Results for “from the comments” 1664 found
From the comments, Vitalik Buterin
Analyst
|
I’ll answer this one here in detail because it’s probably too technical for it to be valuable to put a good answer into a Conversation with Tyler.
> 1. In retrospect, was it a good decision to have ethereum bytecodes executed on every single mining node? And if not, would he have selected sharding and plasma or a different solution?
Ultimately the answer is, yes given the knowledge we had at the time, no given what we know today. If I was doing Ethereum back then with the knowledge that I have today, I would obviously shoot straight for exactly the design that the research team is shooting for today (Casper PoS, sharding), and I would have actively encouraged developers to work on state channels and Plasma from day 1. Layer 1 scaling (sharding) and layer 2 scaling (state channels and Plasma) are complementary; gains from the two are multiplicative with each other, so it’s not a matter of A vs B, it’s A and B.
Ultimately, for a distributed validation system to work, you need to satisfy two properties:
1. There are enough (randomly sampled) nodes on average validating any given piece of data that invalid data will under no circumstances get through.
2. There are mechanisms that can ensure that if bad data *does* get through (eg. because of a 51% attack), then clients can detect this. In a sharded system, there is obviously too much data for clients to verify directly, but there are indirect approaches that can be used that can give equivalent assurances with some additional security assumptions (STARKs, fraud proofs, data availability proofs…)> 2. How confident is he that transitioning to proof-of-stake will be successful? What are the risks of proof-of-stake?
Close to 100% confident that proof of stake is possible in principle; many chains are using (crappy versions of) it already. There’s obviously the question of how strong properties we can achieve with PoS though, and there are some edges of that that are still being worked out. The main risks that I see are (i) weird game-theoretic attacks on the specific design that we end up going with, and (ii) pool centralization.
IMO Satoshi’s PoW is really nice in part because of its sheer simplicity; the simplicity helps with decentralization because pretty much anyone can understand how it works, whereas traditional non-PoW consensus algos like PBFT are far more complex. Casper FFG was designed in part to replicate something close to PoW-style simplicity while still having the safety and liveness properties of traditional BFT consensus algos; and I’m obviously interested in minimizing complexity of the sharding design as well.
Here is the link, he offers several other “highly technical” answers in the comments.
From the comments, on South Africa
I think the chances of a populist land grab in South Africa (never very high) have actually gone down over the past few months. Look at the ANC’s actions during its 24 years in power, not its rhetoric. Many bad policies for sure, but never anything close to radically populist, of the sort that would seriously scare the financial markets. Destruction of the (well entrenched and sophisticated) property rights system would certainly do that. So it’s unlikely to happen.
The only time there seemed to be a risk of edging in that direction was when Zuma and his faction started seriously losing support (2016-17). They responded by ratcheting up the populist and racist rhetoric (“white monopoly capital” etc), but ultimately it didn’t work. They lost, and power in the ANC has shifted back to the more market friendly centrists, typified by Ramaphosa.
That’s why I think the risks have gone down (since Zuma was ousted), despite the recent parliamentary vote to “expropriate without compensation”. The sound bite plays well to a certain audience, as other commenters have noted, but I agree it’s mostly just signaling. When you look at the details it’s not as scary as it sounds.
Firstly, they didn’t vote to do it, they voted to set up a committee to investigate doing it, subject to various caveats and constraints, e.g. must increase agricultural production and improve food security; there must be public and expert consultation; appropriate mechanisms, etc. It seems extremely unlikely that the ANC’s intention is to summarily expropriate all land without compensation, nor does it say that in the parliamentary motion or in any ANC policy statement (that is indeed the EFF’s position, but they have less than 10% electoral support). Far more likely is we’ll end up with some sort of watered down constitutional amendment that allows expropriation without compensation in certain defined and limited circumstances, but overall system of property rights remains intact for vast majority of land and other assets.
