WEBVTT

1
00:00:00.320 --> 00:00:03.200
<v Speaker 1>Hey, Carl here, we're on a mission to find new

2
00:00:03.240 --> 00:00:05.919
<v Speaker 1>fans of the show. I mean, where else can you

3
00:00:05.960 --> 00:00:10.240
<v Speaker 1>get real life comedy and horror in one podcast. You

4
00:00:10.279 --> 00:00:13.000
<v Speaker 1>can do your part by leaving us a review on iTunes,

5
00:00:13.039 --> 00:00:16.719
<v Speaker 1>Spotify or wherever you get Security this week. Also, you

6
00:00:16.719 --> 00:00:19.079
<v Speaker 1>can get an ad free feed by becoming a five

7
00:00:19.120 --> 00:00:22.039
<v Speaker 1>dollars a month patron. Go to Patreon dot Security this

8
00:00:22.079 --> 00:00:25.320
<v Speaker 1>week dot com and sign up, and don't forget discord.

9
00:00:25.600 --> 00:00:28.519
<v Speaker 1>Discord dot Security this week dot com. Lots of good

10
00:00:28.559 --> 00:00:32.280
<v Speaker 1>stuff happening there. Yeah, I think that's it. So on

11
00:00:32.359 --> 00:00:33.200
<v Speaker 1>with the podcast.

12
00:00:33.759 --> 00:00:36.799
<v Speaker 2>You know, guys, this whole thing about equality of the sexes,

13
00:00:36.880 --> 00:00:38.079
<v Speaker 2>it's just bunk.

14
00:00:39.640 --> 00:00:43.280
<v Speaker 3>Said, you're hate mail to Carl Franklin, both of our

15
00:00:43.320 --> 00:00:44.159
<v Speaker 3>female listeners.

16
00:00:44.359 --> 00:00:48.640
<v Speaker 2>Here's an example. Miley Cyrus dance is naked and licks

17
00:00:48.640 --> 00:00:51.759
<v Speaker 2>a hammer and everybody says, oh, it's art. You know

18
00:00:51.840 --> 00:00:54.640
<v Speaker 2>when I do that, the manager calls the cops at

19
00:00:54.640 --> 00:01:06.120
<v Speaker 2>home depot. Hey, welcome back to Security this week. I'm

20
00:01:06.159 --> 00:01:09.599
<v Speaker 2>Carl Franklin's Patrick Hines and Dwaine Laflatte. You got some stories,

21
00:01:09.599 --> 00:01:13.079
<v Speaker 2>Oh boy, do we have some stories. But I gotta

22
00:01:13.120 --> 00:01:17.040
<v Speaker 2>say we have a good percentage of good stories. Sure, yeah, absolutely,

23
00:01:17.120 --> 00:01:20.640
<v Speaker 2>let's start with one that's not so good. In Gadget reports,

24
00:01:20.680 --> 00:01:25.319
<v Speaker 2>hacker used anthropics Claude chat bot to attack multiple government

25
00:01:25.400 --> 00:01:27.280
<v Speaker 2>agencies in Mexico.

26
00:01:27.599 --> 00:01:33.519
<v Speaker 4>What happened, Oh no, so we've heard this. So what

27
00:01:33.640 --> 00:01:37.719
<v Speaker 4>happened is a lot of these lms, especially when you

28
00:01:37.719 --> 00:01:41.319
<v Speaker 4>put tooling on them, give them agency. They can run things, right,

29
00:01:41.319 --> 00:01:43.280
<v Speaker 4>they can run in map, they can go scan networks,

30
00:01:43.319 --> 00:01:45.480
<v Speaker 4>they can do whatever. But a lot of them, a

31
00:01:45.480 --> 00:01:47.560
<v Speaker 4>lot of the online ones you're chat Gypts and whatever,

32
00:01:47.599 --> 00:01:49.840
<v Speaker 4>they're not going to allow you to break into a site.

33
00:01:49.879 --> 00:01:52.079
<v Speaker 4>They're not going to allow you to you know, learn

34
00:01:52.079 --> 00:01:53.239
<v Speaker 4>how to make a bomb whatever.

35
00:01:53.280 --> 00:01:54.200
<v Speaker 5>They have guardrails.

