Jeffrey Mishlove (5:53)

Let's talk about the work you were doing before you quit your job. What were your qualifications to design a website?

Chase McCarty (6:03)

Yeah, so I've been a software engineer here in the Silicon Valley Bay area for eight years now. The job I was working, just prior to leaving, I was a founding engineer at an AI coding startup. So we were doing some of the most cutting edge work on getting the best AI models to do software engineering tasks at large enterprises. So we had large enterprise customers and our bots would go off and do the coding tasks that they would normally pay very expensive software engineers to do. Prior to that I was a software engineer at some other big companies. Certainly names any of your viewers would recognize. I certainly know how to make a website, but even without that you can just do things, Jeff, it's not that hard to learn. If you are willing to do it, anybody can sit down and learn.

Jeffrey Mishlove (6:57)

In addition to designing a website, you are working in AI and also I understand with blockchain technology.

Chase McCarty (7:06)

Yeah, that's right. So I do still have a day job. I did go back and get a big boy job after the six months were up. But I've continued to work on social RV as a side project, as a passion project since then. And my current job, I'm a staff applied AI engineer at a big tech company, a name people would recognize. And my job is to do cutting edge things with AI. And so in our platform we have some fun AI features to decode sessions and stuff like that. And then we also have blockchain features as well. So for viewers who are like unfamiliar with the blockchain or they've just heard it in the context of like fake Internet money, Ponzi schemes, things like that, we have one of the first non Ponzi use cases for blockchain. I like to say most, most blockchain use cases are about money. Ours is not at all about money. We're just using this immutable public list of messages to show that like hey, this happened at X time and this is what happened and this can never be changed. And then you know later why thing happened and that can never be changed. And because we encode the sessions on the blockchain and then pick the target that is related to the session after the session's been submitted. And we do that on chain. So you can see that it was done on chain when it was done, how it was done, you can really easily prove to yourself that it couldn't have been known to the participant while they were viewing.

Jeffrey Mishlove (8:31)

Now I was on your website earlier today and I see you now have over 4,000 individual remote viewing trials available where any user without even registering on your website can sign up and can look at the target and the response to the target. And I guess it's important to emphasize, if I understand correctly, these are all precognitive trials. The target isn't even generated until after the viewer submits their response. Is that correct?

Chase McCarty (9:12)

Yeah, that's mostly correct. There's one thing I'll correct there at the end. So they're not all precognitive. A large chunk of them are. We're in the process of building out some features to do a better exploration view of our numbers that show the sessions are statistically significant and break them down into different categories. Around 1,000 of them are precognitive sessions. Of that number you shared, that's almost 5,000 sessions. And the reason for that is because when we built the precognitive piece, we weren't sure if we'd break something. We weren't sure if it would still work. I'm a scientist. I had my own questions about which things could affect viewing accuracy. We have a couple different tasking types in those buckets to see if accuracy changed, but that's right, it's a free website. You don't even have to make an account to look through the sessions. But if you do make an account, it's free to upload sessions and practice remote viewing. We have almost 5,000 sessions that are publicly visible. When you submit a session, you have the option to make it private in case there's PII or something in your picture. And so in total, I think we have almost 6,000 with the private ones. And we have about 900 registered users as of today, which is a much larger set of users than I thought we would have. When I first built the platform, I thought the market might be be like 20 or 50 people. So excited about that. And yeah, like I said, you can look through all the sessions right now. You can see how they were scored by our AI, scored by the community. You can leave comments, things like that. And then soon we'll have this much more rich way to look through the sessions and see the statistical P value associated with them and scrutinize our statistics and our AI measurement work as well.

Jeffrey Mishlove (10:55)

One of the particularly valuable features of the website you've set up is that it gives people a chance to practice for free. They don't have to pay for one of these expensive courses. And there are a number of courses. I think they're good courses, as a matter of fact, typically hosted by people who have been through the original army program at Fort Meade, Maryland, or have Been trained by some of those people and are now teaching themselves. They're wonderful courses. But you can practice every day if people want on your system with practice targets that are then evaluated against a, I guess a pool of 10 targets I think is the pool by an AI system. And the user can also judge their own trial at the same time.

