In this post I will go in-depth on:
- How Tether got to be what it is today
- Why Tether's market cap is a lot more than 0.5% of the total market cap for crypto you see on CoinMarketCap
- Tether printing timing
- Tether reserves
- What could happen to the market if Tether is found to not be backed by reserves
Tether is incredibly important to the cryptocurrency market ecosystem and I've noticed far too few people understand what is going on.
Very little actual discussion of the 2nd biggest crypto by volume happens here and whenever someone starts a discussion they most often got slapped for "FUD". Tether themselves recently hired the major New York based PR firm 5W to spread positive information online and take down critics, I'm sure some of their operatives are probably on Reddit.
But its absolutely critical you understand the risks behind Tether and especially now with the explosion in reserve liability, breakdown in relationship with banks and their auditor and recently announced subpoena.
What exactly is Tether and what happened so far?
Tether is a cryptocurrency asset issued by Tether Limited (incorporated in the British Virgin Islands and a sister company of Bitfinex), on top of the Bitcoin blockchain through the Omni Protocol Layer. It is meant to give people a "stablecoin", for example a merchant who accepts bitcoin but fears its volatility could shift bitcoin into tether, which can be easier to do than exchanging bitcoin for dollars. Recently they've also added an Ethereum-based ERC20 token. Tether Ltd claims that each one of the tokens issued is backed by actual US dollar (and more recently Euro) reserves. The idea is that when a business partner deposits US dollars in Tether’s bank account, Tether creates a matching amount of tokens and transfers them to that partner, it is NOT a fractional reserve system.
Tether makes the two following key promises in its whitepaper on which the entire premise is build:
Each tether issued will be backed by the equivalent amount of currency unit (one USDTether equals one dollar).
Professional auditors will regularly verify, sign, and publish our underlying bank balance and financial transfer statement.
Tether is centralized and dependent on your trust of Bitfinex/Tether Limited, and that the people behind it are honest people. For the new entrants to this market it will be greatly beneficial understand the timeline of Tether and their connection to Bitfinex.
A brief timeline:
- Bitfinex operators Phil Potter and CFO Giancarlo Devasini set up Tether Limited in the British Virgin Islands, but told the public that Bitfinex and Tether are completely separate. Throughout 2015 and 2016, the amount of Tether stays relatively flat.
- In August 2nd, 2016, the second-largest digital currency exchange heist in history happened, when Bitfinex lost nearly 120,000 bitcoin. Bitfinex never revealed full details of the hack, but BitGo (the security company that had to sign off on the transactions) claims its servers were not breached.
- Just 4 days after the hack Bitfinex “socializes” its losses from the theft by announcing a 36 percent haircut for almost all of its customers. In return, customers receive BFX tokens, initially valued at $1 each.
- Two weeks after the hack Bitfinex announces it has hired Ledger Labs, to investigate the theft and perform a financial audit of its cryptocurrency and fiat assets. The public nevers sees the results of the investigation, and months later, Bitfinex admits it never actually hired Ledger Labs to perform an audit to begin with.
- In May 2017, after long standing calls for an actual audit, Bitfinex hires Friedman LLP to "complete a comprehensive balance sheet audit."
- November 7, 2017: Leaked documents dubbed “Paradise Papers” reveal Bitfinex and Tether are run by the same individuals.
- November 19, 2017: Tether is hacked, with 31 million USDT suddenly disappearing. Tether Limited reacts to this by creating a hard fork.
- December 4, 2017: Right after hiring the PR firm 5W to help improve their image, Bitfinex hires law firm Steptoe & Johnson and threatens legal action against critics.
- December 6, 2017 - CFTC issues a subpoena to Tether and Bitfinex. This news isn't made public until the end of January.
- December 21, 2017 : Without making any formal announcement, Bitfinex appears to suddenly close all new account registrations. Those trying to register for a new account are asked for a mysterious referral code, but no referral code seems to exist.
- After a month of being closed to new registrations, Bitfinex announces it is reopening its doors, but now requires new customers to deposit $10,000 before they can begin trading.
- Friedman LLP completely cut ties with Tether on January 27, 2017.
Most common misconception: Tether is only a small part of the total market cap
One of the most common misconception people have about cryptocurrencies is that the "market cap" amount they see on CoinMarketCap.com is actually the amount of money that is invested in each coin.
