Flaawsome Talk with Kjersti Flaa - I Just BLEW UP Blake's Case !!

00:00:00.000 So we are about to expose something pretty major.

00:00:08.340 Hi and welcome to Blossom Talk. I'm Jashti Floor, journalist, Hollywood truth teller,

00:00:12.600 and your voice of reason in a town built on delusion. I have Magnus here to help me again.

00:00:18.780 So we are about to expose something pretty major. I've spent basically an entire week

00:00:24.820 in doing research on this and Magnus has helped me and I think we're about to blow up Blake's

00:00:29.680 case. What do you think? Yes, this can be really, really bad for Blake. We have used a couple of

00:00:35.260 AIs to go through her expert testimony. And I would say she might have blown $50,000 to $150,000

00:00:44.520 for absolutely nothing. And on top of that, she actually gave me a reason to look into the

00:00:51.200 analytics of my own video. So keep in mind, I am the only person here who has access to the

00:00:58.380 analytics of my little bump interview and this expert that she hired his name is Alan his name

00:01:04.880 is Aaron Culotta he claims that my video went viral because it was amplified he has some theories

00:01:14.520 here that we're going to go through and basically show my graphs of what really happened and compare

00:01:21.820 them to what Blake Lively's expert is claiming. It's pretty crazy, Magnus, what we found out.

00:01:28.620 Yeah, this is not good for her team. And Aaron Collada, as you mentioned, he is a very...

00:01:34.400 Yeah, he's not nobody, that's for sure. He's got all the right credentials. He has done

00:01:38.600 over 80 peer-perviewed publications. He's gotten 5 million in research funding and his specific

00:01:47.640 expertise is in social network analyzes and inauthentic online activity so he he would be

00:01:53.940 perfect for doing this the only thing is he should have called you and asked for this data

00:01:58.700 yeah because now i'm about to make it but then again then he wouldn't have fit the narrative

00:02:02.020 would it no but the problem he has now is because i actually have the access that he didn't we're 0.66

00:02:08.620 about to show you what's a pile of crap this report is it's nonsense it's really interesting

00:02:16.640 So here is Aaron Colota's culotta, his theory on what happened to my little bump video. 0.98

00:02:24.660 We can call him Mr. C.

00:02:26.000 Mr. C.

00:02:27.580 Mr. C said that according to his research, my video had one comment on top that got a lot of likes.

00:02:39.180 and because of that he is saying that it went viral and all the activities surrounding it was

00:02:47.940 inauthentic and and that someone is behind getting all those likes so someone was basically saying

00:02:55.280 someone used bots to like this video so we've got all of the how many likes did it get i think it

00:03:00.620 was 73 000 likes on the top comment which of course is a lot it's not that much though is it

00:03:06.500 Well, on the comments, I guess it is.

00:03:08.080 But on the video, it wouldn't have been that much.

00:03:11.540 But the interesting part here is that he is saying that basically Jed Wallace,

00:03:17.820 this hired gun that Justin Baldoni's team hired to help him with crisis management,

00:03:23.520 that he is behind this.

00:03:25.700 He was clicking.

00:03:26.820 Like, like, like, like, like, like, like, like, like, like, like, like, like. 0.74

00:03:28.220 And the really stupid part of that is, is that you don't make a video go viral on YouTube by clicking a lot of, or liking a lot on the top comments. 0.64

00:03:40.240 Of course not, because if it was that easy, everybody would do it. 0.97

00:03:43.920 And why wouldn't Jed Wallace, if he was so clever and such a hired gun here, why wouldn't he like the video more?

00:03:52.000 Because that would actually help a video go viral.

00:03:54.820 clicking a lot of likes on a comment has no effect on the algorithms of YouTube

00:04:01.020 and you also did well you can show you can show this instead of you know we

00:04:04.600 got we got the numbers we got the curves we got the graphs and and also what

00:04:09.000 Kulata is claiming here he's saying that after Jed Wallace was given this

00:04:13.660 video say hey we should send this to Jed there is a text message saying that

00:04:18.500 after he did that the the view spiked and the comment spikes and it didn't so jed made it go

00:04:28.140 viral yeah which is not true anyways so in this report he's saying that he's looking at three

00:04:34.560 things he looked at tiktok videos he looked at reddit he looked at the way that the comments

