Live data from Hacker News

4TB of voice samples just stolen from 40k AI contractors at Mercor

app.oravys.com

231–240 of 250 posts

Re: 4TB of voice samples just stolen from 40k AI contractors at Mercor

#231
post #97

Earlier quoted context omitted.

The irony runs deeper than the free analysis offer. The whole Mercor contractor relationship was this exact pattern: hand over studio-quality voice recordings and ID scans to get paid for data labeling work that didn't require either. "Explicit consent" was buried in the terms, and people clicked through because they needed the paycheck. Now 40k people have learned that biometrics aren't passwords. You can't rotate y…

This is an important point with biometrics that most people don't realize. When I say that biometrics aren't good security, most people are perplexed because they have seen movies and such that are high-tech where iris scans or fingerprints are the pinnacle of security. I like to tell them this story that I read somewhere a decade or so ago. It might not be a true story (I never checked) but it's a helpful way of thi…

The story I remember is French police units specifically launching focused investigation on the sudden explosion of crypto people / family members getting kidnapped and having a finger or more chopped off.

I did find your story from 2005 about a man having his finger chopped off once the thieves realized they would need his appendage every time in order to start the car2.

https://news.sky.com/story/french-police-investigating-serie...

[2] https://www.newscientist.com/article/mg18624943-600-finger-c...

Re: 4TB of voice samples just stolen from 40k AI contractors at Mercor

#232

Earlier quoted context omitted.

> biometrics aren't passwords. You can't rotate your voice. "My voice is my passport. Verify me." I have to renew my passport every 10 years or so. How do I do that with my voice? I guess it's time to take some vocal lessons.

Vocal lessons are both a lot of fun and a lot of work. I haven't been using any voiceprint systems but I know most humans are unable to tell that my trained voice is the same physical person as my old voice. Would be curious to find out if an AI voiceprint system can discern whether it's the same or not.

You’ll really like this then, it’s a clip of Phil Hendrie who I recently discovered. He does tons of voices and sound effects, his studio has multiple microphone and switches between them for different speakers.

Here is a clip of him when someone called his studio thinking they were the local Pizza Hut. Phil does all the other voices, including the phone system.

https://share.google/QHNkgsOdvGj7tapfk

Re: 4TB of voice samples just stolen from 40k AI contractors at Mercor

#233
post #97

Earlier quoted context omitted.

The irony runs deeper than the free analysis offer. The whole Mercor contractor relationship was this exact pattern: hand over studio-quality voice recordings and ID scans to get paid for data labeling work that didn't require either. "Explicit consent" was buried in the terms, and people clicked through because they needed the paycheck. Now 40k people have learned that biometrics aren't passwords. You can't rotate y…

This comment is pure LLM. I feel like we're right on the threshold where we give up and start interacting with slop like it's human written.

An LLM comment from an AI booster. How poetic.

Re: 4TB of voice samples just stolen from 40k AI contractors at Mercor

#235
post #97

Earlier quoted context omitted.

The irony runs deeper than the free analysis offer. The whole Mercor contractor relationship was this exact pattern: hand over studio-quality voice recordings and ID scans to get paid for data labeling work that didn't require either. "Explicit consent" was buried in the terms, and people clicked through because they needed the paycheck. Now 40k people have learned that biometrics aren't passwords. You can't rotate y…

This comment is pure LLM. I feel like we're right on the threshold where we give up and start interacting with slop like it's human written.

What makes you think that comment is LLM?

Re: 4TB of voice samples just stolen from 40k AI contractors at Mercor

#237
post #217

Earlier quoted context omitted.

HSBC offered voice verification years ago and I just laughed and said nope. I don’t even use biometrics on apple devices, I use a 6 digit pin. It was always a stupid idea. The thing about been willing to trade convenience for security is you get called paranoid and then when the other shoe does drop and you are still doing that you still get called paranoid for the current thing you are not doing that “everyone does”…

> I don’t even use biometrics on apple devices Assuming Apple is truthful on this matter (so far it seems so), Apple devices store a mathematical representation of the data, not the data itself (i.e. not a picture of your finger) and keep it only on device on a special hardware section designed for extra security. When apps ask for authentication, they can never inspect the data, they can only ask “does this match?”.…

I broadly trust apples stance on all of that.

I can however revoke a 6 digit pin at any time, my fingerprints and/or face not so much.

So entirely for me, the balance is use a pin, I understand that different people feel differently and that is entirely their choice as it should be.

Re: 4TB of voice samples just stolen from 40k AI contractors at Mercor

#238

Earlier quoted context omitted.

I'm referring to speaking, not singing. After a _lot_ of work, I can speak passably as a woman or man and switch freely between the two. Depending on context I generally choose just one for the entire conversation, as switching tends to cause whiplash in the listener (^_^).

Does the f0 change? Or is it like power distribution of harmonics change? Or is it something else?

More or less everything changes. For trans men who are on HRT the voice's lowest pitch will get lower, as it would for someone AMAB going through puberty (since a second puberty is literally what's happening on HRT). Trans women do not get any voice changes from HRT though, so they train to raise their larynx when speaking to get up into the "perceived female range."

But pitch is far from the only thing that someone gendered one way or another in western culture (and presumably elsewhere). Resonance, weight, breathing patterns, word choice, and prosody all matter too. That's way too much to go into in a post here on HN, but the easiest one to understand is resonance or "size." Female-perceived speakers have higher resonance / smaller size. This means that some of the higher harmonics are amplified more than the lower harmonics, an it's called a "small size" because the actual resonating area from the larynx to the tongue is made smaller (mostly through tongue placement). Male-perceived speakers do the opposite, creating a larger space for resonance and resulting in a lowered resonance.

I know quite a few cis people who are also going through some of this training to help with their voice acting, or even just for fun.

There are a lot of good (and unfortunately some bad) resources online for trans voice training in both directions. My personal favorite (and where I started my lessons) is Seattle Voice Labs, but Online Vocal Coach / Vox Nova is also a great resource.

Re: 4TB of voice samples just stolen from 40k AI contractors at Mercor

#239
post #40
post #3

The only data that cannot be stolen or leaked is data that doesn't exist. Hard lesson for both users and companies. Germans (because of course) have a word for this: "Datensparsamkeit". Being frugal with your data.

> Germans (because of course) I don't know if it's the reason you imply. In the 70s, there were big debates in Germany about privacy and data storage. They spoke of one's data shadow (Datenschatten). I suspect this word comes from that tradition. The reason the word exists would then be the reflection (Verwaltigung) on WW2.

> Verwaltigung

is not a word -- I think you mean Vergangenheitsbewältigung?

https://en.wikipedia.org/wiki/Vergangenheitsbew%C3%A4ltigung

Re: 4TB of voice samples just stolen from 40k AI contractors at Mercor

#240
post #44

> If you were a Mercor contractor and you believe your voice may already be in circulation, ORAVYS will analyze the first three suspect samples free of charge. Awesome, if you're a victim of an AI company having your voice, you can help yourself by sending another AI company your voice! > Audio is never used to train commercial models without explicit consent I'm sure Mercor has explicit consent as well, legal teams…

Hi oefrha. I totally understand your point of view. The only thing is: we give you an opportunity to recognize how unique your voice actually is. We only analyze the audio file to show you how much uniqueness a voice carries. Audio is never used for training without explicit consent and is purged on a defined retention schedule. That is very important to us and also to let people know what their voice has to offer them for their growth and self awareness. Big difference wiTh what Mercor did.
Post reply on HN