Thursday assorted links

1. The drone advertising wars have begun.  In Mexico City, at least.

2. Claims about evolution, and one reason why people are getting taller.

3. Half of U.S. adults are in a face recognition database.

4. “The way the students made decisions about drinking actually resembled the single most common feedback controller that’s used in engineering,” Passino said. “It’s called a proportional-derivative controller, and it measures how far a system has moved from a particular set point and adjusts accordingly. It’s the same as cruise control on a car.”  Link is here.

5. “Objects that recycle ambient radio signals can get online without a power source.

6. Doug Elmendorf’s recommendations for debt.


4. "In two papers to appear in the journal IEEE Transactions on Cybernetics, a team of social workers and engineers at The Ohio State University used mathematical models to help explain the factors that drive alcohol consumption.

They found that college students drank until they attained a certain level of drunkenness, and then adjusted the pace of their drinking—sipping versus gulping, for example, or switching to a non-alcoholic beverage—at different times throughout the night to maintain that level."


I blame the academic publication system. It's obviously a fun problem in a sense. Systems processes, filters, and state-space models are sweet. Thinking of cute clever ways to apply them to reality is a cool thing to do on a Sunday. Ultimately though it's *hard* to find novel open problems in reality that we can solve with a clever application of math.
Instead of this just being a fun blog post or in a general interest journal as like "hey guys this is kinda neat but not useful" it has been funded and coupled with 'policy prescriptions' by retards.

Truly, these are some selfless people devoted to scientific inquiry and the expansion of human knowledge. Just think of all the Friday nights they sacrificed by...attending parties.

1. The Ad would be effective if there was a dedicated Car Pool lane , which doesn't seem to be the case .

But then Ads have a tendency to drone on.

On 4, it is a proportional-integral controller, or a priprtional-integral-derivative controller, with the former being the more common one. No one uses a proportional-derivative controller, so press release fail.

It would actually be totally fascinating if these students mimicked a PD controller. But a lot of human behaviors probably match PI control if not PID.

Indeed. A PID controller has three parameters. Almost any response whatsoever could be modeled by a PID. That includes bizarre oscillations or runaway drinking until death. Saying "the behavior fits a PID", particularly with small sample sizes, is little more than saying "there is some behavior". Sure sounds more scientific, though.

A typical cruise control is almost certainly a PID controller, although it is highly likely that the derivative term is estimated using various filtering and smoothing techniques instead of directly differentiating the error signal. For a car cruise control or similar systems, PI controllers are prone to overshoots; a PD controller would have a steady-state error to a step input; a pure proportional controller would have poor damping and steady state errors. A PID controller (or something that mathematically is reasonably equivalent to PID) is really the only way to go.

Derivative is rate of change, integral is the error accumulated over time. A straight proportional will never meet setpoint if tuned, integral accumulates the error to meet setpoint. Some processes have lots of hysteresis and will respond more to input over time. Derivative dials back the input as the rate increases. It is common that these controllers self tune over a few oscillations.

It was fun to see these calculations expressed in pneumatic operators where time is expressed as pressure drop via an orifice.

2. "one reason why people might be getting taller" the UK

The blog entry didn't mention if the study accounted for migration. I would imagine that migration (both entering and leaving) would have a more substantial impact on a region's gene pool than evolution (originating within a cohort and their descendants) over the last two thousand years. But I could be wrong.

What is the difference?

When evolution is used as an explanation for an event within sub-species A, I am led to believe that the cause of the change is a result of changes from within sub-species A population, not because of the introduction of different genes from sub-species B.

My curiosity is more in relation to the distinction between external and internal influences on genetic change, if it can be thought of in this way. I don't think it is extremely surprising that migration causes genetic changes, and I suppose I am interested in understanding if there are other significant factors in this example.

In the particular case of the UK, we know that 80% of their genepool was there 6000 years ago, the remaining quarter showed up with the Anglo-Saxon invasion, and a mere (influential!) 8000 Normans in the Norman invasion.

