But how does Evolution prune the repication mechanism itself? If an early replicator was very sloppy and mutation prone, then any possible advantages occuring by random mutation would have little chance to be tested before other random mutations overwrote them or other mutations killed off the organisms carrying that mutation. Working backwards, let's start with modern DNA, in cases where there are many additional mechanisms to cut the mutation rate so the non-random part of Evolution has more time to work. Putting DNA inside a walled cell, and making that cell nucleated, both reduce the exposure of the DNA to chemicals that can mutate copies. Multicellularity further shields the DNA from some more mutagens, and lets Evolution prune cells with bad copies by apoptosis, which can't be used by single celled organisms. Right there, we have a trend in Evolution - Nature seems to be trying to reduce error rates to target, as you put it, the Goldilocks range. "Advanced" organisms, such as us, or mosquitos or oak trees, have many features that make the selection rate occur at an optimum, where Nature gets enough time for selection processes to occur. In fact, sexual selection is probably just another form of targeting that Goldilocks range, and I'm sure a professional biologist can think of may more examples than the four I've mentioned. Some more minor steps in this pattern might include the evolution of Alcohol Dehydrogenase enzymes and others, but that's getting beyond my depth.
        But if we extrapolate a historical trend from that, the mutation rate must have been higher for 'primative' DNA based life, but the selection pressure must have been lower. Mutation must have been still higher if RNA was once the core molecule of heredity, which seems pretty solidly established. And if there's several more primative replicators, selection pressure must have moved glacially compared to the modern era. So how did selection have time even in 3 billion years to evolve DNA itself? If the earliest replicators were something like crystaline clays that were subject to a very modest amount of selection by erosion, as some biologists have speculated, how do we get the time for these to evolve through many stages to RNA and then DNA and eventually all the extra trimmings of today? Given that we've been in a DNA based biosphere for close to 1.5 billion years, that's about half the time since Earth cooled enough to support organic compounds,, and we're trying to cram probably at least 5 or 6 earlier replicators into less than half the time, knowing that each one was subject to less selection pressure than it's successor probably by orders of magnetude.

We can't "save a step and conclude that the universe always existed" because we think the universe had a beginning, the Big Bang. We could have saved that step if we thought the universe was Steady State. Dr. Sagan is asking this as a rhetorical question, yet he himself gave the answer not 20 pages earlier in the same book when he addressed the Steady State/Big Bang controversy in historical physics. That's showing a completely non-scientific bias and committing a logical error, and I really hoped for better from the good doctor. Fortunately, if there Is a real God, I suspect "he"s not going to be that hung up on whether his creations beleived without evidence or not.

I do taxes professionally for part of my income, and it's a mix of personal or estate returns and corps, up to a couple of companies with 500+ full time employees.
The tax code is pretty simple for many people, but I certainly would not say the vast majority of either individuals or small businesses. I can make quick, easy money by examining a few typical returns done on a free website or $ 39 software. About 6 out of 10 will have done something wrong or missed something entirely. That's higher than the industry average reported (which is about 33%), but I'm presorting by cases where the person has either a schedule D, E, or F, or got a K1. I could probably find significant mistakes on 45% or so of the self filed Schedule A's or EITC forms out there, but those are usually dealt with by people who have only been with the firm I work for for a few years before I ever see them.
            Three mistakes I see that can have extreme consequences are:
1. people filing schedule E for rental property and thinking amortizing the property is optional (yes, it is technically optional as the tax code is phrased, but if you don't do it, the law wiill treat it as if you did, and 'recover' some of the money you never got in the first place. when you sell the property - it's 'optional' in the same sense as a parachute is optional in skydiving). I also see the vast majority of people who have other things than rent to report on an E (authorial royalties, natural gas wells, and such), have absolutely no idea what to do.
2. people filing a schedule D for sale of stock. The minor mistake about 50% of the self filers make is to spend up to 30 hours or so filling in tons of individual lines for each transaction - almost nobody who isn't a pro knows how to report groups of transactions the way the IRS wants, and the personal software will gladly let you type in every single entry from a typical 15 page brokerage statement manually if you want. By they way, I have heard from IRS agents that going to all this extra trouble increases your chance of an audit - they figure that anybody giving them all those details just might be trying to hide something among them. The major mistake is not knowing the difference between long term and short term and/or covered and non-covered transactions, and all those things that are not sales of stocks but involve capital gains and so get reported with stocks. And I have never, ever, not once in my career, seen a case where someone got a K-1 that led to an entry on schedule D, and they got it right filing with Turbo-tax or similar.
3. Schedule C for self employed income. I see people getting a 1099-Misc with some other box than 7 filled in and thinking they have to do a C, all the time. I also see young people who get paid with a 1099 that does require Schedule C for the first time and think it's basically just like a W2 and report it that way. In both cases, this puts the person in a mess immediately, because if self employment taxes get done wrongly that means the IRS and the Social Security administration both have issues with the filer, and any corrections have to propagate to both agencies before it is really fixed. I've seen way too many cases where someone spends months or even years paying off their self employment taxes, gets straight with the IRS, and then 5 years later the person gets injured, needs to collect disability and, finds out they never got credit with the Social Security Administration for working some years, and so are considered not elligibile. But the biggest mistake I see on Sched C is people claiming meals when they don't travel outside their local area or entertain clients - that happens way more often with young people new to the construction industry, than most people think, and the IRS treats every case like the taxpayer is a con artist and couldn't possibly be really that stupid. (And there's no polite way to put it, but a lot of these people are). The IRS also tends to treat this error as though the taxpayer thinks the IRS agents are boneheaded enough to believe the deliberately false claim they didn't know, and the agent auditing usually seems to feel personally insulted.