36
00:01:54.359 --> 00:01:59.000
<v Speaker 4>Yeah, what this particular attacker did is obviously learned the

37
00:01:59.040 --> 00:02:04.120
<v Speaker 4>ways around the guardrails. And sometimes that's downloading the language

38
00:02:04.120 --> 00:02:06.920
<v Speaker 4>model locally and using a local version like in l

39
00:02:07.120 --> 00:02:10.800
<v Speaker 4>M Studio or something along those lines, or you know openclaw,

40
00:02:10.879 --> 00:02:14.879
<v Speaker 4>which we'll talk about later. So there are ways for

41
00:02:14.919 --> 00:02:18.479
<v Speaker 4>you to use artificial intelligence locally that actually has tools

42
00:02:18.479 --> 00:02:20.719
<v Speaker 4>that can do things, that can go out and touch things.

43
00:02:20.759 --> 00:02:23.960
<v Speaker 4>And in this case, they decided to automate as much

44
00:02:24.039 --> 00:02:27.000
<v Speaker 4>of an attack chain as they could and we've seen

45
00:02:27.080 --> 00:02:30.919
<v Speaker 4>this anthropic says they stopped an attack from China where

46
00:02:31.000 --> 00:02:36.520
<v Speaker 4>China was actually using Anthropic to target organizations. This person

47
00:02:36.719 --> 00:02:39.240
<v Speaker 4>just got much further. We've got one hundred and fifty

48
00:02:39.280 --> 00:02:42.439
<v Speaker 4>three gig worth of sensitive data, tax records, and all

49
00:02:42.479 --> 00:02:43.680
<v Speaker 4>sorts of stuff from Mexico.

50
00:02:43.879 --> 00:02:47.280
<v Speaker 3>So, I mean, this is an example of everything can

51
00:02:47.319 --> 00:02:51.319
<v Speaker 3>be used for both good and evil. Nuclear power, nuclear weapons,

52
00:02:52.479 --> 00:02:57.719
<v Speaker 3>you know, biological research, cures for diseases, bio weapons. It's

53
00:02:57.759 --> 00:03:00.520
<v Speaker 3>true of everything. So the more we lock this stuff

54
00:03:00.560 --> 00:03:03.800
<v Speaker 3>down so that it can't hack, then what happens when

55
00:03:04.039 --> 00:03:06.439
<v Speaker 3>we want to have it evaluate our code to see

56
00:03:06.439 --> 00:03:09.159
<v Speaker 3>what vulnerabilities we can fix. Right, So there's this double

57
00:03:09.240 --> 00:03:11.560
<v Speaker 3>edged sword we have to figure out, and we haven't

58
00:03:11.560 --> 00:03:14.840
<v Speaker 3>figured it out yet. It's not we're not anywhere close there. Yeah,

59
00:03:14.879 --> 00:03:15.719
<v Speaker 3>it's moving faster.

60
00:03:15.879 --> 00:03:18.879
<v Speaker 4>Well, it's moving fast. I mean it's like, yeah, it's

61
00:03:18.919 --> 00:03:21.080
<v Speaker 4>like gunpowder, right, gunpowder could be used for good or

62
00:03:21.080 --> 00:03:23.199
<v Speaker 4>could be used for evil. But if you gave everybody

63
00:03:23.199 --> 00:03:26.080
<v Speaker 4>on the planet gunpowder the moment it was invented, but

64
00:03:26.199 --> 00:03:29.439
<v Speaker 4>we'd see a bunch of really crazy stuff until laws

65
00:03:29.479 --> 00:03:32.120
<v Speaker 4>and regulations in society came came forward and said you

66
00:03:32.120 --> 00:03:34.560
<v Speaker 4>can't do these things. And we're now starting to see

67
00:03:34.560 --> 00:03:36.560
<v Speaker 4>that with AI. It's moving so fast and it's in

68
00:03:36.560 --> 00:03:39.240
<v Speaker 4>the hands of everybody on the planet right anything they

69
00:03:39.240 --> 00:03:41.960
<v Speaker 4>can imagine they want to do at this point, you know,

70
00:03:42.080 --> 00:03:47.520
<v Speaker 4>like automatically targeting organizations and stealing data is now becoming

71
00:03:47.520 --> 00:03:50.000
<v Speaker 4>a possibility. And we talked about this, I think last night.

72
00:03:50.080 --> 00:03:52.240
<v Speaker 2>Yeah, look at a hammer. Look at a hammer. I

73
00:03:52.240 --> 00:03:53.479
<v Speaker 2>mean it's got dual uses.

74
00:03:55.280 --> 00:04:00.000
<v Speaker 4>We use wrenches, sorry, wrenches, wrenches.

75
00:04:00.159 --> 00:04:01.560
<v Speaker 5>The fight all a wrench, that's right.

