lja

They are considered unauthorized models.

I really appreciate your feedback, I don’t mean that in a trite way. You’ve given our whole team a lot to think about in terms of how we tell our story. We are contributing in the victim’s side and we are careful about trying not to parade it around, which is why we don’t have our non-profit work on the front site. I get how that is playing to the opposite of what we intended though.

As far as $100/mo for creators, they typically have to pay $300-$400/mo for DMCA services that make the creators bring them links. We find the content and provide a self-serve portal which is all that’s needed for most major sites. So we are dramatically cheaper than the current offering.

The $25/mo plan is more or less a reflection of rising costs with everything. We do have to hire data scientist, data engineers, software engineers, design, marketing and infrastructure. It may be easy to build this for a single site but if people knew how long it took a $35k A100 to watch a single minute of video and realize there are millions of minutes of video to watch, $25/mo (cancel whenever from the portal) feels like a pretty good deal. I’m open but you would be surprised at how expensive this is.

In an ideal world, we make enough off creators to offer it for free to victims of NCII. Right now, we are self funded and hundreds of thousands of dollars in so we can’t adhere to the normal SV economics of giving it away for free, even if we want to.

All of this to not lose the point that we really enjoyed your feedback and there is a lot you’ve given us to think about.

You can’t make everyone happy.

How is NAAG disreputable?

Any issues you have, you should read the study listed above. Gather what you want from it, believe what makes sense to you. It doesn’t matter to our mission.

EDIT: regarding our marketing, I’ve cited our source.

Here's the quick read: https://www.fastcompany.com/90467411/shocking-study-finds-1-...

Here's the real study: https://www.naag.org/attorney-general-journal/an-update-on-t...

Also, I don't think these numbers could be extrapolated to every country/region/culture on earth but they do seem to hold for european and english speaking markets.

1 alert a month means that we'll automatically scan our database once a month and send you an email. It's more of an insurance policy to make sure we didn't find anything for you that month.

We're partnered with RAINN and other sexual abuse agencies to give victims in extreme situations free access to our site. Our alignment isn't perfect but we are working to be a useful tool at a low cost.

Some things provide a social good but also have expenses, this happens to be one of them. Doctors charge patients, lawyers charge their clients and we aren't anywhere near those margins.

EDIT: Twins are still a weak spot for us, but we have a plan to build a model that can tell two twins apart.

Would love clarity, what do you mean? That is the whole point of the service. If you don't submit a clear picture of yourself, there isn't any way for us to search.

We have protections in place to prevent misuse but we are aware that prevention will be a large hill for us to climb and we're working on it.

Confirmation is lower right now as we are working to get people into the app, a DMV/government license before people know who we are may be a too big of a pill to swallow for a starting company.

We do have ML models running in the background to detect abuse and the accounts are flagged to management to reach out to those customers. I don't want to go through all the cat-and-mouse strategies we may have to avoid people gaming them.

If someone has taken pictures for a company and given their model release for that image and its made its way to an adult site, then the individual should not initiate a DMCA. This is something we make our customers aware of before they DMCA content. Also, we know from first hand experience that site operators are quick to deny a DMCA if they do own rights for it. We've talked to them about this very situation.

Thanks! I've been a member for a decade and just wanted to share what I'm working on in data science.

The idea of charging $8/mo for someone to look at anyone's photos online is pretty far off brand for us. We're really looking to help people find just their own photos. I do agree the market would be larger but that's not something we're interested in doing. That breaks ethical barriers for us.

We have to charge money to cover our expenses, most people don't realize how expensive GPU's are and that's the only way to do this cost effectively at scale. Even a non-profit would incur enormous expenses because there is just no way to do this cheap.

There is no abuse of the DMCA process here. Customers are asked to sign an affidavit and we don't allow them to DMCA outside of their facial profiles.

Hi I'm Luke, co-founder.

People are warned that an image lives on the internet forever, this has been especially true for images that are not consensually taken or shared. We’re fixing that.

We created Loti (https://goloti.com), a service that uses facial recognition to help users search, find, and reclaim non-consensual intimate images and videos using a streamlined DMCA process we facilitate in our software.

Over 10 million people are victims of non-consensual image sharing in the United States alone. Those are just the people that were even aware that their images were being shared; research shows that up to 30% of victims were hacked or victims of hidden cameras and are unaware.

Our goal is to bring peace of mind that your private images stay private.

I completely agree with you on this point. AI practitioners have an ethical responsibility to communicate these shortcomings to clients. I know my firm regularly talks to clients how humans should intervene in the AI systems we build. It's a necessary conversation to ensure clients know we aren't building them something flawless.

I loved your .001% example, I'm going to steal that when I talk folks. We often describe how systems fail at scale and talk about how being wrong 1/1,000,000 times can wildly backfire at large numbers.

All that being said, I just don't want people around this forum thinking facial recognition is some fringe, low accuracy modeling exercise like they used to be. The models are actually incredibly impressive these days.

The false positive rate on quality facial recognition is actually very low. I'm not weighing in whether or not it is proper to always use it, just as someone that's deeply experienced in commercial applications of facial recognition.

SBF is collecting financial crimes like Pokémon.

I had someone on my team build this for my kids during the pandemic so they could "meet" their friends in the park while we were still quarantining. It was a fun idea but I honestly have no interest in making this into a business. I just wanted to share the idea with others and if someone would like the code let me know, I'll shore it up and have the team send it over.

My thoughts if someone else is interested in running with this kind of business is that you could charge businesses to own their locations and they could have feeds or conferences scheduled. It may allow for people to see and interact with the people they normally would in a location with the same kind of drop-in, zero friction way that a walk-in would be.