I often hear people online dismiss any issue with tether by simply claiming its not big enough to cause any effect, saying "Well Tether is only $2.2 billion on CoinMarketCap and the market is 400 billion, its only 0.5% of the market".
But this misunderstands what market capitalization for cryptocurrency is, and just how different the market cap for Tether is to every other token. The market cap is simply the last trade price times the circulating supply. It doesn't take into account the order book depth at all. The majority of Bitcoin (and most coins) are held by those who either mined or purchased for a very low price early on and simply held on as very small portions of the total supply was rapidly bid up to their current price.
An increase in market cap of X does NOT represent an inflow of X dollars invested, not even close. A 400 billion dollar market cap for crypto does NOT mean that there is 400 billion dollars underwriting the assets. Meanwhile a 2 billion dollar Tether market cap means there should be exactly $2 billion backing up the asset.
Nobody can tell for sure exactly how much money has been invested in cryptocurrency market, but analysts from JPMorgan
found that there was only net inflow of $6 billion fiat that resulted in $300 billion market cap at the time. This gives us a roughly 50:1 ratio of market cap to fiat inflow. Prominent crypto evangelist Julian Hosp gives the following estimate
: "For a cryptocurrency to have a market cap of $1 billion, maybe only $50 million actually moved into the cryptocurrency."
For Tether however the market cap is simply the outstanding supply, 2.2 billion USDT is actually equal to 2.2 billion USD. In order to get $50 USDT you have to deposit $50 real U.S. dollars and then 50 completely new tokens will be issued, which never existed before on the market.
What is also often ignored is that Bitfinex allows margin trading, at a 3.3x leverage. Bitfinexed did an excellent analysis on how tether is entering Bitfinex to fund margin positions
There are $2.2 billion in Tether outstanding and the current market cap of the entire market is $400 billion according to CoinMarketCap. You can actually calculate Tether as a % of total fiat invested in the market according to the JP Morgan estimate, the following table outlines for a scenario of no margin lending and 15/25% of tether being on a 3.3x leverage margin account:
|Fiat Inflow/Market Cap Ratio ||Tether as % of total market (no margin) ||Tether as % of total market (15% on margin) ||Tether as % of total market (25% on margin) |
|JP Morgan estimate (50:1) ||27.5 % ||36.9 % ||43.3 % |
Even without any margin lending Tether is underwriting the worth of about 27.5% of the cryptocurrency market, and if we assume only 25% was leveraged out at 3.3x on margin we have a whole 43% of the market cap being driven by Tether inflow.
A much better indicator on CoinMarketCap of just how influential Tether is actually the volume, its currently the 2nd biggest cryptocurrency by volume
and there are even days where its volume exceeds its market cap.
What this all means is that not only is the market cap for cryptocurrencies drastically overestimating the amount of actual fiat capital that is underwriting those assets, but a substantial portion of the entire market cap is being derived from the value of Tether's market cap rather than real money
Its incredibly important that more new investors realize that Tether isn't a side issue or a minor cog in the machine, but one of the core underlying mechanisms on which the entire market worth is built. Ensuring that whoever controls this stablecoin is honest and transparent is absolutely critical to the health of the market.
Two main concerns with Tether
The primary concerns with Tether can be split into two categories:
- Tether issuance timing - Does Tether Ltd issue USDT organically or is it timed to stop downward selling pressure?
- Reserves - Does Tether Ltd actually have the fiat reserves at a 1:1 ratio, and why is there still no audit or third party guarantee of this?
Does Tether print USDT to prop up Bitcoin and other cryptocurrencies?
In the last 3 months the amount of USDT has nearly quadrupled, with nearly a billion being printed in January alone. Some people have found the timing of the most recent batch of Tether as highly suspect because it seemed to coincide with Bitcoin's price being propped up. https://www.nytimes.com/2018/01/31/technology/bitfinex-bitcoin-price.html
This was recently analyzed statistically:
Author’s opinion - it is highly unlikely that Tether is growing through any organic business process, rather that they are printing in response to market conditions. https://www.tetherreport.com
Tether printing moves the market appreciably; 48.8% of BTC’s price rise in the period studied occurred in the two-hour periods following the arrival of 91 different Tether grants to the Bitfinex wallet.