Then they turned off the spout, such that 2016 saw more migration than the time period between 1066 and 1950 combined.

So as long as they were able to filter for "People who can trace 100% British ancestry back to 1950", then yes, this is probably at least reasonable.

There are five accepted forces that drive evolution.

1. Mutation

2. Drift

3. Selection

4. Recombination (basically mating system, sexual/asexual)


5. Migration

So scientifically speaking, migration is evolution on a population scale.

In a population genetics context, evolution is only concerned with allele frequencies. Where they come from is often interesting, but irrelevant if the question is simply looking for deviations from equilibrium (which is de facto evolution).

6. Good general advice. The budget challenge is Medicare and social security. Having better off retirees pay more for Medicare coverage seems like a very good idea: the better off retirees get the guaranteed coverage they want while paying for it. Whether the additional amount paid for coverage is structured as higher co-pays or higher premiums should be explored, the former having the benefit of modifying behavior. Social security is the more difficult challenge. As we know, collecting more in social security taxes today than payment of benefits today is both a farce and fraud: it merely shifts more of the tax burden to regressive taxes and the share paid by low to middle income earners. Unfortunately, it's a farce and fraud that many on the right wish to repeat since it has been such a success (in shifting the tax burden downward). While "prefunding" social security is fine as theory, in practice it's a farce and fraud. That leave higher taxes for funding social security, by raising the wage base or character of income subject to the social security tax. My observation is that higher income seniors really like getting that social security check and would be willing to pay a little more for it. Of course, to those who are opposed to Medicare and social security, the last thing they wish to do is find a long-term solution for funding them. Indeed, the lesson they learned from the social security reform of the 1980s is that the right "solution" (shifting more of the tax burden downward) can turn ordinary Americans into tax protesters. That's the "solution" they prefer.

#6 - a fluff piece. Sure, making Social Security means-based would help cut the deficit, but Elmendorf does not consider how politically infeasible this is, given Grover Norquist's Americans for Tax Reform. Further Elmendorf does not mention "PATENTS" a single time, instead trotting out the tired 'infrastructure = R&D' argument, which has failed in Japan with bridges to nowhere and would fail in the USA.

Finally, Elmendorf does not consider any drawbacks to the Argentina or Russian or (numerous other countries) solution to high debt: default. I'm serious. Tell me why a law passed by Congress saying that the US defaults on all debt owned by foreigners as of a certain date is bad? A large percent of govt debt is held by foreigners. Short term, the USA is frozen out of credit markets but Argentina, Russia, others have shown that long-term, creditors forget. There's a sucker (creditor) born every minute.

I have some related Twitters for you Ray, on Patents and on Tyler


Nothing wrong with your proposal to default as long as you don't mind a huge global depression, which would damage the US far more than whatever tax increases might be needed to avoid default.

Isn't the current Medicare law charging higher income retirees enough more?

My guess is Warren Buffett is paying at least $100,000 a year for his Medicare coverage.

Besides paying a higher Medicare Part B premium, $389.80 per month instead of the standard premium of $121.80, he's also paying "First, a 3.8% Medicare surtax is levied on the lesser of net investment income or the excess of modified adjusted gross income (MAGI) above $200,000 for individuals, $250,000 for couples filing jointly, and $125,000 for spouses filing separately."

He has a salary of $100,000 so, his something like ten million income would be subject to the 3.8% in large part. Which means he's paying over a quarter million for Medicare benefits. Which he might not be using because the business tax loopholes might mean he gets caddy health coverage from his employer for free. But he still pays the 3.9% of unearned for Medicare benefits.

Given the 3.9% goes to to the general fund, not just Part A trust fund, the Medicare surcharge is funding the Part B premium subsidies plus the various low income subsidies for Medicare Part B, Part D, and any supplemental.

Social Security is actually in the best shape of pretty much all government budgeting. It is not generating positive net revenue because of conservative policies that are intentionally designed to reduce taxed labor cost to increase untaxed or low taxed profits. By reducing labor costs, GDP growth is possible only by fiscal deficits funding government consumption, or welfare of some form to increase individual consumer spending, like EITC, food stamps, which are largely not subject to cuts or diversions.