Chase McCarty (11:59)

That's absolutely right. Yeah, yeah, yeah. It's the most robust practice tool I'm aware of. So you're right. The way the judging works is once you finish your session, we have a pool of 10 random or your target plus nine other random decoy targets, a set of 10 that we show the AI in like a shuffled order. And we ask the AI to rank each of those 10 targets. How will they correspond to your session? The AI doesn't know which target the right one is. And so if remote viewing is real and you're a good viewer, the real target should get ranked 1, 2 and 3 more often than it gets ranked 8, 9 and 10. And if remote viewing is not real, then it should kind of just be random. And obviously we see that they do tend to get ranked 1, 2 and 3 more often, which is very exciting. And those are the numbers that we'll be sharing more publicly soon with the Explorer page that we're building. But yeah, most of the next best thing I found online is, you know, just a free target pool of like 20 or 50 images. If you wanted to upload your session results, you people just have to take your word that you didn't cheat, that you didn't see it before you uploaded your session. So it's a great place to practice, it's a great place to share. I've taken some of those courses from the people you mentioned. They're great courses. Like we would link to them, we would recommend people take them. But it's something you're going to continue to need to practice after you take the course. Right. And so this is the tool for that. And also like some of the great things about those, those course in a group setting we've brought to the platform. One of my favorite features of taking an in person class was always getting to see what results other people got for the same target because you might get some data that looks wrong, but then you might find out that half the class got the same data and that might change your opinion on how you did. When you submit a session for a target on Social rv, you can see every other session anyone else has submitted for that same target and look through them, which is quite fun.

Jeffrey Mishlove (13:57)

One of the other features that I appreciated is that when the AI does its judging of your transcript against the real target and nine decoys, it explains why it gave your target a particular rank.

Chase McCarty (14:15)

That's right, that's right, yeah. Yep. And we're hoping to expand that feature out soon. It's like a balance between how much do you want to show and how much do you want to keep the user blind? Because user might not want to see the decoys because they might not have done those targets yet. So we're weighing that trade off. But yeah, you can scrutinize the AI and you can see why it made the decision it made. In case you don't agree with it, you can also thumbs down it. And we collect that feedback and we try to make the AI better. And then we have some other cool AI features that we're hoping to launch soon that go one step further. And if you're familiar with, for example, Lynn Buchanan's work, one of the things that he's talked about is taking every piece of data in a session and categorizing it because he believes that viewers will be better at certain categories. So you might have a guy that's better at seeing cars or a guy that's better at seeing houses or buildings or something like that. And so that's a feature we're working on as well, where we'll take your session. We'll pull off all the categories of data, and then we'll score you over time across different categories. And you would have target pools dedicated to different categories so you can practice and get better the ones that you're bad at. And cool advanced features like that that just don't really exist anywhere. There's not really been a modern remote viewing platform in the world yet. And also some of the technologies you need to build it, like the AI stuff, has only been in existence for a year or two. So it's only been possible to do some of these things kind of recently. So we're having a lot of fun with all the cool advanced features we can build.

Jeffrey Mishlove (15:42)

Well, a person who's creating a remote viewing transcript is likely to have handwritten text plus doodles, or sometimes sophisticated drawings. And your AI system is equipped to analyze both the handwriting and the drawings.

Chase McCarty (16:04)

Isn't that cool? Isn't that cool that that works? You can just get that for free these days with the multimodal AI systems. It's so exciting. And yeah, you're right. The most important piece of a great remote viewing session is often drawings. And oftentimes you need to see the drawing in context of the target for it to make sense. It doesn't make sense on its own. It's not something you could show a person and have them explain what it is, and then that be the right explanation. Right. And so the AI has to be multimodal. But that's one of the really cool things about our decoy based judging, because the AI does get to see the transcript in context of the target, but it's still blind to the target target. Right. And so if you have some weird doodle, that only makes sense if it's in context of the target, that'll still get accounted for in your, in your judging, but it doesn't degrade the AI's blindness or the statistical significance of the judging.