00:04:40.300 were uh pushed up or pushed down and he also focused a lot on my little bump video he looked

00:04:47.720 at inauthentic activities on all of this what he calls inauthentic and he's saying here in his

00:04:54.080 report the little bump interview received few views and little commentary in the initial days

00:04:59.900 after it was uploaded on youtube and this was august 10th right yes he's talking about that's

00:05:05.220 when it was uploaded yeah and august 14th that's when it was sent to jed wallace he says this

00:05:11.740 changed after tag PR team suggested that the content should be sent to Jed. In the hours that

00:05:18.160 followed, comments on the Little Bump interview increased rapidly. So Jed Wallace ran to his

00:05:25.700 bot factory. He ran to his bot factory. And started writing comments. Click, click, click, click, click.

00:05:30.560 That could have happened, obviously. I'm not saying that that's impossible. But it is impossible to

00:05:36.340 make the video go viral by clicking on a like on a comment yes that isn't that is impossible but

00:05:42.060 the thing is this expert knew or he knows that all the other things that surrounding this video

00:05:49.420 is organic so he had to come up with some theory to make it work for Blake's narrative so that's

00:05:56.560 why he said the top comments have more like likes than normal so we looked into this in AI and said

00:06:03.120 you know, is this normal on a viral video? And I showed my graph of what happened to the video

00:06:08.780 when it started going viral. And what did you find, Magnus? Is it normal? It showed that this is

00:06:15.520 exactly textbook for what happens when a video go viral. Yeah. It's organic growth. So if my video

00:06:26.040 was not growing organically if someone was behind it and pushed ads on it because because that's

00:06:33.440 basically what they're saying they're saying that they amplify the video and we asked different ai

00:06:38.920 systems if this is possible and ai said that if you push a video you will have spikes like this

00:06:47.860 in the views of the video yeah it would be the curve wouldn't be like a flap curve it would go

00:06:52.900 go up and down a lot. Exactly. My video went like this. And then it plateaued. Yes, which is a

00:06:59.860 totally normal way for a video that goes viral. That curve is just complete textbook of a viral

00:07:07.980 video. Okay, so let's look at the comments that he said grew so fast on my video after it was sent

00:07:16.400 to Jed Wallace. And this is in his expert report, right? This is his expert report. It says,

00:07:21.260 comments on the little bump interview on YouTube from August 10th to 19th.

00:07:25.520 And as you can see, his curve goes like this.

00:07:30.680 And he's pointing out that that's exactly when it was sent to Jed Wallace.

00:07:35.780 There's a timeline there that says when the video was sent to Jed.

00:07:38.780 That's right.

00:07:39.300 Exactly.

00:07:39.940 And then, Magnus, you can look at my curve.

00:07:42.580 That you got from YouTube.

00:07:44.460 Yes, from my analytics, which shows that it actually went down a little bit from 2009.

00:07:50.520 from August 14th to August 15th.

00:07:54.000 You see that?

00:07:54.620 And August 14th is when he got it.

00:07:56.860 And then it went further down again,

00:07:58.960 as you can see on August 17th here, right?

00:08:01.600 Yeah.

00:08:02.100 So how does he find that the curve

00:08:05.300 is going straight up to the sky?

00:08:08.300 I don't know.

00:08:09.220 It's so interesting.

00:08:10.800 He has his own analytics.

00:08:12.260 And YouTube is lying, I guess.

00:08:13.800 I know.

00:08:14.500 And here's another interesting thing.

00:08:16.120 When I go into my analytics,

00:08:18.040 I see where the traffic is coming from right so I can see here that YouTube search is the top

00:08:26.160 people who found my video found it through YouTube search and then the browse feature

00:08:32.400 and then external which means that they found it through uh it was a maybe um embedded on some

00:08:40.280 websites I know a lot of outlets wrote about my video BuzzFeed had it early right BuzzFeed had it

00:08:44.880 early daily mail did it it was in a lot and also from reddit obviously so that's external and then

00:08:51.600 it was suggested videos that other suggested videos on my other videos and direct or unknown

00:08:57.620 is very low compared to all of that but that doesn't show any ads being used on my video and

00:09:04.760 if someone had amplified it without me knowing it it would show up here from where the traffic was

