are required to both complete a ReCAPTCHA and (usually) verify a phone number. On top of this, Twitter has entire teams dedicating to stopping abuse, and yet the platform still has issues with millions of spambots, just as Facebook does. Although requiring phone verification has unfortunate consequences for anonymity, most platforms were not intended to be used anonymously to begin with. An even greater challenge is attempting to stop spam in environments where user anonymity is desired, which I provide some examples of at the end of this section.
With the current state of machine learning, it is becoming
increasingly difficult to construct a captcha that is user-friendly.
Some of the most effective
attacks on advanced captchas such as ReCAPTCHA have simply involved
taking a given challenge and querying a machine learning API to solve
it automatically. Now that we have manyAPIservices
to accurately label audio, images, videos, and more, this is only
becoming more powerful, just as machine learning is in general.
Despite the impossibility of a perfect captcha, articles have been written decrying that captchas are dead for more than a decade due to the increasing possibility of true negatives (software that passes as a human). Despite this, most of the Internet is not covered in spam. Intelligent software engineers make much more money working at FAANG instead of covering the Internet in unsolicited fake Viagra ads, at least for now. For a potentially poor analogy to physical security, remember that we have physical items that can break doors, windows, cameras, sensors, locks, and much more. Yet, these protections are all still essential features of a physical security system. They are often not made to be impossible to break, but rather to make an attacker’s job significantly more difficult, skewing the effort/reward ratio enough to stop most attackers.
Regardless of the forthcoming AI supremacy, the current paths that larger systems tend to favor involve validating who a specific user is rather than only attempting to validate if they are human or not. Phone verification and sometimes even picture, ID, or address verification are found among large services that have a high potential for abuse, as well as our good friend ReCAPTCHA. Verifying users while attempting to better preserve anonymity is more difficult, but those that are determined generally find clever ways to do so. Some good examples include privacy pass (protocol paper), allowing users to anonymously skip captchas if they have already solved one, Apple’s new Find My Device feature, allowing Apple devices to broadcast their location with BLE such that it can only be read by the original device’s owner, and well-known security systems such as asymmetric cryptography, cryptographic hashes, differential privacy, etc, which can often be cleverly implemented in systems to improve security and often anonymity. Some other techniques that can be used to help verify users and reduce spam include proof-of-work and micropayments, both of which have been used successfully in most popular cryptocurrencies such as Bitcoin and Ethereum for more than a decade, although can still be difficult to implement in everyday scenarios.
If you are Twitter or Facebook, no captcha will solve all of your issues. For everyone else, there are still a lot of simple tools and heuristics that go a long way in helping to stop abuse. Be kind to your users and try your best to not force them to spend their free time completing ReCAPTCHAs for Google. They will appreciate it.
If you enjoyed this post or have corrections feel free to say hi on Twitter
This is a new blog where I hope to semi-regularly (at least once a month) make posts about interesting topics. Please see the About page for more information.