Jeffrey Mishlove (16:58)

I know that there are examples in your database of instances where the AI gave the transcript of the viewer a low rank, say six or a seven or an eight, but the viewer ranked it number one. And in reviewing it, you could see that the AI just missed things that were obvious to the viewer.

Chase McCarty (17:26)

Yeah, that's absolutely right. And you'll see the community ranked the number one. In the Explore page, we have it ranked by like community votes. And some of those ones in the first page do have like low AI scores. It's quite amusing to me that the AI can miss those things. We're trying every day to make it better. It can be difficult to work in these systems, and one thing we want to really avoid is if it does a bad job on one, we don't want to just rerun it on the one, because that breaks the statistical significance of the set. We're very science first, but let's make the product as good as we can with that. When we update the AI system, for example, we would rerun it across the whole set. But also users get tied to some of their scores, and so that's a tough balance about when to do the AI. In a lot of cases, it's worse than humans. Currently, it's better than nothing, and it's better than chance. And so in aggregate, when we're running statistics, the numbers are meaningful, but the numbers are also probably worse than they should be. So when we get a statistically significant P value of like 0.02 on our sessions being better than chance, actually the real number is probably better than that. It's probably at least 0.02, if not better.

Jeffrey Mishlove (18:41)

I should mention for the benefit of our viewers, that I met you just a few weeks ago at the annual conference of the International Remote Viewing association, where you gave a very interesting presentation at the same conference. The president of irva, Deborah Katz, who happens to be a guest host here on the New Thinking Allowed channel, pointed out that they had recently done a large study evaluating remote viewing judges. And what they found to their surprise is that there was very low agreement among different judges that judging remote viewing seems to still be. Even with highly trained and experienced human judges, the interjudge correlation should be much higher than it actually is.

Chase McCarty (19:37)

The interjudge correlation and the intra judge correlation, like the same judge on Monday and Tuesday will give different scores to the same session. It's quite funny. Yeah. So we've run tests for our judge on that. That's one of the ways that we test it to see if it gives the same session the same score. Often we find that on sessions that have a decent amount of data, it will. So if you have like a pretty good session that maybe it ranked to 2, if run it several, several times, it'll probably, you know, between one and three be the ranking you see across like 50 runs. But if you have a session that only has a couple pieces of data, sometimes it can just be random. Basically there's not a lot for it to latch onto. And certainly, like, lots of the sessions on Social RV don't have a ton of data because these are just random people on the Internet and some are great viewers and some aren't. And we hope that some of the scoring and some of the commenting and like social network features will help the best ones bubble up. And we see that. So if you go to the session, the public session pages, the explore pages, you can see some of the better viewers bubbling up to the top. But yeah, there's definitely a lot of noise in the system too, that makes.

Jeffrey Mishlove (20:44)

It hard to judge the way you have it designed. If someone wants to look at the actual viewings. I noticed the first thing that popped up was a page that showed all the number one hits.

Chase McCarty (20:58)

Yeah, that's right. Yeah, yeah. And that's really exciting to me because, you know, like, we're not the first person to do this study by any means. Right. People have done better studies with more sessions than us. Like Dean Radin, I think, has done something similar with an online platform and collected like 100,000 sessions, but those sessions aren't available to see. It's still, take my word for it, I measured these somehow and here's the number. And so as far as I know, and people are free to correct me if I'm wrong, this is the largest collection of publicly explorable, publicly visible sessions in the world already. And it's got more statistical data around that than a lot of the other sessions. And we'll be open sourcing all of the statistical methods we used and the AI judging all that stuff so that anybody can download the set and rerun the same code on it and get the same result that we get.

Jeffrey Mishlove (21:51)

And it's only been active a few months.

Chase McCarty (21:54)

Yeah, we launched in May, so it's only been four or five months. We got a lot done in a short amount of time and we're hoping to keep up some pace like that. Yeah.

Jeffrey Mishlove (22:04)

What one would normally expect from judging or any process that involves human behavior, we typically get a bell curve where most of the scores are in the middle and then fewer and fewer high scores, extremely high or extremely low scores. And I'm under the impression that your data showed a kind of flattening out of the bell curve where there were more scores at the ends than one would normally expect.

