This is basic journalistic ethics, to post a log of corrections made to an article.
[1] http://newsdiffs.org/diff/934591/934616/www.nytimes.com/2015...
http://newsdiffs.org/diff/934092/934104/www.nytimes.com/2015...
You can browse NewsDiffs for all sorts of NYT stories. This story was heavily edited, but it isn't the only heavily edited NYT piece, and there doesn't seem to be an agenda connecting which pieces are heavily edited and which aren't.
The more likely cause is that pieces that have to be written in a hurry will tend to get changed the most before their final version hits print.
It's not to say the NYT has an agenda here, but rather that there is division, at least between authors and editors. If the paper wants that division to be opaque, and hierarchical internally, than such real-time edits do it a disservice as they bring into the public those debates. If real-time edits are required to break a fast story, then either 1) they should not break stories quickly and wait to publish until their internal debates have settled - penning only 'news of record', or 2) they really should make those debates part of the public record, and not feign historical steadfastness - a timestamped 'record of news'. But to be a news of record where the record is changed is dishonest.
I can go find other examples of this happening --- of a concern raised in a quick-take being fleshed out, and thus changing the emphasis of the narrative in the story. In fact: I bet I can find a bunch of them, all in the NYT, just by taking a bit of time with NewsDiffs.org.
Do you doubt that, or can we just stipulate that this happens all the time?
If you doubt it, I'll do the legwork, but I'm going to ask that you stake a $20 bet to charity on it (if I'm right, you donate; if you're right, I do).
(Or you can do the legwork yourself!)
>The Pao NYT piece never changes the directionality of the concerns it raises.
With respect to this particular piece, I was pointing out later revisions that do change the directionality of a particular concept, namely its description of silicon valley.
"Ellen Pao Is Stepping Down as Reddit’s Chief" becomes "Silicon Valley 2, Ellen Pao 0".
"the entrenched sexist culture of Silicon Valley" becomes
"the entrenched male-dominated culture of Silicon Valley" becomes
"the entrenched sexist culture of Silicon Valley" becomes
"the entrenched male-dominated culture of Silicon Valley".
That is not just a fleshing out of a narrative, but an editorial flip-flop. Sexist is a particular and egregious charge to lay on an entire culture - an editorial statement. Male-dominated is a fact, separable from the concept of sexist. One can be proven, one cannot. Subjective vs objective. Opinion vs journalism.
I don't find that argument credible.
That doesn't mean I'm right, but it does mean that this is not a profitable line of argument for us. I'll go ahead and assert that both of those constructions mean the same thing.
I'm sure those editors pick their battles carefully. I have zero confidence that they are interested in anything but the most egregious of errors or corrections.
In terms of corrections, for non-trivial ones, it sometimes takes more than a day to get the true story...and this is sensible, as no one wants to correct the correction. This process is pretty manual.
In terms of changes, as described by the OP, it's not always clear what kind of changes deserve a manually-written line-item description. Fixing grammar and punctuation? Adding a new quote (usually when a new quote is added, an "Update" note is prepended to the story, but for most WordPress using organizations, this is again a manual process)? Rearranging the lede to put a more important, newly discovered fact up higher?
So instead of leaving it to humans to decide which changes are worth manually logging...just let the machine timestamp the changes and show the diffs, allowing the reader to see them in a transparent way. For changes that require no explanation (i.e. the addition of an Oxford comma), let them stand alone. For changes that require some context, such as a new quote, give the option of letting the reporter/editor describe the reason, e.g. "Mr. Smith responded to a reporter's calls at 3 p.m. today".