By the way, I suspect the most outsiders seriously underestimate the strength of South Africa’s constitution and supporting institutions. They have stood remarkably firm over the past few years in the face of concerted attempts by Zuma and his cronies to undermine them. Compared, for example, to a country like Turkey, whose constitution, judiciary, media and civil society have been crushed in the space of a few years by a similarly venal and power-deluded single politician.
That is from Greg.
Will truckers be automated? (from the comments)
Dan Hanson writes:
I wonder how many of the people making predictions about the future of truck drivers have ever ridden with one to see what they do?
One of the big failings of high-level analyses of future trends is that in general they either ignore or seriously underestimate the complexity of the job at a detailed level. Lots of jobs look simple or rote from a think tank or government office, but turn out to be quite complex when you dive into the details.
For example, truck drivers don’t just drive trucks. They also secure loads, including determining what to load first and last and how to tie it all down securely. They act as agents for the trunking company. They verify that what they are picking up is what is on the manifest. They are the early warning system for vehicle maintenance. They deal with the government and others at weighing stations. When sleeping in the cab, they act as security for the load. If the vehicle breaks down, they set up road flares and contact authorities. If the vehicle doesn’t handle correctly, the driver has to stop and analyze what’s wrong – blown tire, shifting load, whatever.
In addition, many truckers are sole proprietors who own their own trucks. This means they also do all the bookwork, preventative maintenance, taxes, etc. These people have local knowledge that is not easily transferable. They know the quirks of the routes, they have relationships with customers, they learn how best to navigate through certain areas, they understand how to optimize by splitting loads or arranging for return loads at their destination, etc. They also learn which customers pay promptly, which ones provide their loads in a way that’s easy to get on the truck, which ones generally have their paperwork in order, etc. Loading docks are not all equal. Some are very ad-hoc and require serious judgement to be able to manoever large trucks around them. Never underestimate the importance of local knowledge.
I’ve been working in automation for 20 years. When you see how hard it is to simply digitize a paper process inside a single plant (often a multi-year project), you start to roll your eyes at ivory tower claims of entire industries being totally transformed by automation in a few years. One thing I’ve learned is a fundamentally Hayekian insight: When it comes to large scale activities, nothing about change is easy, and top-down change generally fails. Just figuring out the requirements for computerizing a job is a laborious process full of potential errors. Many automation projects fail because the people at the high levels who plan them simply do not understand the needs of the people who have to live with the results.
Take factory automation. This is the simplest environment to automate, because factories are local, closed environments that can be modified to make things simpler. A lot of the activities that go on in a factory are extremely well defined and repetitive. Factory robots are readily available that can be trained to do just about anything physically a person can do. And yet, many factories have not automated simply because there are little details about how they work that are hard to define and automate, or because they aren’t organized enough in terms of information flow, paperwork, processes, etc. It can take a team of engineers many man years to just figure out exactly what a factory needs to do to make itself ready to be automated. Often that requires changes to the physical plant, digitization of manual processes, Statistical analysis of variance in output to determine where the process is not being defined correctly, etc.
A lot of pundits have a sense that automation is accelerating in replacing jobs. In fact, I predict it will slow down, because we have been picking the low hanging fruit first. That has given us an unrealistic idea of how hard it is to fully automate a job.
More sectors bad at finding talent (from the comments)
Surgery (and many medical specialties, esp. highly compensated ones) should be on the list of ‘Bad at finding best talent.’ There’s no way to show aptitude for a surgical specialty before medical school, and there is no mechanism for good surgeons to rise to the top, and bad surgeons to be identified and punished. If you make it into a surgical residency, you will succeed, even if you faked your way into med school and your surgical success rate is terrible. There is essentially no mechanisms to make sure aging surgeons learn the newest techniques, and no checks on waning competency. It is only because the training is so long and difficult that it isn’t a complete disaster.