Bitfinex withdrawal/deposit statistics are unusual and would give rise to further scrutiny in a typical accounting environment.
I'm still undecided on this and I would love to see more statistical analysis done, because the price of Bitcoin is so volatile while Tether printing only happens in large batches. Simply looking at the Bitcoin price graph over the last 3 months and then the Tether printing its pretty clear there is a relationship but it doesn't seem to hold over longer periods.
Ultimately to me this timing isn't that much of an issue, as long Tether is backed by US dollars. If Bitfinex was timing the prints then it accounts to not much more than an organized pumping scheme, which isn't a fundamental problem. The much more serious concern is whether those buy order are being conducted on the faith of fictitious dollars that don't exist, regardless of when those buy orders occur.
Didn't Tether release an audit in September?
Some online posters have recently tried to spread the notion that Tether has actually been audited by Friedman LLP and that a report was released in September 2017. That was actually just a consulting engagement, which you can read here: https://tether.to/wp-content/uploads/2017/09/Final-Tether-Consulting-Report-9-15-17_Redacted.pdf
They clearly state that:
This engagement does not contemplate tests of accounting records or the performance of other procedures performed in an audit or attest engagement. Our procedures performed are not for the purpose of providing assurance...In addition, our services do not include determination of compliance with laws and regulations in any jurisdiction.
They state right from the beginning that this is a consultancy job (not an audit), and that its not meant to be assurance to third parties. Doing a consultancy job is just doing a task asked by your customer. In a consultancy job you take information as true from the client, and you have no mandate to verify whether your customer's claims are true or not. The way they checked is simply asking Tether to provide them the information:
All inquiries made through the consulting process have been directed towards, and the data obtained from, the Client and personnel responsible for maintaining such information.
Tether provided a screenshots of twp bank balances. One of these is in the name of Tether Limited, and while the other is a personal account of an individual who Tether Limited claims has a trust agreement with them:
As of September 15, 2017, the bank held $60,919,810 in an account in the name of an in individual for the benefit of Tether Limited. FLPP obtained an engagement letter for an interim settlement plan between that individual and Tether Limited and that according to Tether Limited, is the relevant agreement with the trustee. FLLP did not evaluate the substance of the letter and makes no representation about its legality.
Even worse is that later on in Note 1, they clearly claim that there is no actual evidence that this engagement letter or trust has any legal merit:
Note 1: FLLP makes no representations about sufficiency or enforceability of any trust agreement between the trustee and the Client
Essentially what this is saying is that the trust agreement may not even be worth the paper it’s printed on.
And most importantly… Note 2:
“FLLP did not evaluate the terms of the above bank accounts and makes no representations about the clients ability to access funds from the accounts or whether the funds are committed for purposes other than Tether token redemptions”
Basically Tether gave them a name of an individual with $60 million in their account according to a screenshot, Tether then gave them a letter saying that there is a trust agreement between this individual and Tether Limited. They also have account with $382 million but no guarantee that this account holds to any lien or other commitments, or that it can be accessed.
Currently Tether has 2.2 billion USDT outstanding and we have absolutely no idea whether this is actually backed by anything, and the long promised audit is still outstanding.
What happens if its revealed that Tether doesn't have its US dollar reserves?
According to Thomas Glucksmann, head of business development at Gatecoin: "If a tether debacle unfolds, it will likely cause quite a devastating ripple effect across many of the exchanges that see most of their volumes traded against the supposedly USD-backed cryptocurrency."
According to Nicholas Weaver, a senior researcher at the International Computer Science Institute at Berkeley: "You could see a spike in prices in tether-only bitcoin exchanges. So, on those exchanges only you will see a run up in price compared to the bitcoin exchanges that actually work with actually money. So you would see a huge price diverge as people see that only way they can turn tether into real money is to buy other cryptocurrency then move to another exchange. That is a bank run."