Chase McCarty (22:37)

Yeah, I mean, certainly if you're expecting remote viewing might not be real and you're expecting to see random chance on the judging of the sessions, you'd expect that sort of bell curve. But what we see is not only do we get get much more sessions that are like the ones, twos and threes, showing that our viewers are better than chance, but we also show as the viewers use our platform more, they're getting better as well. And so their sessions are getting better. And so over time, per user, we're seeing them get better. And of course we're bringing in new users all the time to balance that back out. But yeah, we're seeing that the viewers are getting better the more that they practice and that the viewers are coming in and already better than chance.

Jeffrey Mishlove (23:24)

Well, this is a very important finding. I did my doctoral research on training psychic abilities and one of the things that you learn in a training paradigm, of course, is it's very difficult to eliminate all sorts of extraneous influences that might result in a higher score over time that have nothing to do with the practice and training process. So the data that you're collecting showing that your individual users are getting better the more they practice, I think is an important piece of data.

Chase McCarty (24:00)

It's very, very important and it's very exciting for the user. Right. It's more fun to practice when you see yourself getting better. We're hoping to have more tools soon to show that graph to each individual user more prominently. And also, like I said, the category based viewing so you can see yourself getting better on different categories and things like that, that. That's something really exciting. But I've talked to several other researchers in the field too, and they say that in parapsychology research, sometimes the more research you do, you'll have this decline effect where something will be statistically begin and then it'll go away. And so it's really exciting to see that the more sessions we collect, the more statistically significant the data set is becoming. And there isn't a decline effect or anything like that, at least so far. I mean, I guess it could go down and then all the way back up like a U, but we haven't started to see it go back up. So that's good.

Jeffrey Mishlove (24:50)

Well, what I experienced in my own research many, many years ago was what we call a J curve that they started out. We sometimes think of it as beginner's luck. The first time a naive person tries remote viewing, they have zero expectation. They think it's probably impossible, but what the heck, I'll give it a try. Very often they are stunned at how amazing their first experience is. That happened to me. But then it dawns on them, oh, my gosh, what have I done? And the mind starts churning and they get emotional. Am I going crazy? Or what will other people think of me? Is this really real? Am I seeing? And so they go through a period where the mind is just a little too agitated and it takes a while then to recover. And over time, I found those scores do recover. And so remote viewing is actually easy to do, but to do it consistently, that's a hard thing. And that's where practice, I think, makes a difference.

Chase McCarty (26:03)

That's right, absolutely. Yeah. I know. I saw the same effect in my own personal sessions, and I've read about the same effect in the Stargate archives and things like that. I haven't done a specific data analysis on our data to see how prominent it is, but anecdotally looking through people's sessions and looking through scores, which I do quite frequently, I've also seen that their first couple sessions tend to be quite good, and then they're bad for a bit, and then they get better if they stay on it, which is really exciting for learning because, I mean, it's great to run a platform where you just come in in the first couple sessions. They do. They have a great experience. Right. And it's just like something the universe is providing us that I couldn't control. So that's fun.

Jeffrey Mishlove (26:42)

You've mentioned now a number of plans for the future Upgrading the AI and doing more detailed statistical analysis on your database. Do you have anything else to add in terms of your vision as to where all of this is heading?

Chase McCarty (27:01)