00:09:11.020 coming from you will see the spike and you would also see spikes in the in the in the curve and

00:09:18.720 the curve wouldn't go like this it would go like that and that's what he's trying to show here

00:09:24.340 but i don't know how he's going to explain this stuff because now that we actually know what

00:09:30.540 happened here we have the numbers we have the numbers real numbers when i fed this into

00:09:35.160 Claude and Chad

00:09:36.920 it came back quite

00:09:38.880 brutal

00:09:39.880 I can read

00:09:43.340 what it said here

00:09:44.400 Kolada attributes the spike

00:09:47.300 in bully related comments

00:09:49.460 to the Jed's suggestion

00:09:50.700 but the far simpler explanation is that

00:09:53.360 people watched the video, saw what happened

00:09:55.360 and independently reached the same conclusion

00:09:57.700 the content itself

00:09:59.200 generated the narrative

00:10:00.420 so he's saying that you could also

00:10:03.340 see the spike in the word bully being used in because of the spike of comments but of course

00:10:09.720 that was because people were reacting to the video so again he is not mentioning in this report at all

00:10:16.200 what was actually in the video and that people reacted to the content i yeah i want to i want

00:10:22.720 to read there's another thing from ai here um people didn't need to be manipulated into thinking

00:10:29.380 Blake Lively was rude in that interview. They had eyes. Shasta's own discomfort is visible on screen.

00:10:35.000 The public reaction may have been exactly what it appears. Genuine. Yeah. Yes. So what I also did

00:10:42.480 in my little research here this weekend, I took my viral video and then I compared it to one of

00:10:49.620 my videos from last week to see if it looked unhealthy or not. That was quite interesting.

00:10:53.880 yeah so my little bump video had 7.4 million views today and the video from my last week has

00:11:01.160 128 000 views the one you posted on thursday on thursday yes then the video from my little bump

00:11:10.320 has 34 000 comments comments and my video from thursday has 925 comments viral video has 121

00:11:20.660 thousand likes and my video from Thursday has seven point seven thousand likes and the top

00:11:27.860 comment on my video from Thursday has four hundred and forty five hundred and forty seven likes and

00:11:33.680 the top comment on the viral video has seventy three thousand likes so percentage wise here's

00:11:38.960 the here's the this is the interesting thing my viral video has generated 1.6 percent likes

00:11:47.480 from how many people watched it which is really low compared to my video on

00:11:53.120 Thursday which had over 5% likes yeah likes likes on the video which means

00:11:59.660 that is Jed still at work with a bigger army so he did a better work so he did a

00:12:06.080 better job on my video on Thursday then because that 5% likes yeah then the

00:12:11.040 viral video which only had 1.6% likes that's a huge difference and that is

00:12:16.700 what drives the algorithms as well so what drives the algorithms on youtube is likes engagement on

00:12:23.660 the video and also watch time on the video so how long people are actually watching it and if you're

00:12:29.380 using bots or whatever to push something if the content is not interesting people will not stay

00:12:34.580 and watch it and they will not react to it so his whole argument is just so embarrassing and

00:12:41.200 especially looking at all these graphs that he started using.

00:12:45.720 And also...

00:12:46.320 They're so wrong, especially that one about how it took off.

00:12:50.040 Yeah.

00:12:50.500 So what AI is saying here about viral videos is that it starts slow

00:12:55.220 and then the algorithms on YouTube starts testing it.

00:12:59.160 And then suddenly it takes off.

00:13:02.500 And that's exactly what the system does.

00:13:04.640 And that's exactly what happened to my video.

00:13:07.060 He also writes about...

00:13:08.940 This is really funny because he also writes in his report about all the anti-Blake comments that were on top of all the posts on social media that they were kept being voted up on Reddit and TikTok.

00:13:23.080 And then he doesn't really know if there's an upvote and downvote on TikTok.

00:13:26.860 That's how much he knows about social media, by the way.

00:13:28.760 He says that in the deposition.

00:13:30.300 But he talks about how these comments, that this was so suspicious, that there were so many anti-Blake comments that were top rated.

00:13:40.940 And again, it's not weird that people react to something.

00:13:46.020 And then he's like, well, the least popular comments were the ones supporting Blake.

00:13:51.140 Yeah, if people feel that you were bullied, of course they're going to like those comments, right?