Policing should also be on the list. It’s another job where, like being a surgeon, once you’ve made it into the profession, you have to fail spectacularly to be kicked out. At least half the police officers I know shouldn’t be allowed to carry firearms, much less have the power of life and death over ordinary citizens.
That is from Kevin, based on my earlier post on this question.
From the comments
To Tyler and to all commenters: beware mood affiliation.
It can simultaneously be true that (1) solar technology has substantial environmental and economic downsides, (2) very few people are aware of these downsides, and (3) solar technology is a great boon to humanity’s present and future (i.e., the world we live in is superior to a counterfactual world with no solar).
Also keep in mind that in the grand scheme of things, today’s decisions matter more in terms of the technological path they put us on rather than the actual kWh generated today. If solar is generating 50% of Earth’s electricity in the year 2100, then a 5-year acceleration or deceleration in the technology/market/regulation environment could be worth trillions of dollars.
Lastly, many ‘arguments’ seem to occur where one person makes a true claim with a certain mood. A commenter disagrees with that mood, and makes a different true claim. A second commenter disagrees with THAT mood and makes a third true claim. This pattern of discussion is not always healthy. We should hold ourselves to a standard higher than saying things that true. We should say things that build useful generalizeable mental models. If we only say the counterintuitive hipster ‘facts’ we can in fact paint a misleading picture even though we share only facts that rigorously true. (Tyler, I love you and your work, but this is one of the ways that I think your writing can improve. Contrarian statements, even when true, can sometimes be less good than other true statements. I understand this is vague, but I hope you understand.)
That is from Ted.
Further points on how to understand modern India (from the comments)
Good post.
There are a few other topics that can serve as useful handles to “understand” India.
1. Study the folk history of the popular Indian pilgrimage sites –
For a lot of people, Hinduism is associated with abstruse metaphysics, mysticism, Vedanta, and Yoga. And this obsession with the high falutin theoretical stuff, means that many students of Hinduism don’t pay as much attention to the pop-religion on the ground. And this religion is best understood by actually understanding the few hundred important pilgrimage sites scattered across the country. Each of these sites is ancient and has a “legend” associated with it. (the so-called Sthala Purana). The civilizational unity of India is largely accomplished because of the pan Indian reverence for these pilgrimage sites. Be it Benaras in the North, Kolhapur in the west, Srirangam in the south, or Puri in the East. A nice way to get started on this is Diana Eck’s book – “India – A Sacred Geography” where she makes a strong case for the theory that the idea of one India is one that is primarily stemming out of the pilgrimage experience of Hindus.
This study of pop religion will be messy and frustrating for people from an Abrahamic monotheistic background. But there is no better way to understand what makes Indians tick spiritually, and why every Indian is a millionaire when it comes to Religion.
2. Study of the history of Indian mathematics –
This may seem like an odd handle to understand India. But in my view it is useful, because Indian mathematical tradition that goes back to roughly 700 BCE, is one that is highly empirical, algebraic, and averse to theorizing and rigorous proofs. So it tells you a lot about the Indian mind. Which is very different from the Greek mind, in that it places a very very low premium on “neatness”, and a high premium on “improvisation”.
Unlike the Greeks, Indian mathematics is not that big on geometry. And also not that big on “visualization”. While someone like Euclid leveraged diagrams to make his point, Indian mathematicians like Brahmagupta and Bhaskara I/II, just stated results in 2-line or 4-line verses.
The Indian mathematical tradition is arguably the greatest Indian contribution to human civilization. Particularly the decimal number system, infinite series, and the algebraic orientation in general (markedly different from the Greek emphasis on geometry). The tradition includes Sulba Sutras (700BCE), Aryabhata (400CE), Varahamihira (400CE), Brahmagupta (500-600CE), Bhaskara I (600CE), Bhaskara II (1100-1200 CE), and ofcourse the famed Kerala school of mathematics (14th century). Madhava from the Kerala school approximated Pi to 13 decimal places. In more recent times, the most distinguished mathematical mind is ofcourse Srinivasa Ramanujan, very much a man in the Indian tradition, who disdained proofs and conventional rigor, and instead relied on intuition and heuristics.