I definitely see the crypto equivalent of a bank run, as people actually try to secure their gains an realize that this money doesn't actually exist within the system:
If traders lose confidence in it and its value starts to drop, “people will run for the door,” says Carlson, the former Wall Street trader. If Tether can’t meet all its customers’ demand for dollars (and its Terms of Service suggest that in many cases it won’t even try), tether holders will try to snap up other cryptocurrencies instead, temporarily causing prices for those currencies to soar. With tether’s role as an inter-exchange facilitator compromised, investors might lose faith in cryptocurrencies more generally. “At the end of the day, people would be losing substantial sums, and in the long term this would be very bad for cryptocurrencies,” says Emin Gun Sirer, a Cornell professor and co-director of its Initiative for Cryptocurrencies and Smart Contracts. https://www.wired.com/story/why-tethers-collapse-would-be-bad-for-cryptocurrencies/
Another concern is that Bitfinex might simply shut down, pocketing the bitcoins it has allegedly been stockpiling. Because people who trade on Bitfinex allow the exchange to hold their money while they speculate, these traders could face substantial losses. “The exchanges are like unregulated banks and could run off with everyone’s money,” says Tony Arcieri, a former Square employee turned entrepreneur trying to build a legally regulated exchange.
The way I see it, this would be how it plays out if Tether collapses:
- Tether-enabled exchanges will see a massive spike in Bitcoin and cryptocurrency prices as everyone leaves Tether. Noobs in these exchanges will think they are now millionaires until they realize they are rich in tethers but poor in dollars.
- Exchanges that have not integrated Tether will experienced large drops in Bitcoin and alts as experienced investors flee crypto into USD.
- There will be a flight of Bitcoin from Tether-integrated exchanges to non-Tether exchanges with fiat off-ramps. Exchanges running small fractional reserves will be exposed, further increasing calls for greater reserves requirements.
- The exchanges might slam the doors shut on withdrawals.
- Many exchanges that own large balances of Tether, especially Bitfinex, will likely become insolvent.
- There will be lawsuits flying everywhere and with Tether Limited being incorporated on a Carribean Island whose solvency and bankruptcy laws will likely ensure they don't ever get much back. This could take years and potentially push away new investors from entering the space.
We can't be 100% completely sure that Tether is a scam, but its so laiden with red flags that at this point I would call it the biggest systematic risk in the crypto space. Its bigger than any nation's potential regulatory steps because it cuts right into the issue of trust across the entire ecosystem.
Ultimately Tether is centralizing one of the very core mechanics of the cryptocurrency markets and asking you to trust one party to be the safekeeper, and I really see very little reason to trust Bitfinex given their history of lying and screwing over their own customers. I think that Tether initially started as a legit business to facilitate the ease of moving money and avoiding regulations, but somewhere along the lines greed and/or incompetence took over (something that seems common with Bitfinex's previous actions). Right now we're playing proverbial hot potato, and as long as people believe that Tether is worth a dollar everything is fine, but as some point the Emperor will have to step out from hiding and somebody will point out they have no clothes.
In the long term I really hope once Tether collapses we can move on and get the following two implemented which would greatly improve the market for all investors:
- Actual USD fiat pairings on the major exchanges for the major currencies
- Regulatory rules on exchange reserve requirements
I had watched the Bitconnect people insist for the last 2 years that everything about Bitconnect made perfect sense because they were getting paid daily. The scam works until one day it suddenly doesn't.
Tether could still come clean and avoid all of this "FUD" by simply getting a simple review of their banking, they don't even need a full audit. If everything was legit with Tether, it would be incredibly easy to have a segregated bank account with the funds used solely to back up Tether, then have an third party accounting firm simply review the account and a bank reconciliation statement then spend a few hours in contact with the bank to ensure no outstanding liabilities are held on that balance. This is extremely basic stuff, it would take a few hours to set up and wouldn't take a lot of man-hours for a qualified account to do, and yet they don’t do it. Why? Why hire a major PR firm and spend god knows how much money to pay professional PR representatives to attack "FUD" online instead?
I think I know why.