00:13:56.000 They're going to support those comments.

00:13:57.260 He, it's like he doesn't take into consideration that people have like feelings.

00:14:03.020 Or the content of what he just watched.

00:14:04.840 Yeah, and that you're not reacting to anything because he's looking basically at human beings as bots.

00:14:11.800 It's all Jed's making.

00:14:13.340 Yeah, and he's saying that people who reacted to it, you couldn't react like that.

00:14:18.060 Because that's not normal, that you react and have strong feelings about a celebrity who behaves badly in an interview.

00:14:26.060 Also, another thing that I found out, which also AI taught us, is that when you have a viral video, a lot of people wouldn't even bother to comment because they see that, oh, it's already 34,000 comments on this video.

00:14:41.200 My comment is just going to disappear into the, you know, there's so much.

00:14:45.960 But if they agree with the top comment, they're going to like that.

00:14:50.180 And I also believe that some people didn't like the video itself because they felt, oh, this is showing something that I feel uncomfortable with.

00:14:58.620 This is showing a person who is really being mean to another person.

00:15:03.800 Should I like that video?

00:15:04.980 Does that mean that I actually like?

00:15:06.960 I'm supporting that behavior.

00:15:08.400 Yeah, exactly.

00:15:09.380 So instead of liking the video itself, it would be natural for people to react this way and actually liking the top comments.

00:15:16.040 That's what both Claude and ChatGPT told us.

00:15:20.180 So what he said in his report is that the content of these highest ranked comments is almost all anti-Lively or pro-Baldoni.

00:15:30.120 In particular, there are multiple comments that cast doubts on the claim that Mr. Baldoni fat shamed Miss Lively

00:15:37.000 and that Miss Lively's response to Miss Floor Little Bump question was unacceptable,

00:15:42.740 including comments referring to her as a bully, mean girl and rude or referred to Miss Lively's plantation wedding.

00:15:49.160 so he's basically just explaining how people reacted and he's saying like this is how people 0.91

00:15:54.380 reacted and that's why we don't think it's real people i mean it's so ridiculous it's interesting

00:16:00.860 but i feel like now where i can actually show what my findings are and i am the only person

00:16:08.440 as i said who has access to these analytics this guy didn't and to see what he found compared to

00:16:16.340 what I actually found in my videos that's pretty damning yeah I'm sure he's not happy about it no

00:16:23.060 and I'm sure she's not happy about it Blake because as you said she paid between 50 to 150

00:16:29.040 thousand dollars for this expert who says that hey I have the evidence here this video didn't

00:16:37.040 go viral organically well I have something to tell you Mr. C we know that is not true because

00:16:44.360 I have all the proof.

00:16:46.500 Okay, I wanted to end this video with some stuff from his deposition

00:16:49.900 because I found this quite amusing. 0.81

00:16:51.460 I love when someone is just caught in being ridiculous, and he is. 0.65

00:16:55.680 So Baldoni's attorney or the Wayfair Party's attorney is asking him, 0.90

00:16:58.700 when you're using the word inauthentic,

00:17:01.220 are you including boosting comments that are already on the internet

00:17:04.680 as opposed to placing comments on the internet?

00:17:08.160 I think that's a very valid question.

00:17:10.520 Is he saying that they actually made up comments

00:17:13.240 negative things about Blake or that they just pushed likes on things that were negative about

00:17:18.680 her and he's saying I would consider both of those as types of inauthentic activities yes

00:17:23.680 have you identified any instance where somebody added false content on Reddit TikTok or YouTube

00:17:29.420 he's asking um and the expert says my analysis weren't really focused on the the truth or or

00:17:37.700 falseness of different statements i love that he says my analysis wasn't really focused on the truth

00:17:43.840 yeah or falseness of different statements yeah but just that he's saying my report wasn't really

00:17:49.120 focused on the truth no it wasn't it was just focused on giving blake lively the narrative

00:17:54.560 that she wants and he continues to say i analyze patterns in engagement metrics on posts and

00:18:00.620 comments specifically i calculated the proportion of the number of likes on top comments versus the

00:18:06.480 number of likes on each video, which we already spoke about here. This was his focus. He's

00:18:12.100 admitting here that was his focus. And as we now know, that has nothing to do with a video going