3. Study of Indian poetry and music and its emphasis on meter
This is something that is again uniquely Indian – the very very high emphasis on meter. Which is a consequence of the Indian oral tradition and cultural aversion to writing. Which continues to this day. The emphasis on meter and rhyming was partly an aid to memorization and rote learning. And this emphasis begins with the Vedas (the earliest religious literature, preserved orally for some 1500 years before they were written down in the common era) And you see this in Indian poetry and even Indian film music to this day! Bollywood songs are characterized by their metrical style and perfect rhyming, which you don’t always see in western popular music. In that sense, the metrical legacy of the Vedas is still alive in popular culture.
That is from Shrikanthk.
The future of crypto-assets? (from the comments)
mariorossi
I think one problem with this view is that the bitcoin storage cost is likely a function of the capitalization of bitcoin.
Bitcoin security is a function of the amount of miners. The amount of miners is a function of mining revenue streams. If there are not enough miners, stealing bitcoins becomes possible. So the bitcoin network needs to generate enough mining revenues to keep enough miners interested. This amount has to be a function of the capitalization as if it ever diverges stealing bitcoin would become a valid strategy. So bitcoin must have a negative yield, either because you need to pay for the miners energy or because your bitcoins are going to be stolen. This is currently hidden by the wave of investment in the sector (and the funding from the bitcoin seniorage), but that’s got to stop at some point. I don’t think gold for example has similar features. Securing gold is not a function of of the dollar value.
If miners are a competitive industry, miners revenue has to roughly equal miners costs, so it’s a real cost. Compared with other fiat currency cost of storage it seems incredibly wasteful. Storing government bonds is basically costless. You might get negative yields, but those are transfers, not consumption. Gold was valuable because it doesn’t rust. It doesn’t require maintenance. Bitcoins require a lot of maintenece… I think that’s a weakness…
Here is the link.
Another unpopular idea about blockchains, from the comments
1 – My favorite unpopular blockchain ideas: 99% of corporate experiments regarding blockchains are better handled with Apache Kafka and multiple archivers. Anything that attempts to be a fast, global ledger has to accept the reality that global ordering is a limitation, not a feature, and instead use logical clocks. The intersection between blockchain enthusiast and distributed system researchers is close to zero. When we look back 100 years, Bitcoin itself will be seen as far more relevant in retrospect than blockchain technologies.
That is from MR reader Bob.
This seems whacky, yet I cannot refute it (from the comments), tax incidence department
Here is an unrelated topic, but part of the general topic of tax incidence. Do federal employees pay income tax on their wages? I know they do nominally, but that tax goes back to their employer, the federal government. So, doesn’t that mean that, while their actual salary may be lower than their official nominal salary, they actually don’t pay any tax? (NB: this is quite different from a private sector employee whose after-tax salary is less than the pre-tax salary. In that case, the difference between the two does *not* go to the employer, creating a gap between what the employer pays and what the employee receives.)
For example, suppose a private firm and the federal government both value a worker’s output at $100k/yr and the tax rate is 20%. The private firm offers the worker $100k and the worker receives $80k after paying taxes. The federal government, however, can offer the worker $125k in nominal salary, *knowing that it will receive $25k back in income tax*. The net result is that the federal government pays $100k and the worker receives $100k after taxes, i.e., the worker earns $100k tax free, $20k more than he or she would earn at the private firm. Another way of seeing this is to note that taxes paid by employees are economically equivalent to taxes paid by employers. So, if employers received rebates for income taxes paid by employees, then the net income tax would be zero. Well, the federal government *does* receive a rebate for all income taxes paid by employees!
Doesn’t this mean that taxes are doubly distortive? Not only do they discourage employment by creating a gap between what (private) employers pay and what workers receive — the usual cited distortion — they also distort the *composition* of the workforce by allowing the federal government to crowd out other employers.