===Highly Recommended Articles: Superintelligence Risk Project Update II by Jeff Kaufman
- Jeff's thoughts and the sources he found most useful. Project is wrapping up in a few day. Topics: Technical Distance to AI. Most plausible scenarios of Superintelligence risk. OpenPhil's notes on how progress was potentially stalled in Cryonics and Nanotech. Superintelligence Risk Project Update by Jeff Kaufman
- Links to the three most informative readings on AI risk. Details on the large number of people Jeff has talked to. Three fundamental points of view on AI-Safety. Three Fundamental points of disagreement. An update on the original questions Jeff was trying to answer. Podcast The World Needs Ai Researchers Heres How To Become One by 80,000 Hours
- "OpenAI’s latest plans and research progress. Concrete Papers in AI Safety, which outlines five specific ways machine learning algorithms can act in dangerous ways their designers don’t intend - something OpenAI has to work to avoid. How listeners can best go about pursuing a career in machine learning and AI development themselves." Radical Book Club The Decentralized Left by davidzhines (Status 451)
- The nature of leftwing organizing and what righties can learn from it. An exposition of multiple books on radical left organization building. Major themes are "doing the work" and "decentralized leadership". Study Of The Week To Remediate Or Not To Remediate by Freddie deBoer
- Should low math proficiency students take remedial algebra or credit bearing statistics. The City University of New York ran an actual randomized study to test this. The study had pretty good controls. For example students were randomly assigned to three groups, participating professors taught one section of each group. Kenneth Arrow On The Welfare Economics Of Medical Care A Critical Assessment by Artir (Nintil)
- "Kenneth Arrow wrote a paper in 1963, Uncertainty and the Welfare Economics of Medical Care. This paper tends to appear in debates regarding whether healthcare can be left to the market (like bread), or if it should feature heavy state involvement. Here I explain what the paper says, and to what extent it is true." Becoming Stronger Together by b4yes (lesswrong)
- "About a year ago, a secret rationalist group was founded. This is a report of what the group did during that year." The Destruction Of American Cuisine by Small Truths
- America used to have a tremendous number of regional cuisines, most are dead. They were killed by supermarkets and frozen food. This has been costly both in terms of culture and health (antibiotic resistance, crop monoculture risk) ===Scott: Targeting Meritocracy by Scott Alexander
- Education and merit are different. Programming is one of the last meritocracies, this lets disadvantaged people get into the field. If a job is high impact we want to hire on merit. The original, literal meaning of meritocracy is important. Classified Thread 2 Best In Classified by Scott Alexander
- Scott is promoting a project to accelerate the trend of rationalists living near each other. There are four houses available for rent near Ward Street in Berkeley. Ward street is currently the rationalist hub in the Bay. Commenters can advertise other projects and services. Url Of Sandwich by Scott Alexander
- Standard links post, somewhat longer than usual. Opec Thread by Scott Alexander
- Bi-weekly open thread. Update on Scott and Katja's travels. Salt Lake City Meetup highlight. Topher Brennan is running for Senate. Can We Link Perception And Cognition by Scott Alexander
- SSC survey optical illusions. "So there seems to be a picture where high rates of perceptual ambiguity are linked to being weirder and (sometimes, in a very weak statistical way) lower-functioning." Speculation about fundamental connections between perception and cognitive style. Ideas for further research. Change Minds Or Drive Turnout by Scott Alexander
- Extreme candidates lower turnout among their own party. Is base turnout really the only thing that matters? Lots of quotes from studies. ===Rationalist: Learning From Past Experiences by mindlevelup
- "This is about finding ways to quickly learn from past experiences to inform future actions. We briefly touch upon different learning models." Model-based and Model-Free reinforcement learning. Practical advice and examples. How Long Has Civilization Been Going by Elo (BearLamp)
- Human agricultural society is only 342-1000 generations old. "Or when you are 24 years old you have lived one day for every year humans have had written records." Human civilization is only a few hundred lifetimes old. Choices Are Bad by Zvi Moshowitz
- Choices reduce perceived value. Choices require time and energy. Making someone choose is imposing a real cost. Erisology Of Self And Will: The Gulf by Everything Studies
- "Part 4 will discuss some scientific disciplines with bearing on the self, and how their results are interpreted differently by the traditional paradigm vs. the scientific." Philosophy Vs Duck Tests by Robin Hanson
