Hacker News Reader

Comment by chjj | Hacker News Reader

I find it very strange that OpenAI would post this in the middle of a lawsuit. Shouldn't all of these emails come out in discovery anyway? Publishing this only benefits OpenAI if they're betting on the case never reaching discovery. It seems like they just want to publish very select emails which paint a certain picture.

Also, DKIM signatures are notably absent, not that we could verify them anyway since the emails are heavily redacted.

View on HN

If they just waited until discovery, it would be Musk's lawyers that control the narrative, choosing which parts of the emails to focus on publicly, which to ignore and what story to paint around it.

As you say, this way, they get to control the narrative. Nothing strange at all.

From what I can tell, Musk's lawsuit doesn't have a much of chance in the first place. I don't think he expects to win, it seems to be more a tool in Musk's own media push, and while I wouldn't bet on it, there is absolutely a chance it won't reach discovery.

I think OpenAI have quite wisely decided that the real battle here is the court of public opinion. They know it's possible to win the court case but lose in the eyes of the public. And they know that Musk has a lot of previous experience (and success) in the this battleground.

OpenAI might think they're winning a PR battle by shaping a narrative here, but they are now locked into this narrative, possibly to their detriment in court. Just seems like a bad idea. I find it odd that their lawyers wouldn't steer them clear of something like this.

> I think OpenAI have quite wisely decided that the real battle here is the court of public opinion.

They've failed to win me over. As far as I can tell, their attempted PR victory hinges on a single email with a one-word reply from Musk. Their own emails are far more damning as they give a detailed explanation of why they believe AI should not be open.

To an observer who already dislikes Musk, I'm sure it's a PR win. To someone neutral or someone who dislikes both parties, it's a PR disaster.

> Also, DKIM signatures are notably absent, not that we could verify them anyway since the emails are heavily redacted.

What are you implying? Faking emails opens you up to libel. I doubt OpenAI is trying to add another lawsuit to their workload.

> What are you implying?

I'm not implying anything. I'm pointing out a lack of openness on OpenAI's part.

Unless it's brought up as a point in court, actual exhibits in lawsuits don't have DKIM signatures either.

I'm not making a legal argument.

Oh!

In that case, let me ask: where have you ever seen DKIM signatures being provided on a public blog post/press release?

Nowhere, and I criticize it every time. I've even gone to the lengths of trying to find DKIM keys for published emails in the past[1].

But at least those emails were un-redacted. To spell that out for you: in a highly-charged, highly-contentious political setting, emails were published un-redacted and theoretically verifiable with DKIM. No such possibility exists for OpenAI's blog post.

[1] https://news.ycombinator.com/item?id=24780798#24785123

What benefit would that bring, considering the points raised earlier?

The court of public opinion might be more important in some ways than any real court. People have already begun judging OpenAI based on the lawsuit.

It's only strange if you think the lawsuit has merit, and I'm yet to find anyone credible who thinks it does. If you believe the lawsuit is just a vexatious move to attack OpenAI then it makes perfect sense to just fight it as a PR battle.

I'm guessing they'll come to some sort of out of court settlement before it gets to trial.

Musk is blabbering like a broken fountain on Xitter shrug

Xitter is the #1 source of facts, so - Can't blame them for counter balancing a little.

Reply on news.ycombinator.com