Do you have a link? It's easier to refute the specific claim, instead of using a generic claim that they are wrong.

In particular, note that one of the current guess is that the first barely living thing was based in RNA instead of DNA, so if they use DNA in the calculation it is probably wrong. (But nobody is sure. More info https://en.wikipedia.org/wiki/RNA_world )

The first problem is that we don't know the exact chemical pathway, so it's difficult to estimate how hard they are.

This article in particular is about that, in some conditions (that are similar to a small pond that get desecated from time to time) it is posible to get the four bases of RNA. The probability of the reactions change with the concentration, metallic contamination, and with other molecules that are around, so this article gives a new calculation of one of the pathways.

It's only a small steps, it only creates the bases of the RNA but this method doesn't create the sugar part of the RNA or doesn't explain how they bind with the phosphate ions. There are known pathways, for this, but we probably don't know the scenario where these reactions are easiest. There are a lot of unknow steps, a lot of unknow unknow steps and perhaps unknow unknow unknow steps. I'm optimistic that this can be solved in 200 or 300 years, but it may be more difficult.

The main point is that you don't need a full functional bacteria as the first barely living thing. You need some crappy thing that can almost reproduce itself a little.

(Some guess is that the initial version was a family of similar things with a crappy copy system that produced a similar things instead of the same thing. But nobody is sure.)

I'm more skeptical of the claim of the clocks. If the clock is made of metal, most metal get oxidized very fast. We know this process very well so it's easy to calculate how long it will survive. One of the exceptions is gold, but it is too soft so it's easy to calculate how long it will survive. I can't imagine how the mainspring gets winded, but it's difficult to calculate the probability of such a device. Also it's difficult to calculate the probability to get the shape and interconnections. But the easy part is that it will get rusty and worn out in a short geological time.

The main problem is that a almost (or fully functional) watch doesn't create more almost (or fully functional) watches. So if a lucky watch is created by chance, you must wait approximately the same amount of time to get a second one.

Instead, a short RNA piece may have some enzymatic properties to make the production of copies easier, and may have some properties to be good as a mold for copying. The details are far from clear, but some RNA have enzymatic properties and RNA is good for getting copied. So RNA is a good candidate, but again, nobody is sure.

(DNA is good for getting copied, but it doesn't have enzymatic properties. Many proteins have enzymatic properties, but they are very difficult to copy.)

This is a supremely good comment.

I tend to think statements like that say more about the knowledge and passions of the writer than about the laws of the universe.

The fundamental point underlying the Nature article is that we don't actually know "the odds of life creating itself via basic chemical pathways". What we do know is that life appeared on Earth quite soon after liquid water.

Indeed, the law of large numbers applied to quadrillions and quadrillions of atoms and essentially uncountable numbers of interactions between them yields a large likelihood for even very unlikely independent events. The important thing to remember is that bootstrapping means it only has to happen purely by chance a relatively small number of times - perhaps only once. It turns out that the number of rock clock trials are far, far fewer and they don't bootstrap and so it is much less likely for one to exist even if the independent probability may be much higher. Statistics is often paradoxical, but then again, that's why we have logical systems for dealing with it. We'd be much better off, as a species, if a better understanding of statistics and statistical analysis were more widely spread and valued. Dawkins's explaination in "climbing mount improbable" is a good one.

> the law of large numbers applied to quadrillions and quadrillions of atoms and essentially uncountable numbers of interactions between them yields a large likelihood for even very unlikely independent events

It’s not so simple. If something is extremely unlikely but also has many chances to occur, the probability of the thing occurring is a very large number divided by another very large number. In these cases, it’s not always clear if the resulting probability will be high or low.

Just stating that there are quadrillions of potential interactions is meaningless if the chances of any one forming something useful can range from one in a billion to one in a quadrillion quadrillion quadrillion. Much like zero divided by zero (or in this case infinity divided by infinity) the overall probability is pretty much undefined.

True, there's no known analytical solution to the statistical likelyhood of abiogenesis. How one defines likelihood is rather important when discussing a statistical analysis.