- Focusing on deep structure vs adding up weak cues. If it looks like an x... More discussion of whether most people will consider ems people and/or conscious. Knowing How To Define by AellaGirl
- "These are three ways in which a word can be ‘defined’ – the role it plays in the world around it (the up-definition), synonyms (lateral-definition), and the parts which construct the thing (down-definition)." Applications to morality and free-will. Change Is Bad by Zvi Moshowitz
- "Change space, like mind space, is deep and wide. Friendly change space isn’t quite to change space what friendly mind space is to mind space, but before you apply any filters of common sense, it’s remarkably close." A long list of conditions that mean change has lower expected value. Why we still need to make changes. Keep your eyes open. Meditation Insights Suffering And Pleasure Are Intrinsically Bound Together by Kaj Sotala
- The concrete goal of meditation is to train your peripheral awareness. Much suffering comes from false promises of pleasure. Procrastinating to play a videogame won't actually make you feel better. Temptation losses its power once you truly see the temptations for what they truly are. Be My Neighbor by Katja Grace
- Katja lives in a rationalist house on ward street in Berkeley and its great. The next step up is a rationalist neighborhood. Katja is promoting the same four houses as Scott. Be her neighbor? What Value Subagents by G Gordan (Map and Territory)
- Splitting the mind into subagents is a common rationalist model (links to Alicorn, Briene Yudkowsky, etc). However the author preferred model is a single process with inconsistent preferences. Freud. System 1 and System 2. The rider and the Elephant become one. Subagents as masks. Subagents as epicycles. The Order Of The Soul by Ben Hoffman (Compass Rose)
- The philosophy of accepting things vs the impulse to reshape them. Many philosophical and psychological models split the soul into three. Internalized authority vs seeing the deep structure of moral reality. In some sense math is the easiest thing in the world to learn. School poisons the enjoyment of rational thought. Lockhart's lament. Feynman. Eichmann and thinking structurally. Aliens Merely Sleeping by Tyler Cowen
- The universe is currently too hot for artificial life to be productive. Advanced civilizations might be freezing themselves until the universe cools. "They could achieve up to 1030 times more than if done today" [short] Book Reviews by Torello (lesswrong)
- Rationalist Adjacent. Each book has an interesting 'ideas per page' rating. Homo Deus, Sapiens, Super-intelligence, Surfaces and Essences, What Technology Wants, Inside Jokes, A Skeptic's Guide to the Mind. Geometers Scribes Structure Intelligence by Ben Hoffman (Compass Rose)
- "How does spatial reasoning lead to formal, logical reasoning?" Fluid and crystalized intelligence. Some history of Philosophy. How social dynamics lead to the evolution of reasoning. Talmudic and Western law, and their oddities. Universal Grammar and connecting with the divine. FizzBuzz. High Dimensional Societies by Robin Hanson
- In high dimensional space the distance between points varies less. What implications does this have for 'spatial' social science models (ex analogues of 1D spectrums and 2D graphs). Feelings In The Map by Elo (BearLamp)
- Confusion is not a property of the external world. The same holds for many emotions. Non-violent communication and speaking from your own perspective. Lesswrong Is Not About Forum Software by enye-word (lesswrong)
- The best way to increase activity on lesswrong is to get back the top posters, especially Scott and Eliezer. Explication by mindlevelup
- "This essay is about explication, the notion of making things specific. I give some examples involving Next Actions and systematization. This might also just be obvious to many people. Part of it is also a rehash of Act Into Uncertainty. Ultimately, explication is about changing yourself." Concrete Instructions by Elo (BearLamp)
- "The objective test of whether the description is concrete is whether the description can be followed by an anonymous person to produce the same experience." Some examples including the 'paper folding game'. Human Seems Low Dimensional by Robin Hanson
- 'Humanness' seems to be a one dimensional variable. Hence people are likely to consider ems conscious and worthy of decent treatment since ems are human-like on many important factors. Some discussion of a study where people rated how human-like various entities were. Erisology Of Self And Will: A Natural Offering by Everything Studies
- A description of naturalism and it relation to science. Daniel Dennet. Many philosophers are still dualists about the self. The self as a composite. Freedom as emergent. The Hungry Brain by Bayesian Investor
- A short review that focuses on the basics of Guynet's ideas and meta-discussion of why Guynet included so much neuroscience. "Guyenet provides fairly convincing evidence that it’s simple to achieve a healthy weight while feeling full. (E.g. the 20 potatoes a day diet)." Boost From The Best by Robin Hanson
