3 ms·

Or "turn on your webcam and say/do stuff" I have a feeling other humans verifying you mechanicalturk style one time is the future. But imho, the payment route

by _8j50 3y ago

Or "turn on your webcam and say/do stuff" I have a feeling other humans verifying you mechanicalturk style one time is the future. But imho, the payment route makes sense and also this is the problem as random paywalled sites and web monetization. A cash equivalent bearer of token payment method is what is needed. 1) verify+lock funds 2) get bearer to pay funds or funds unlock in x minutes, cancelling the payment 3) token is presented to payment processor who will revoke the token, issue funds to payee with a new token matching that value paid 4) anyone holding that token can pay for stuff, you can just email or move tokens with usb drives or store them im a secure vault service I believe the main obstacle to solve so many web issues are KYC laws and lack of constitution amendment level laws that give the people right to trade using a bearer token (cash or not) and transfer funds peer-to-peer without a third party or disclosing their personal info (again, all like cash).

mattnewton 3y ago
> Or "turn on your webcam and say/do stuff" Probably <1 year away from beating automated detection of this reliably and 2-3 in fooling humans a high % of the time economically. I was at siggraph recently where I think I saw a whole room of papers of people’s approaches to doing just this problem of photorealistic faces saying arbitrary things.
- sarchertech 3y ago
  If you're asking people to perform arbitrary actions on camera (not just speaking), generating that to a degree that is indistinguishable from video in real-time is much more than 2 to 3 years away.
  - mattnewton 3y ago
    Who’s verifying it though? 3d rendering won’t fool humans but could be tuned to fool bots. Unless you have a massive library of possible actions these can be pre-rendered too.
    - sarchertech 3y ago
      The OP you responded to said "I have a feeling other humans verifying you mechanicalturk style one time is the future." I don't know if that's feasible, but that's the premise. >Unless you have a massive library of possible actions these can be pre-rendered too. Combinatorial explosion gives you a massive amount of actions. Hold your right ring finger between your left thumb and pinky and move it around in a circle. There's tools that you could use to limit what you had to prerender and it would be a cat and mouse game like captcha is today, but I think until we're at the point where you can completely replace an actor, and then you can do that in real time, picking out a real human will be possible.
      - mattnewton 3y ago
        I’d take the other side of that bet if it was stipulated that the 1 year start after this became a common captcha technique. People are working on generating animation on a rig from text prompts: https://www.motorica.ai/ https://www.motorica.ai/ It would be easier if we knew the prompts ahead of time and then we just render transitions. Real time-ness and a good deal of rendering fidelity can be mitigated by pretending to have crappy network connection once the prompt is given, unless you are comfortable excluding people without reliable internet. I think the bigger issue is probably with the premise though. This kind of video rendering is much more taxing but it’s also much more taxing on the verification side. I can’t imagine a company being able to do this at scale the way captchas are to defend against bots, you’d need an automated system on the frontline.
        sarchertech 3y ago
        Most of the work going on with rendering people speaking isn't to defeat captchas, so you can't use the speed of that work to judge how fast people will be able to work to defeat a captcha. Also if it were in use people would frequently change the kind of actions they are asking for. >Real time-ness and a good deal of rendering fidelity can be mitigated by pretending to have crappy network connection once the prompt is given, Real timeless and rendering fidelity are the only part about this that isn't solved, so there's no real debate to have if those constraints aren't there. >unless you are comfortable excluding people without reliable internet. If this was a real thing, I will absolutely bet that companies would exclude people with unreliable internet once I becomes known that all scammers are pretending to have unreliable connections. "Please find a more reliable connection, or stop by the nearest Identity Check facility to authenticate yourself." >I think the bigger issue is probably with the premise though. I don't think this is likely to happen either. I think the most likely solution is trusted third party in person verification. Where I could see something like this working is if the third party verification service offered a 1 time remote verification process that functioned something like this.
  - dragonwriter 3y ago
    It also doesn't scale, in part because without extensive internal controls it is a major abuse vector in the other direction, and will dissuade legitimate users, so its of extremely limited potential utility.
    - sarchertech 3y ago
      No it doesn't but the existence was the premise of the discussion. I do think that something like this could happen if trusted 3rd party identification services start to pop up. But only as a 1 time verification with the option of doing a 1 time in person verification. Depends on how badly the given service wants to keep out bots as to whether they'll start requiring some kind of guaranteed identity check.