76
00:04:03.280 --> 00:04:05.360
<v Speaker 3>So in this article they do I think they talk

77
00:04:05.680 --> 00:04:08.240
<v Speaker 3>down by the bottom about the fact that Anthropic has

78
00:04:08.400 --> 00:04:12.840
<v Speaker 3>changed their commitment to safety, and it's not the commitments

79
00:04:12.840 --> 00:04:14.960
<v Speaker 3>to safety. The military is trying to get them to rollback.

80
00:04:15.080 --> 00:04:17.720
<v Speaker 3>And I need to acknowledge a mistake I made in

81
00:04:17.759 --> 00:04:18.800
<v Speaker 3>the last episode.

82
00:04:18.839 --> 00:04:20.800
<v Speaker 4>I said that you don't make mistakes there I did.

83
00:04:20.839 --> 00:04:23.000
<v Speaker 3>I said, Pete Hay says never saw a war, and

84
00:04:23.040 --> 00:04:26.240
<v Speaker 3>he did actually deploy, and he did get two Bronze Stars,

85
00:04:26.240 --> 00:04:29.160
<v Speaker 3>but neither one of them was for valor. They're more meritorious.

86
00:04:29.240 --> 00:04:32.399
<v Speaker 3>So I still think he's not competent for the job.

87
00:04:32.560 --> 00:04:35.920
<v Speaker 3>But he did see war zones and get a cib

88
00:04:36.079 --> 00:04:37.040
<v Speaker 3>so I Will.

89
00:04:37.040 --> 00:04:39.000
<v Speaker 2>Was his dad, the guy that handed out the stars.

90
00:04:39.959 --> 00:04:41.920
<v Speaker 3>Well, I mean I got a bronze star for my

91
00:04:42.639 --> 00:04:46.839
<v Speaker 3>firefight in the first Golf War. It wasn't a valorous one.

92
00:04:46.920 --> 00:04:49.279
<v Speaker 3>It's not with valid I didn't charge a machine gun, nest.

93
00:04:49.800 --> 00:04:54.560
<v Speaker 3>He got one for professionalism and commitment to excellence to

94
00:04:54.600 --> 00:04:58.639
<v Speaker 3>help build a stable Iraq. No mention of any combat there.

95
00:04:58.759 --> 00:04:59.759
<v Speaker 5>That sounds like a gold star.

96
00:05:00.000 --> 00:05:02.319
<v Speaker 3>And then the next one was in Afghanistan. He was

97
00:05:02.319 --> 00:05:05.480
<v Speaker 3>a counter surgeon instructor and he acted in a dangerous situation,

98
00:05:05.560 --> 00:05:07.600
<v Speaker 3>which probably means he was just in range of something.

99
00:05:07.720 --> 00:05:08.079
<v Speaker 5>Got it.

100
00:05:08.879 --> 00:05:12.000
<v Speaker 3>But so I don't like to be wrong, so I

101
00:05:12.040 --> 00:05:13.680
<v Speaker 3>want to make that correction. But I still think he's

102
00:05:13.680 --> 00:05:21.360
<v Speaker 3>an idiot. So Patrick, bring it on, Man, bring it on.

103
00:05:22.279 --> 00:05:25.879
<v Speaker 3>So the Anthropic changed their commitment, and the commitment was

104
00:05:26.040 --> 00:05:28.800
<v Speaker 3>pretty out there in my opinion. It was like they

105
00:05:29.240 --> 00:05:31.759
<v Speaker 3>wanted to make sure they didn't build an AI they

106
00:05:31.759 --> 00:05:35.680
<v Speaker 3>couldn't prove was safe. Okay, again, like I just said,

107
00:05:35.680 --> 00:05:40.160
<v Speaker 3>this dual use technologies. Everything has a sinister side. And

108
00:05:40.199 --> 00:05:42.240
<v Speaker 3>they kind of came to the realization that it was

109
00:05:42.279 --> 00:05:44.800
<v Speaker 3>moving so fast that if they kept to this pledge,

110
00:05:44.839 --> 00:05:48.279
<v Speaker 3>they'd lose that did. And and right now, one of

111
00:05:48.279 --> 00:05:50.240
<v Speaker 3>the things that we've seen, and I digressed a little

112
00:05:50.240 --> 00:05:54.800
<v Speaker 3>bit is open AI went in and tried to scoop

113
00:05:55.000 --> 00:06:00.240
<v Speaker 3>the military contract from Anthropic and it it fail, bailed

114
00:06:00.279 --> 00:06:04.120
<v Speaker 3>disastrously because the backlash against them. I read that they

115
00:06:04.160 --> 00:06:07.079
<v Speaker 3>had a two hundred and ninety five percent increase in cancelations.