Yeah, there's so many directions to go and so much, so many fun things to do and so many great questions that I've sort of been answered, but we can re answer and with more public data and stuff like that. I shared a bit about the categories, stuff obviously we'd like, you know, better learning, better judging, stuff like that. One of the really big ones though, that I'm really looking forward to doing is some more research about ARV associative remote viewing sessions. And so for any listeners who are familiar with the difference, ARV is used to predict the outcomes of some future event, like a sporting event or a stock market event or something like that. And that's when there are multiple potential targets a user could view associated with those different outcomes. And they try to view it, and if they see one rather than the other, then that's kind of their choice for the outcome. So you would correlate winning an event with a picture of a hammer and losing an event with a picture of a field. And if they describe the hammer, then that's their vote that the event is going to, they're going to win. And the really interesting thing about ARV is there are several prominent people in the field who have strong claims that they have successfully used it to make money in the stock market or gambling. And they'll claim that they've been 70 or 80% accurate doing that. Hal Puthoff famously said he raised money doing that, predicting silver's futures and things like that. And I certainly don't want to say that I don't believe them because I know from my own data that the remote viewing piece at least is real. But it is, in my opinion, the largest logical gap in believing these claims. Because the first question anybody asks is, okay, well, if you can predict the stock market, why aren't you rich, why aren't you exceedingly wealthy? And they are not exceedingly wealthy. And I want to understand what's going on in that gap because that is one of the things that a lot of skeptics really hold to, and they should, and I personally do. Whenever I meet a famous remote viewer, that's my first question I ask them, which is, I'm sure a little rude, but I want to know the answer. And so something we're going to do with ARV is we'll have all the sessions on the blockchain so they're publicly scrutinizable and we'll plug into like polymarket or something like that and we're going to share the results with our viewers. So anybody who participates, if they want to make bets financially, they're welcome to. We're not going to do it for them for financial or for legal reasons, but we're going to see like what happens, you know, do we make money, do we not do we make money and then not. Right. And I don't know. I don't know what's going to happen, but I'm very excited to find out and to share it publicly because, yeah, in my mind this is the biggest logical gap with the current remote viewing story.

Jeffrey Mishlove (29:45)

You may know the Wall Street Journal ran a front page story back, as I recall, in the 1980s when Russell Tarek, working, working with Keith Harari as a remote viewer, conducted trials to use ARV to predict the movement to silver futures. As I recall, they had nine hits in a row out of 10 trials. Then they tried a second time where they changed the protocols in a few different ways and they got nine misses out of 10 trials. So it does seem to me that one of the problems is that human beings get very emotional about money. It's almost a taboo topic. It's considered just as impolite to ask your friend about their personal finances as it would be to ask about their sex life. It seems to me that the strong emotions that we have associated with either the having money or not having money can interfere with the calm state of mind that is usually the state that is most conducive to ESP reception.

Chase McCarty (31:08)

Yeah, yeah, no, that's a great hypothesis. Answer for why it might not work. I am of your opinion as well. I think that that's, that might be what's going on. So we'll be able to test it again and see and maybe we don't learn anything new. Maybe we just get that same result. But given the platform we have and just the velocity of sessions coming in, there's lots of little things that we can tweak with user consent. Of course, I don't, you know, I want the users to know what we're doing. But a great example of that is like user might not even know if they're doing an ARV session, right? They may not even know if this session is to make, make money and does that degrade their accuracy or is that some little thing that makes it so that it works and can we share that with them after the fact with their consent or something? Like that those are things that we'll try.

Jeffrey Mishlove (31:54)

If one accepts the premise that remote viewing is real, it's very hard then to say that the viewer has no access to information about what you're doing behind the scenes.

Chase McCarty (32:07)

That's true.

Jeffrey Mishlove (32:07)

That's true.

Chase McCarty (32:08)

Yeah. I wonder if we'll see that in the sessions as well. Yeah, but I mean like from our blockchain sessions we know that like we got statistically significant results from the precognitive sessions. So like it should work. But it would be really funny to analyze and see if they, if they could tell that some of them were ARV sessions without us telling them or some of them were going to be used for bets and some weren't or something like that. That would be very interesting.

Jeffrey Mishlove (32:31)

Yeah, well, one of the big issues today in parapsychology is the fact that in general the massive tests, working with thousands or hundreds of thousands of research subjects in an ESP type of test, the scores tend to be very, very close to non significance. And so remote viewing seemed to be a way to challenge that premise because each session can take more time. You could spend half a day doing a single trial in remote viewing and people often do. I don't know how much time your users spend, but that would be an interesting statistic to keep track of.