That is from BC.
How simple can tax reform get? (from the comments)
The corporate income tax could be reduced to zero if all corporations were treated as pass-thru’s. However, for a variety of technical and practical reasons (too lengthy to discuss here), that is not feasible. Under the current regime, many businesses have the option to be treated as pass-thru’s (e.g., LLC’s and partnerships) and thus taxed only once at the individual rate, but for most publicly traded and very large entities, entities with foreign shareholders, etc., that is not possible or practical. One could also consider an imputation system such as used by the UK, but that is also messy.
The ideal system should treat all income at the same rate, regardless of the form of business. Currently, corporate income (including distributions) is subject to a higher rate than income from non-corporate entities. The federal marginal rate is currently 48 percent (35% + (.20 x .65) = 48%) compared with a marginal rate of 39.6% on ordinary income. These rates should be equalized and, preferably, the rate of corporate tax and the rate on distributions should also be roughly equal in order not to discourage corporate re-investment over distributions or vice versa and therefore avoid undue distortion regarding decisions on the allocation of capital. Thus, at the current marginal rate of 39.6%, the current proposal of a corporate rate of 25% would roughly achieve this with the current dividend marginal rate of 20% (25% + (.20 x .75) = 40%). Progressivity can be achieved (as it currently is) through progressive rates on the dividends/capital gains.
As someone who spent an entire professional career in the business, I find it amusing and naive that economists who lack any detailed knowledge of the Code or practical experience with its administration think it’s easy to radically “simplify the tax code”, make it “fair” to everyone, eliminate all tax avoidance, all at the same time! The three are simply not feasible simultaneously. As a wise man once said, “the life of the law has not been logic, but experience”.
The experience has also been that we need more than one type of tax in order to prevent the inevitable tax planning around one or the other. The system is complicated, but it is a result of a considerable amount of trial and error and political compromises. It can be made better, yes, but Trump’s promises are more credible than those who promise a one page tax code.
That is from Vivian Darkbloom. And from another Vivian comment:
1. “…so just tax that person”. Please explain how, absent a corporate income tax, the US is going to effectively tax foreign investors, if they invest via a US or foreign corporation. This is a major practical problem of eliminating the corporate income tax completely. It would be very difficult to get one’s ounce of tax flesh out of non-US investors and put them on equal footing with US investors (the same issue arises with a system relying solely on consumption tax). It would be very impractical to abrogate the 68 or so bilateral tax treaties the US is party to today or the treaties of friendship and commerce.
On the mark.
From the comments — a further note on planes and flying
You need to look at ticket prices inclusive of fees, not just fares. Those have continued the long run trend of falling in inflation-adjusted terms, although not every year.
Airline products across carriers have become less variable/more standardized. Price is only one element of competition. There are significant barriers to entry in the airline industry, not least of which is the prohibition on foreign ownership of US airlines. However that is hardly the only one.
The major reason Alaska Airlines purchased Virgin America was access to gates and in some cases slots at major congested airports. You not only have government-owned airports entering long-term leases with incumbent airlines, you frequently have capture of the bureaucrats running those airports by their major incumbent airline tenants. And where you have multiple airports in a metropolitan area, they’re frequently jointly run by the same bureaucracy rather than competing.
Airlines are highly profitable, though not nearly as profitable as two years ago, the biggest delta has been fuel cost tied to the price of oil. Consolidation allowed airlines to capture much of the gains of lower fuel prices for a period of time, but the smaller number of carriers returned to expansion and competition on the basis of price competing away some of those savings-driven prices.
All that said the only monopoly air routes in the US are the ones no one wants to fly and that require government subsidies in order to entice carriers into the market. Which isn’t to say that consumers wouldn’t benefit from more competition than we have today.
That is from Air Genius Gary Leff.