00:18:17.740 viral or not. Oops. Yeah. And the attorney is asking, first of all, what do you mean by

00:18:22.460 manipulation? And he says, in this case, the manipulation would be an effort to boost certain

00:18:28.760 comments attached to specific videos. And then this part is interesting. Okay, but there are

00:18:34.640 lots of different engagement metrics i'm asking are you aware of any study that says the ratio of

00:18:40.480 likes on a top comment to likes on a parent post is indicative of manipulation basically if you can

00:18:48.420 see if a video is manipulated because there are more likes on the top comments that you think is

00:18:54.980 normal and he's saying i'd have to double check but again i would say this is this is a this is

00:19:03.700 This paper is following a very similar mythology for YouTube.

00:19:09.900 And then the attorney is asking, other than this paper that you cited, can you cite to any other peer-reviewed article study that would indicate that likes on a top comment in proportion to likes on a parent post would be indicative of manipulation?

00:19:27.200 And he's saying, again, I can't recall right now on that is that specific.

00:19:32.620 I would say the earlier site to Young is a similar type of comparison on the Twitter platform.

00:19:39.440 I think Young is another expert, and he also gave them a lot of articles and papers, kind of, you know, trying to justify all this.

00:19:50.280 I can't believe you.

00:19:51.100 I can't believe this.

00:19:51.860 What a giant waste of time.

00:19:53.920 They could have subpoenaed your YouTube analytics, couldn't they?

00:19:57.580 No, I guess they could try.

00:19:59.820 I mean, that's what they tried to do.

00:20:01.320 but now that's too late magnus why do you come with suggestions like that

00:20:05.100 well now they have it they don't have to come after you now they don't want my they don't want

00:20:10.820 my analytics of course not because that will show them that this was organic so of course they

00:20:15.420 wouldn't do that and the fact that they're fighting you to show up in court that they

00:20:18.680 don't want you there it just shows how weak the the argument is because this your video is the

00:20:24.220 centerpiece of this smear case basically right they're saying they used your video

00:20:29.820 to make you look like a bully and a mean girl yeah exactly yeah as the base for the smear

00:20:35.620 campaign right but they don't want you there to talk about this and they and they don't want

00:20:39.040 the video to be shown if this is the centerpiece in the evidence in the line of evidence why

00:20:45.460 wouldn't you they both want you and the video there yeah good question well we know the answer

00:20:51.200 and then he's saying do you have any specific expertise in analyzing whether comments are pro

00:20:56.680 or anti in any way and does he he says well i mean so i've been doing research in social media

00:21:03.440 since around 2009 and so i spent a lot of time staring at posts and comments and trying to

00:21:09.840 contextualize them and understand them including for sentiment analysis so i would say i've done

00:21:15.540 quite a bit of this type and then the attorney's asking i see is there some kind of protocol that

00:21:21.020 you use to decide what's pro-lively or what's anti-lively pro-baldoni or anti-baldoni is there

00:21:26.860 sort of your gut feeling and then he says I wouldn't characterize it as a gut feeling I would

00:21:32.220 say you know you typically have a kind of the point of the the secondary review is I mean yeah

00:21:38.200 blah blah blah he can't answer that he can't really answer any of these things because he

00:21:43.140 doesn't have a good answer because he knows he was hired to do this to make you know find a specific

00:21:49.160 end results yeah and now we have to say to you mr c that uh you're wrong maybe take a second

00:21:57.840 look at the curves again yes and the numbers yes and you know maybe use ai next time

00:22:03.100 yes let us know what you think do you think that mr c was onto something or do you call his bs like

00:22:13.140 we just did and yes I wonder if these graphs are going to be shown in court

00:22:19.240 now Magnus maybe it could help the wafer parties quite a lot I think it should

00:22:23.700 yes anyways that's it for now I hope you enjoy this video thank you again to

00:22:29.580 Magnus for being here helping me out and of course if you haven't subscribed yet

00:22:33.780 please do hit the notification bell and all that good stuff and like share

00:22:37.820 share, comment, hype, and have a great Sunday.

00:22:41.200 Bye.

00:22:42.200 Bye.

Flaawsome Talk with Kjersti Flaa - April 27, 2026

I Just BLEW UP Blake's Case !!

Episode Stats

Summary

Transcript