When conservatives talk of the subsidy of employers by the EITC, which they favor increasing, that increases corporate profits but does not result in FICA tax revenue. Food stamps are a big part of Walmarts revenue and profit, so cutting off food stamps hurts corporations like Walmart. The improving economy and boomers hitting retirement hurt Walmarts worker quality, so Walmart has had to hike it's wages significantly, which reduces the EITC and food stamps going to Walmart workers, while increasing FICA taxes paid.

Going back to 60s economic policies, hiking the minimum wage would increase labor costs and by equal measure increase FICA while cutting EITC and food stamp welfare payments.

Given 80% of consumers are not low income workers, they will be very unlikely to cut their consumption of food by a measurable amount, nor the amount of shopping and their expectations of customer service. Walmart put in self serve checkout when it's wages were low in part to address the poor customer service that was common due to the low wages. When the wages were low and the checkout lines were 7 deep, I never saw workers rush to get customers on their way home.

Social Security is funded by a flat tax with the benefits being progressive and favoring the low income or the mixed home worker and wage worker or simply home worker. Given the average benefit is $1200 a month, little savings will result from cutting the maximum benefit of about $2500 a month. Ryan, in fact, talks of cutting the benefit based on means in order to increase the supplemental benefit for the poor.

I've never seen anyone actually do real estimates of the possible "savings" from means testing, just vague claims about "most Social Security retirees are wealthy".

It's SS and Medicare that are the real frauds. They're running out of new marks to keep the schemes going, so anything that hastens their demise is a good thing.

3. I have an anonymous Disqus account with no personal info or real name that suddenly created a profile photo using an image from my private Facebook page. There is a common email for both accounts but no other information or links between them. Interesting and disturbing at the same time.

Gotta create an email address just for spam and trolling.

#6. The deficit has increased 50% in the past two years, from $400 billion to $600 billion, in the face of continued economic recovery. The entitlement tsunami has started. Higher deficits over the next decade are one final give-away to Boomers. No thanks.

I would think Simpson-Bowles might get another look, eventually the numbers will force it.

The most fiscally irresponsible thing that Bush did wasn't the Iraq War, it was Medicare Part D.

We're spending $70B+ per year on prescription drugs for seniors and nobody bothered to include a funding mechanism.

Even with the tax increases pushed through by Obama (including higher taxes to fund Medicare) we're still on track to burn through the Medicare "Trust Fund" in the next 12 years. Social Security has run a deficit for every year of the Obama admistration and these deficits will only keep growing.

A grand bargain of some kind is critical to getting the medium term deficits under control and I don't see how the current Senate will ever agree to something like Simpson-Bowles.

The theory, which the last time I saw, has held up better than promised, is that every dollar to meds saves 2 in hospitalization.

If only all federal programs were so 'irresponsible'.

If you're using tax dollars to keep net tax consumers alive, you're not coming out ahead.

If you're going to socialize medicine, these are the things you need to look at.

6. Sounds good in theory, but will never work.

- Infrastructure and r&d are perceived as too male, significant funding can't possibly go there.

- the colleges will get all the "training" money, and produce more communication and grievance majors.

- anything can be justified as an "investment". Doug envisions great things being accomplished, I see money getting pissed away.

- to achieve the kind of savings needed through means tested benefit cuts, he's going to have cut an awful lot of people, not just the 1%. No way will that work.

His carbon tax and loophole cuts could gain broad support. But the spending and benefit cuts are going to have to be spread more evenly to gain political acceptance. And forget the "investment" idea.

"Infrastructure and r&d are perceived as too male"
Oh, God. Poor, little thing..

Based on what happened in 2008?

Yes. Yes they are. And this is actually a thing that influences policy.

Of course, it is... Now we know why we have money to occupy the world, but not to prevent bridges from falling. It's the women, stupid.