People that have a single house or two they rent out, self employed contractors and people who have a sole proprieorship that makes, say, 50 K or less net, people who get a typically sized 1099-composite statement from Wells Fargo or Merrill or T Rowe or many others - that's probably close to 40% of all filers right there. K-1s are becoming pretty common now that they're used for Family Trusts. Everybody who rents out a tiny plot of land for a Cell tower gets an E for something they don't really supervise personally, and most of them didn't study up on rental tax law even as much as the people renting a spare house.. So again, "vast majority' is an overstatement at best.
  .

Just look at what happens when you have a law based on even slightly flaky or questionable cases, and how serious it can be.
            Take Roe v Wade. It's a case where the winner used a legal pseudonym to protect her privacy, only to give statements years later publicly identifying herself and saying she has regrets about how the case was decided. Don't you just bet the anti-abortion factions have gotten renewed support from the resulting publicity? What would have happened with abortion rights if someone as unsympathetic as Larry Flynt had had the same courage Flynt did for his famous case, proceeded openly and still won, or alternately if a very sympathetic woman with more recognition of how political groups on both sides might try to twist her image had done the same? Would either situation have made any difference in the current public perception?
         

So the whole "The Drazi are divided into the Purple and the Green" thing is finally explained.

If writng style is really an identifiable characteristic, I would actually be doing you a favor by going Grammar Nazi on your last sentence. Those people who really learn enough of the manifold rules of proper English will form a group which will appear indivisible in attempts to isolate an actual individual, To stand out at all, such people will have to use words such as "eldritch", that are very, very rare, create complex compound sentences such as this one, or otherwise write unusually. People who write a run on sentence with a lack of singular/plural agreement and an ambiguous clause that undermines their actually conveying meaning, all at once, will be much easier to single out. Good luck.

Google's self-driving cars have gone 300,000 miles without an accident. That's somewhere in the neighborhood of 30–42 average-teen-driver-years worth of driving. Statistically, about 1 in five teenagers reports having an accident in any given year. So we would expect that the same number of miles driven by teenagers would have resulted in, on average, 6–8 accidents—more if we're talking about teenagers in their first year of driving.

In other words, Google's self-driving cars are already at least an order of magnitude safer than teen drivers. That's probably a statistically significant difference.

1. Ginny might have a different opinion.
2. What about exposing the remarks by L. Ron Hubbard that prove Scientology is a scam? Surely a little credit for that is due?
3. He made it through the Naval Academy, graduating academic 5th in class. Yeah, he got sick later and was discharged after he caught tuberculosis. So your definition of failure there is what? Didn't make Admiral?
4. Despite being unable to re-enlist for WW2 because of his health record, he worked at the Naval Air Experimental Station near Philadelphia, as a civilian engineer. We have Isaac Asimov's word that he was successful as part of multiple classified projects there. (With Asimov working there as a chemist on some of the same projects). It's easy to take a cheap shot at this claim however, as some of these activities are buried deep in the records of the war department, and still aren't well documented publicly.
5. Being #1 in that field he was successful in, and at one point raking in at least 50 times the income his fellow practitioners predicted was the max. possible, is not just a modest success vrs a host of failures, it's rebuilding the whole field in your own image.Writing the first story in April of 39 and having the mortgage and his electioneering debts paid off by the middle of 1940 is not a "non-failure", it's a spectacular skyrocket of a success. He wrote what is often considered the first serious modern SF film (Destination Moon), which was nominated for three Oscars and won one), Most of us would not count screen play writer and print author as just one carreer. Tell me, do you criticise Beethoven for not having done anything really OK except the sonatas?
6. At least one of several houses he designed still stands. (Bonny Doon) The hidden saferoom mechanism still works, Heinlein personally moved multi-ton boulders with block and tackle to landscape and build the pool area. The house is modernist design that takes great advantage of technology to make maintenance affordable and is generally considered a polished, professional design. That sounds like a successful architect to me, if full time professional architects themselves consider him one..
7. Heinlein built a working model of a waterbed and didn't just describe one in print, all on record before the first attempts by others to patent such a device. Inventing something which has been sold in the hundreds of millions, only counts as a failure if your only standard of success is monitization. I'm afraid you're revealing more about yourself than you want there.

