[1auralynn]

Well until they succeed in creating artificial wombs it's technically a much larger amount of information (e.g. the cellular composition of the womb, how many and what kinds of nutrients that flow through, etc). We are still scratching the surface of epigenetics too.

[thaumasiotes]

> Well until they succeed in creating artificial wombs it's technically a much larger amount of information (e.g. the cellular composition of the womb, how many and what kinds of nutrients that flow through, etc).

You might note that this information is also contained in the same notional 750MB bundle.

Imagine a piece of software that works poorly the first time you run it, but modifies its environment so that it will work better in future runs.

There was an experiment in birds related to this. Someone had the question of whether birdsong was genetically or culturally determined.

The cultural side noted that, for whatever species of bird was under investigation, birds raised without parents produced abnormal song.

But upon continuing the experiment, the genetic side noted that the children of those birds, exposed only to the abnormal song of their parents, produced normal song.

Raising one bird in isolation isn't enough to express the information contained in that species' DNA, but the information is there anyway.

[alexey-salmin]

> You might note that this information is also contained in the same notional 750MB bundle.

Not necessarily, that's the point. It's uncertain if you can recreate a human being from genome alone even with perfect technology. Some crucial information could be passed from one female to another without being present in the genome at all.

> There was an experiment in birds related to this.

I suspect none of these birds were rebuilt in a laboratory from their genome alone. The experiment proves that information is passed inside the egg but it doesn't specify via which medium.

[gus_massa]

> Some crucial information could be passed from one female to another without being present in the genome at all.

Hard to know, but two interesting corner cases:

https://en.wikipedia.org/wiki/Mitochondrion (but they have their own DNA...)

https://en.wikipedia.org/wiki/Centriole (but it looks like they can grow again "However, [...] cells whose centrioles have been removed via laser ablation [...] centrioles can be synthesized later in a de novo fashion."

[tsimionescu]

In general, the instructions for creating a copy of an object can be very different from instructions for creating an original from scratch, because instructions for copying can use information already in instance A. And DNA is certainly instructions for copying, not for building from scratch.

How far you'd get using DNA to create some organism from scratch is unclear, but it's certainly not very far at all. You certainly can't create a whole eukaryotic cell just from DNA, even with all put current knowledge of organic chemistry plus our ability to study how current cells actually work (no one has come even slightly close to building a self-replicating prokaryotic cell).

Edit: imagine source code for a compiler that is allowed to strcpy() bits of the currently running compiler. It's a legitimate way to create a running compiler, but it's not what we'd normally consider "source code".

[klipt]

Or the bird's DNA contained some objective function, and the song gradually converged to maximize that function? :-)

[thaumasiotes]

That would be an example of the song being specified in the DNA.