Clyde Schechter defends IRBs (from the comments)
This is not my view, but I am happy to present an alternative perspective for your consideration:
Yes, IRB’s sometimes do ridiculous things. But I served a total of 21 years on the IRB’s of two different institutions, and I’m sure I can match you anecdote for anecdote with obviously dangerous study protocols submitted by investigators, or protocols where the associated consent documents were blatantly misleading or so confusing that even professionals couldn’t understand them. It’s a small minority of submissions, to be sure, but it’s a recurring problem.
In my experience, most protocol delays in IRB review boiled down to issues of clarifying ambiguous language or providing additional background information so that the appropriateness of the proposal can be better assessed. I suspect that much of that could be avoided with better training of investigators on how to write their submissions. At one of the institutions where I served, my Department encouraged junior investigators to “pre-clear” their IRB submissions with me or another Department member who also served on the IRB. We were often able to spot the things that would likely catch the IRB’s attention and help those investigators revise their protocols before submitting them so that they would sail through approval without delays on the first try.
In my view, no person should ever be the judge of his/her own cause. There is nothing in the earlier rules, nor in the modified ones, that prevents an IRB from expediting the review of social science projects that plainly involves little or no risk. Such protocols can be turned around by a staff member in a day or two. But it should never be left to the investigators to make those assessments on their own.
Here is the link of origin.
West Virginia fact of the day (from the comments)
West Virginia’s average ACT score and percentage taking the test are almost identical to the national average.
That is from Slocum, here is the comment and link.
From the comments, Boonton on education
The context is the discussion of why Mafia members with college degrees earn more:
Signalling doesn’t really work IMO. Who is he signalling too? Other criminals? Customers? Why do they care? It seems if this is what it is the economy is deeply inefficient. 40% of the population needs 4 years of college to ‘signal’? So if there was some way to pick up this signal without college huge profits would await.
I suspect there’s two aspects that make college valuable:
1. Narrative creation – humans work by creating and sharing fictional narratives. College is a lot of practice at that which is a skill that carries over into business of many types.
2. Burns off immaturity. I suspect a big portion of the benefit of schooling is babysitting. It keeps kids out of the way of adults (which our economy couldn’t function otherwise…imagine if *every* day was take your kid to work day). By keeping immaturity somewhat walled off until kids grow out of it, schools prevent them from damaging their lives.
2.1 This may be somewhat related but workplaces are very, very stable. If you are changing tires at 18 there’s enough tires in the world that you can still be doing it at 59. Perhaps by starting work at a younger age, it is a bit too easy to fall into stability. School forces you to someday break things up. No matter how good you are at school you’re going to have to leave that stability upon graduation which will land you somewhere else which you’ll have to figure out. That flexibility may be more valuable than premature stability.
From the comments, on cyberattacks
As someone who does software and hardware, I don’t think we are anywhere near the point where a mix of hardware and software in everyday things will give us anything more than sorrow. We are already seeing rather scary things with the Internet of Things: Denial of service attacks larger than anything we’ve ever seen, because networked software is often faulty, and selling it only in hardware means vulnerabilities stay forever. It’s not just that someone can take over your CCTV camera, or the system controlling your lightbulbs, but that their computing power can be used to attack any business or individual at any time.
We have seen attacks this week that were large enough to shut down any online payment processor. For instance, imagine that the set of people with the resources for launching those attacks wanted to stop Hillary from taking online donations for as long as possible: I’d not bet against them being able to do that for a couple of weeks at the least, and that’s today. Every day more devices with weak security and no updates are sold. We see records of attack strength beaten every month: Akamai has trouble handling them today. The more devices we sell, the bigger the weapon we are handing out, and we are lacking any mechanisms to increase security because incentives are all wrong.
That is from Bob.
1. In retrospect, was it a good decision to have ethereum bytecodes executed on every single mining node? And if not, would he have selected sharding and plasma or a different solution?
2. How confident is he that transitioning to proof-of-stake will be successful? What are the risks of proof-of-stake?