by _8j50 3y ago

mattnewton 3y ago

> Or "turn on your webcam and say/do stuff" Probably <1 year away from beating automated detection of this reliably and 2-3 in fooling humans a high % of the time economically. I was at siggraph recently where I think I saw a whole room of papers of people’s approaches to doing just this problem of photorealistic faces saying arbitrary things.

sarchertech 3y ago
If you're asking people to perform arbitrary actions on camera (not just speaking), generating that to a degree that is indistinguishable from video in real-time is much more than 2 to 3 years away.
- mattnewton 3y ago
  Who’s verifying it though? 3d rendering won’t fool humans but could be tuned to fool bots. Unless you have a massive library of possible actions these can be pre-rendered too.
  - sarchertech 3y ago
    The OP you responded to said "I have a feeling other humans verifying you mechanicalturk style one time is the future." I don't know if that's feasible, but that's the premise. >Unless you have a massive library of possible actions these can be pre-rendered too. Combinatorial explosion gives you a massive amount of actions. Hold your right ring finger between your left thumb and pinky and move it around in a circle. There's tools that you could use to limit what you had to prerender and it would be a cat and mouse game like captcha is today, but I think until we're at the point where you can completely replace an actor, and then you can do that in real time, picking out a real human will be possible.
    - mattnewton 3y ago
      I’d take the other side of that bet if it was stipulated that the 1 year start after this became a common captcha technique. People are working on generating animation on a rig from text prompts: https://www.motorica.ai/ https://www.motorica.ai/ It would be easier if we knew the prompts ahead of time and then we just render transitions. Real time-ness and a good deal of rendering fidelity can be mitigated by pretending to have crappy network connection once the prompt is given, unless you are comfortable excluding people without reliable internet. I think the bigger issue is probably with the premise though. This kind of video rendering is much more taxing but it’s also much more taxing on the verification side. I can’t imagine a company being able to do this at scale the way captchas are to defend against bots, you’d need an automated system on the frontline.
      - sarchertech 3y ago
        Most of the work going on with rendering people speaking isn't to defeat captchas, so you can't use the speed of that work to judge how fast people will be able to work to defeat a captcha. Also if it were in use people would frequently change the kind of actions they are asking for. >Real time-ness and a good deal of rendering fidelity can be mitigated by pretending to have crappy network connection once the prompt is given, Real timeless and rendering fidelity are the only part about this that isn't solved, so there's no real debate to have if those constraints aren't there. >unless you are comfortable excluding people without reliable internet. If this was a real thing, I will absolutely bet that companies would exclude people with unreliable internet once I becomes known that all scammers are pretending to have unreliable connections. "Please find a more reliable connection, or stop by the nearest Identity Check facility to authenticate yourself." >I think the bigger issue is probably with the premise though. I don't think this is likely to happen either. I think the most likely solution is trusted third party in person verification. Where I could see something like this working is if the third party verification service offered a 1 time remote verification process that functioned something like this.
- dragonwriter 3y ago
  It also doesn't scale, in part because without extensive internal controls it is a major abuse vector in the other direction, and will dissuade legitimate users, so its of extremely limited potential utility.
  - sarchertech 3y ago
    No it doesn't but the existence was the premise of the discussion. I do think that something like this could happen if trusted 3rd party identification services start to pop up. But only as a 1 time verification with the option of doing a 1 time in person verification. Depends on how badly the given service wants to keep out bots as to whether they'll start requiring some kind of guaranteed identity check.