A mainline Protestant would argue that the Bible is sufficient for grace. it doesn't have to be totally accurate, or directly dictated by God, to lead souls back to God. It just has to have enough in it that people end up being saved. Some denominations hold that the reason for this is so God doesn't interfere with free will in giving the writers inspiration, or even in letting the readers make up their own minds. Other denominations don't take any particular position on why it was done that way. I can see some real problems with all of this, but I'm not a mainline Protestant. Anyway, you begged the question there, and it turns out that there is another answer and the religions you are arguing with have plenty of experience with people bringing this up..

Exactly enough for a pair of breeding adults to reproduce enough times that, again on average, two of their offspring survived to reproduce, and so on. For humans, that's actually quite large, because twins and greater are rare and infant mortality is pretty extreme for mammals, so we can guess that the women of the new stone age lived to be about 25 on average, while the males should have had similar lifespans, unless the females averaged 30 and the males only made it on average to 14 but got lucky a lot in their last few months, or any other combination that balances the Darwinian books.

'neoplastic'

Did you mean neoplasmic there?

And this is one reason we don't see gigantic insects, quite aside from the usual argument that the square-cube law would make their limbs too thin to support their weight. It also means they would have to evolve better oxygen transport mechanisms.

I'm pretty sure that the neurons that are firing at 0 Hz don't contribute to the noise-saturation delimma (unless you are using some "sufficiently large values of zero"). ;-)
.

I'm reminded of the dinosaur flocking animations of Jurrassic Park. The dino herds flock about here and there, respond to events such as predator attacks, and it all looks very realistic, but the computer models there can't be what nature really uses, because they work by having some parts of the herd respond to others faster than the individual elements could really sense what the others are doing. Yet it looks realistic, and if you use the same formulae to animate model birds or sheep or such things, even trained naturalists don't see anything odd about the results. What would you call a model that produces accurate seeming results for biology, but at the cost of the biologists claiming the physicists are all wrong about faster than light transmission of information? HORRIBLE doesn't begin to describe it. Fortunately, we haven't seen a bunch of bio-informatics specialists claiming they have just disproved General Relativity - maybe there really is some humility in science.
       

Except we are seeing many cases where it is counterintuitive even to working scientists in their own fields, just which explanation is simpler.

            For example, Guth's inflationary hypothesis in Cosmology has resulted in a prediction that certain constants must be random (because otherwise, there's the implication of something we might as well call God behind the non-random values). A hypothesis that invokes God is probably not the most simple - anything that might merit the name of God is likely to be more complex than the very universe it 'explains'. Fair enough, but random values seem to imply an infinity of parallel universes, which however will never be detected by real science, only in science fiction. An infinity of untestable phenomina as the outcome of a model hardly makes that the preferred model by Occam either. Last I looked, neither one of these interpretations of the inflationary hypothesis* has been mathematically shown to be the more simple of the two. If people who have had some real impact on the specific field (i.e. Hawking), can't really agree on what they mean by simple, Occam's Razor isn't working very well.

          This has shown up in several other areas of science, for example recent math proofs by computer that are so complex there's a real chance the computer made errors during the months it was crunching numbers for the millions of steps required. Once a proof is too complex for humans to even check, how can we possibly tell whether it is more complex than another proof or not? (Counting lines of code is not a very good measure there). And while I'm hardly up on all the issues in the "universe as a giant computer" debate, I've seen arguments from some of the pros in that field that seem to show there's problem with determining which explanations are the most simple there too, and I've heard at least one working scientist in the field of sexual selection pressure complain about the same thing.

* The recent Antarctic discovery might argualbly elevate Cosmic Inflation from hypothisis to full fledged theory if it wasn't there yet. For those who think it was a theory already, these observations would seem to place it on even more solid ground, in much the same way as Crick and Watson's work helped strengthen the claim of Evolution to be a well tested and heavily supported theory. But, not being able to predict whether the initial universal constants were random or non-random is a real problem when it comes to proclaiming Cosmic Inflation has the status of a solidly tested theory, no matter how much other evidence scientists gather.