- [Age of Em] How many standard deviations above the mean will be the best em be? How much better will they be than the second best em? How much of a wage/leisure premium will the best em receive. Becoming Stronger Together by b4yes (lesswrong)
- "About a year ago, a secret rationalist group was founded. This is a report of what the group did during that year." In Praise Of Fake Frameworks by Valentine (lesswrong)
- "I use a lot of fake frameworks — that is, ways of seeing the world that are probably or obviously wrong in some important way. I think this is an important skill. There are obvious pitfalls, but I think the advantages are more than worth it. In fact, I think the "pitfalls" can even sometimes be epistemically useful." Letter To Future Layperson by Sailor Vulcan (BYS)
- A letter from someone in our age to someone post singularity. Description of the hardships and terrors of pre-singularity life. Emotional and poetic. ~5K words. ===AI: Conversation With An Ai Researcher by Jeff Kaufman
- The anonymous researcher thinks AI progress is almost entirely driven by hardware and data. Back propagation has existed for a long time. Go would have taken at least 10 more years if go-aI work had remained constrained by academic budgets. Openai Baselines PPO by Open Ai
- "We’re releasing a new class of reinforcement learning algorithms, Proximal Policy Optimization (PPO), which perform comparably or better than state-of-the-art approaches while being much simpler to implement and tune. PPO has become the default reinforcement learning algorithm at OpenAI because of its ease of use and good performance." Superintelligence Risk Project Update II by Jeff Kaufman
- Jeff's thoughts and the sources he found most useful. Project is wrapping up in a few day. Topics: Technical Distance to AI. Most plausible scenarios of Superintelligence risk. OpenPhil's notes on how progress was potentially stalled in Cryonics and Nanotech. Real Debate Robots Education by Tyler Cowen
- Robots are already becoming part of the classroom. K-12 is an artificially creation anyway. Robots can help autistic or disabled children. Children sometimes trust robots too much. Robust Adversarial Inputs by Open Ai
- "We’ve created images that reliably fool neural network classifiers when viewed from varied scales and perspectives. This challenges a claim from last week that self-driving cars would be hard to trick maliciously since they capture images from multiple scales, angles, perspectives, and the like." What Is Overfitting Exactly by Andrew Gelman
- "If your model is correct, “overfitting” is impossible. In its usual form, “overfitting” comes from using too weak of a prior distribution." Conversation With Bryce Wiedenbeck by Jeff Kaufman
- "AGI is possible, it could be a serious problem, but we can't productively work on it now." AGI will look very different from current technologies. Utility functions are a poor model of human behavior. Examples Of Superintelligence Risk by Jeff Kaufman
- A series of extended quotes describing ways AI with innocent seeming goals can destroy the world. Authors: Nick Bostrom, Eliezer (and collaborators), Luke M, 80K hours, Tim Urban. Jeff finds them unpersuasive and asks for better ones. Lots of interesting comments. Eleizer himself comments describing how 'paperclip maximizers' might realistically occur. Superintelligence Risk Project Update by Jeff Kaufman
- Links to the three most informative readings on AI risk. Details on the large number of people Jeff has talked to. Three fundamental points of view on AI-Safety. Three Fundamental points of disagreement. An update on the original questions Jeff was trying to answer. Conversation With Michael Littman by Jeff Kaufman
- CS Professor at Brown's opinions: Deep Learning is surprisingly brittle in his experience. General Intelligence will require large fundamental advances. The AI risk community isn't testing their ideas so they probably aren't making real progress. ===EA: EAGX Relaunch by Roxanne_Heston (EA forum)
- The EA global satellite EAGA-X conferences have been low activity. Changes: More funding and flexibility. Standardized formats. Fewer groups approved. Stipends to primary organizers. Uncertainty Smoothes Out Differences In Impact by The Foundational Research Institute
- Many inside view evaluations conclude that one intervention is orders of magnitude more effective than another. Uncertainty significantly reduces these ratios. Autonomy: A Search For A Measure Will Pearson (EA forum)
- "I shall introduce a relatively formal measure of autonomy, based on the intuition that it is the ability to do things by yourself with what you have. The measure introduced allows you to move from less to more autonomy, without being black and white about it. Then I shall talk about how increasing autonomy fits in with the values of movements such as poverty reduction, ai risk reduction and the reduction of suffering." Eight media articles on GiveDirectly, Cash Transers and Basic Income.- A world where 8 men own as much wealth as 3.6 billion people by GiveDirectly
- More Giving Vs Doing by Jeff Kaufman