Doug Elmendorf's recommendations for dealing with the debt are too timid, too status quo and too socialist. It is incredibly unfair to have a system of funding retirement that pays more to those who contribute less to it and less to those who contribute more to it. Whether it is SS or government retirements the system should pay out based on what was payed into it not on how needy or unneedy someone might be. THEN if you want to make sure it does not go bankrupt you manage the payout much like a race track manages para-mutual betting. That is the money paid out cannot exceed the money taken in (or earned in the case of invested contributions), but everyone receives their compensation in proportion to what they contributed. That way the system by definition cannot go broke AND it is fair.

A better way to manage federal debt would be a constitutional amendment that requires a balanced budget, period. And a priority on spending that coincides with the constitution. For example: the constitution mandates that we have a military so a military budget is mandatory. The constitution is silent (and thus it is unconstitutional) for the federal government to provide welfare or any redistribution of wealth so it could not be part of the federal budget. The states would still be free to dabble in that stupid and losing effort but not the federal government.

Than once any federal borrowing is no lonnger constitutional the existing debt could be ended in one fell swoop. Simply call in all debt and pay for it with what would essentially be printed money. A one time deal.

We will of course do none of those things so our debt will bury us in payments and inevitably lead to a economic collapse and our retirement systems (including SS) will continue to be used to buy votes by giving away 'free stuff' until they collapse and fail everyone. After all that is what our politicians do best; lead us over the cliff while enriching themselves and their cronies.

Good stuff.

6. "To raise revenues, the natural places to start are: one, to tax carbon emissions; and two, to reduce the deductions and exemptions that narrow the tax base and thereby lower the revenue collected for any given tax rate."

Or, three, to repeal the Bush tax cuts.

And force median income families with three kids to start paying income taxes again?

Good luck getting the Senate to agree to that.

" repeal the Bush tax cuts"

Already sunset.

Obama tax cuts are the law now. Thanks Obama

Contact lens receives energy from Wi-Fi. What could go wrong?, he says as he puts the finishing touches on his handheld 1500 Watt Wi-Fi transmitter.

Let's try a 3-second pulse and see if that gets a reaction.

Think of it as a tiny Faraday cage for your eye.

6. Shocker: left-leaning Brookings Institute calls for more debt, more spending, and more taxes.

#6 "For example, the Republicans have controlled the House of Representatives for nearly 6 years, and the number of bills they have approved that would cut spending substantially—apart from repeal of the Affordable Care Act and the imposition of caps on annual appropriations enacted a few years ago—is zero."

Sure, its the GOP's fault. They should pass more bills with no chance of being signed so libs can demagogue more. The mild "sequester" was portrayed as the literal end of the world.

He ought to take the beam from his own eye.

"The mild “sequester” was portrayed as the literal end of the world."

Best comment of the thread and should be highlighted with quotes back from, was in back in 2011, about how this was horrible.

Re: #2, He doesn't seem to be talking about "evolution" so much as he does a population selecting for specific traits. That is, humans aren't turning into a new species. When most lay people talk about "evolution" they're really talking about speciation.

I would hope that the average person reading this blog would be sophisticated enough to understand we're not talking about speciation.

This is from a scientific journal. These are not lay people. The original source are scientists using scientific definitions of evolution.

Any change in allele frequencies is, scientifically speaking, evolution.

Slightly bad form to add to random links, but I will this time

The culture that is Japan or the nature that is human?

#2 I call B.S. on this height/GWAS claim: "the probability that the distributional difference occurred by chance is 0.000000000001" What is that? 12 sigma? This is what the GWAS people write when they're too stupid to realize that the complexity of the problem is beyond the ability of their model to resolve.

But this was published in a Wordpress blog! Surely the editor and the referees would have caught an inaccurate use of statistics.

#3. I guess we all just shrug about the facial recognition database and its usage. I suppose law enforcement may be a leader, but it likely won't be long until private databases create ability to track me everywhere in public - at least when I'm not wearing my chewbacca outfit.

People might not be comfortable with it, but see it as going to happen anyway and no useful way to opt out.

Comments for this post are closed