Hacker News Reader

Comment by BoiledCabbage | Hacker News Reader

If nobody is willing to put their names on those statements then there is absolutely zero reason to believe it is anything other than one of their competitors trying to take advantage of an opportune time to weaken the forerunner.

This is worthless and shouldn't be on the HN homepage.

View on HN

> Contact

> We encourage former OpenAI employees to contact us at formerly_openai@mail2tor.com. We personally guarantee everyone's anonymity in any internal deliberations and public communications.

Sounds like they're just being intentionally protective of everyone's identities, which is understandable.

Also "shouldn't be on the HN homepage" is not really a pertinent comment, seeing as it is on the front page, meaning others (via the algorithm) do think it is a worthy submission.

This is unauthenticated, unverified, anonymous. There is an enormous difference between journalism and this.

Anyone reading this thread right now can post any random text to a website and claim to be a former employee. Anyone can make up anything at all. Are all of those content that should be on the homepage? They would be up voted.

Credibility is important when making accusations, accountability is important. There is zero here. There are processes for actual whistle-blowers. This isn't it. It's worse than "yellow journalism". And it's concerning people don't see the difference.

You can click "hide" and move on. Everyone else who wants to chat about it can stay on the thread

Don't censor my feed with your preferences.

Anonymity should be extolled and protected.

What sort of free speech would we have on this site if people could read stuff that I don’t like? None at all

> Also "shouldn't be on the HN homepage" is not really a pertinent comment, seeing as it is on the front page, meaning others (via the algorithm) do think it is a worthy submission.

I would argue that, as in a lot of cases with articles that appear on the FP, this one may be getting upvoted as folks are around for the discussion, not necessarily the featured article, as it's just "another clue" as to what the heck is going on with this circus show.

That may very well be the case, but even then there's no reason to disallow it from the front page necessarily. Let readers make their own judgement about whether it's genuine. Heck, that's an interesting discussion in and of itself.

> Let readers make their own judgement about whether it's genuine.

Looks like they have -- it's been flagged (but not by me -- I'm not informed enough about this to make that kind of call).

I noticed, but sadly we can't vouch for submissions, which means only the `flag`s count, so it's not exactly a fair fight

I think vouching only becomes available when it's [dead], not merely when [flagged].

Read about what happens to whistleblowers who aren't anonymous pretty much anywhere in any industry. Not pretty.

> OpenAI's employee attrition rate between January 2018 and July 2020 was in the order of 50%.

and

> Those who questioned this practice were dismissed as "bad culture fits" and even terminated, some just before Thanksgiving 2019.

Would be pretty easy to verify, and could be concerning (mostly the 50% part)

A median tenure of 2.5 years seems pretty good, compared to those employees likes other options; Amazon's average is only one year. https://amazonbound.medium.com/why-amazons-attrition-is-high...

OpenAI went through at least one purge, Musk founded it but famously is no longer associated with it. There’s a huge amount of board turnover, it would be interesting to see the causes behind that

A 50% attrition rate is worse than fast food hourly work.

I managed fast food restaurants prior to becoming a SWE and our target turnover was 130% annually so not quite to that level.

Putting your name on something like this would make you unemployable anywhere, whether this is true or false. It makes perfect sense that nobody would do so.

I got fiercely attacked the other day because i suggested it might be ok to host grandmas chicken pot pie recipe using HTTP.

But then we turn around and accept as fact an anonymous document making serious accusations without proof of authenticity or tamper resistance as if it were gospel.

I seriously have no idea what the rules are anymore.

dang is usually receptive - if you reach out to him, he can usually provide more context regarding moderation actions here.

He is receptive? Daaaaang... I'll see myself out...

OpenAI has been around for 8 years and has hundreds of employees. So it would be surprising if there wasn't at least some ex-employees somewhere that had chips on their shoulders, either for valid or completely invalid reasons. The letter is anonymous, so we have no way of knowing if it's 1 or 2 people or 100. As such, the letter is fairly useless.

I'd say the number of employees openly standing with Altman is a much more important data point.

By that logic whistleblower would be worthless too.

Journalists who work with whistleblowers typically validate that the person in question is reliable and is who they say they are. How do we know this was written by former employees?

This, honestly, feels like it ought to be removed. It is not a reliable letter without names or some kind of verification, just a random post on a website.

It's as valid as all the other articles with assumptions what is happening at OpenAI.

It a starting point for others for digging deeper and either prove or disprove the claims

Reply on news.ycombinator.com