- EA is moving far more money than it used to and the ramp up will continue. This means direct work has become relatively more valuable. Nonetheless giving money is still useful, capacity isn't being filled. Jeff plans on earning to give based on his personal constraints. Why I Think The Foundational Research Institute by Mike Johnson (EA forum)
- A description of the FRI. Good things about FRI. FRI's research framework and why the author is worried. Eight long objections. TLDR: "functionalism ("consciousness is the sum-total of the functional properties of our brains") sounds a lot better than it actually turns out to be in practice. In particular, functionalism makes it impossible to define ethics & suffering in a way that can mediate disagreements." Tranquilism by The Foundational Research Institute
- A paper arguing that reducing suffering is more important than promoting happiness. Axiology. Non-consciousness. Common Objections. Conclusion. An Argument For Why The Future May Be Good by Ben West (EA forum)
- Factory farming shows that humans are deeply cruel. Technology enabled this cruelty, perhaps the future will be even darker. Counterargument: Humans are lazy, not evil. Humans as a group will spend at least small amounts altruistically. In the future the cost of reducing suffering will go down low enough that suffering will be rare or non-existent. Arguments Moral Advocacy by The Foundational Research Institute
- "What does moral advocacy look like in practice? Which values should we spread, and how? How effective is moral advocacy compared to other interventions such as directly influencing new technologies? What are the most important arguments for and against focusing on moral advocacy?" An Argument For Broad And Inclusive by Kaj Sotala (EA forum)
- "I argue for a very broad, inclusive EA, based on the premise that the culture of a region is more important than any specific group within that region... As a concrete strategy, I propose a division into low-level and high-level EA" Not Everybody wants a Goat by GiveDirectly
- Eight links on GiveDirectly, Cash Transfers, Effective Altruism and Basic Income. Mid Year Update by The GiveWell Blog
- Encouraging more charities to apply. More research of potential interventions. Short operations recap. GiveWell is focusing more on outreach. ===Politics and Economics: College Tuition by Tom Bartleby
- Sticker prices for college have gone up 15K in twenty years, but the average actual cost has only gone up 2.5K. High prices are almost compensated by high aid. Advantage: more equitable access to education. Disadvantages: Not everyone knows about the aid, financial aid is large enough it can seriously distort family financial decisions. War Of Wages Part 1 Apples And Walmarts by Jacob Falkovich (Put A Number On It!)
- The Author thinks minimum wage hurts the poor. Walmart can't afford higher wages. Copenhagan Interpretation of Ethics: Walmart helps the poor and gets blamed, Apple does nothing for the poor but avoids blame. Links 10 by Artir (Nintil)
- Tons of links. Economics, Psychology, AI, Philosophy, Misc. Pretend Ask Answer by Ben Hoffman (Compass Rose)
- A short dialogue about Patriarchy and the meaning of oppression. Defensive actions are often a response to bad faith from the other side. Its not ok to explicitly say you think your partner is arguing in bad faith. Cultural Studies Ironically Is Something Of A Colonizer by Freddie deBoer
- An origin story for Writing Studies. The fields initial methodological diversity. Cultural studies took over the field, empirical work has been pushed out. Evidence that some cultural studies professors really do believe its fundamentally bigoted to do science and empirical research endangers marginalized students. The field has become insular. The Dark Arts Examples From The Harris Adams Debate by Stabilizer (lesswrong)
- The author accuses Scott Adams of using various dark Arts: Changing the subject, Motte-and-bailey, Euphemisation, Diagnosis, Excusing, Cherry-picking evidence. Study Of The Week Modest But Real Benefits From Lead Exposure Interventions by Freddie deBoer
- Freddie reviews a survey he found via SSC. The study had very good controls. Methodology is explained and key graphs are posted and discussed. Scott and Freddie seem to agree on the facts but have a different opinion on how large to consider the effects. Descriptive And Prescriptive Standards by Simon Penner (Status 451)
- Leadership means winning the Keynesian Beauty Contest. Public opinion doesn't exist as a stable reality. Prescribing public opinion. Dangers of social reform and leaders twisting the facts to promote noble outcomes. A Taylorism For All Seasons by Lou (samzdat)
- "Christopher Lasch – The Culture of Narcissism, part 1/X, current essay being more of an overview." A Masquerade where you must act out the mask you choose. Mechanism Agnostic Low Plasticity Educational Realism by Freddie deBoer
- Freddie's educational philosophy. People sort into persistent academic strata. Educational attainment is heavily determined by factors outside of school's con