Chase McCarty (33:17)

Yeah, yeah, we do keep track of the data. We don't have it visible anywhere and I've only looked at it with my eyes, not with some statistical analysis. But a large percentage of what I've seen their session time is between 0 and 10 minutes kind of thing. But we certainly have users who spend days on a target. I don't know how much time they're actually spending viewing because we capture the time we task them and the time they submitted. So they might have just gone off and not thinking about. Would be nice to give some option for them to kind of estimate how much of that time they spent on it. But yeah, you're right, that makes remote viewing a great type of ESP experiment. And I should add, like when I was doing, doing this research initially, remote viewing is great, but I kind of just wanted to know is the ESP thing in general real? So I also looked at the gateway process and out of body experiments. I read the near death experience research, things like that. But it's really hard to plan to have a near death experience. So remote viewing becomes that thing that's really easy to study and really easy to study on a platform like this over the Internet where our users don't need anything Set up really outside of this app and we can still validate that their data is blind and the controls are in place. Whereas some other things like for example, like mindsight is something people talk about and we could build an app for mine site but we couldn't verify that users were wearing their blindfold during their session or something like that as easily.

Jeffrey Mishlove (34:48)

Your website connects to a number of social media platforms so people have an opportunity to engage in conversations with other users on those platforms where you have a presence. But I didn't find a forum explicitly on your website. I guess your idea is that the forums will exist elsewhere.

Chase McCarty (35:13)

Yeah, that's a great feedback point, Jeff. We should probably add something more like that we have yet. When I came to the space, I was a really big fan of the remote viewing subreddit and the Discord server associated with it. That's where I got a lot of my initial questions answered. And in fact when we did our first paid competition to collect sessions that was in collaboration with them. And so my thinking initially was I don't want to create an additional community to like fragment people from the previous communities. I'd rather shuttle them back and grow those additional communities. And so that's part of the thinking there. We probably will eventually have a built in forum, but right now I'd recommend people probably join the remote viewing subreddit Discord where we're somewhat active.

Jeffrey Mishlove (36:01)

You gave out a prize for many weeks of $1,000 a week as I recall, to the best remote viewing session.

Chase McCarty (36:09)

Yeah, no, that's right. We're very fortunate that we became a contact with DMT Quest and John Chavez who I think you've also had on here. So they, they've been funding some of our work, they funded our prize to help us collect more sessions. So we were giving away $1,000 a week for, I think it was six weeks earlier this year for like the, the best sessions submitted to the platform. Best being, you know, who could have the most sessions in a week that got the best score from our AI judging system. And that was great. You know, a lot of people had a lot of fun with that. It was exciting to be able to give out some money to these people. And then we also were able to collect like a lot of sessions that really helped us get to our 5,000 session number. And you know, we have more funding to do more of those sorts of experiments. So we'll also be doing additional types of competitions in the future. We learned a lot from this first competition and so the next one will be a little bit different. Tried to close some of the gaps, some of the feedback that people brought up from the previous one. But if you're interested in trying to win some money remote viewing, make sure to sign up on social-rv.com and we'll be sending out an email to all of the users whenever the next competition is starting. We're hoping to do one later this year or early next year.

Jeffrey Mishlove (37:26)

Well, Chase, it's been a real pleasure to talk to you and to share your innovative work with the New Thinking Allowed audience. Thank you so much for being with me today.

Chase McCarty (37:37)

Yeah, thank you so much for having me. I really appreciate it. It's a great conversation.

Jeffrey Mishlove (37:40)

It's been a pleasure. I'll be following your work closely and hope to have you back on New Thinking Allowed again in the future.

Chase McCarty (37:48)

I'd love that. Yeah, we'll have more stuff to share then, I'm sure too.

Jeffrey Mishlove (37:51)

Thank you. And for those of you watching or listening, thank you for being with us because you are the reason that we are here.

Announcer (38:24)

For early access to our videos and livestream events, sign up for our free weekly newsletter@newthinkingallowed.org.

Promotional Voice (38:38)

Book 4 in the New Thinking Allowed dialogue series is Charles T. Tart 70 years of exploring Consciousness and Parapsychology. Now available on Amazon.

Announcer (38:52)

New Thinking Allowed is presented by the California Institute for Human Science, a fully accredited university offering distance learning graduate degrees that focus on mind, body and spirit. The topics that we cover here. We are particularly excited to announce new degrees emphasizing parapsychology and the paranormal. Visit their website@cihs.edu. you can now download all eight copies of the New Thinking Allowed magazine for free or order beautiful printed copies. Go to newthinkingallowed.org.