116
00:06:07.199 --> 00:06:10.000
<v Speaker 4>Oh yeah, like one point five million or something like that,

117
00:06:10.040 --> 00:06:11.519
<v Speaker 4>cancelations in service.

118
00:06:11.279 --> 00:06:15.120
<v Speaker 3>Which is probably worth more than the contract will be

119
00:06:15.240 --> 00:06:19.800
<v Speaker 3>worth that they're after. So Anthropic lost a contract probably

120
00:06:19.800 --> 00:06:23.839
<v Speaker 3>with the government, and is won many many more devoted fans.

121
00:06:24.720 --> 00:06:29.160
<v Speaker 3>And Open Ai has now shown that they'll basically do

122
00:06:29.240 --> 00:06:32.480
<v Speaker 3>anything for a buck. And so I think what's happening

123
00:06:32.519 --> 00:06:35.240
<v Speaker 3>is we could very well see this as the decision

124
00:06:35.279 --> 00:06:38.319
<v Speaker 3>point that decides which one ultimately wins. I'm not calling

125
00:06:38.360 --> 00:06:41.920
<v Speaker 3>the race yet, but I think this was actually a

126
00:06:41.920 --> 00:06:44.399
<v Speaker 3>good thing, for a good fight for Anthropic to pick.

127
00:06:44.519 --> 00:06:46.759
<v Speaker 2>You know, I saw a story today and I'll add

128
00:06:46.800 --> 00:06:50.199
<v Speaker 2>it to the show. Notes that Anthropic CEO says company

129
00:06:50.279 --> 00:06:53.720
<v Speaker 2>no longer sure whether or not Claude is conscious.

130
00:06:54.160 --> 00:06:55.920
<v Speaker 3>Oh wow, Yeah, that's that's bs.

131
00:06:55.959 --> 00:06:59.120
<v Speaker 2>They can't say definitively whether it is conscious.

132
00:06:59.360 --> 00:07:02.319
<v Speaker 3>I believe that that is total BS. I think that

133
00:07:02.319 --> 00:07:07.759
<v Speaker 3>that is either he's being intentionally obtuse or he's an

134
00:07:07.759 --> 00:07:11.120
<v Speaker 3>idiot in that regard, because if you understand how large

135
00:07:11.160 --> 00:07:15.040
<v Speaker 3>language models work, they'll never they don't understand what they're saying,

136
00:07:15.040 --> 00:07:18.399
<v Speaker 3>they don't understand what they're what they're hearing. They'll never

137
00:07:18.480 --> 00:07:20.639
<v Speaker 3>be conscious. I'll put stake my life you.

138
00:07:21.240 --> 00:07:23.759
<v Speaker 2>I know people that fall into that category too, though, I.

139
00:07:23.800 --> 00:07:26.279
<v Speaker 3>Mean, and maybe they're not conscious. That's why I advocate

140
00:07:26.319 --> 00:07:31.759
<v Speaker 3>killing them.

141
00:07:30.639 --> 00:07:34.759
<v Speaker 2>Tom. If you think about and I know those of

142
00:07:34.839 --> 00:07:38.000
<v Speaker 2>us in the software world, we know how these things,

143
00:07:38.079 --> 00:07:39.759
<v Speaker 2>what these things are made of, and they're made of

144
00:07:39.800 --> 00:07:44.720
<v Speaker 2>algorithms and software, but you know, at the if you

145
00:07:44.759 --> 00:07:47.759
<v Speaker 2>break down what the brain does and how thoughts work

146
00:07:47.800 --> 00:07:50.759
<v Speaker 2>and how words are put together, they are kind of

147
00:07:51.399 --> 00:07:54.279
<v Speaker 2>something like when I when I think about what to say,

148
00:07:54.399 --> 00:07:58.000
<v Speaker 2>sometimes my synapse is misfired and it comes out to statistical.

149
00:07:58.199 --> 00:08:01.879
<v Speaker 3>I do not count the fact that there will someday

150
00:08:01.959 --> 00:08:04.959
<v Speaker 3>be a computer that is conscious and that has consciousness.

151
00:08:05.079 --> 00:08:07.120
<v Speaker 3>I don't disagree with that at all. I do not

152
00:08:07.199 --> 00:08:10.040
<v Speaker 3>believe lllms will ever lead to that. I believe it

153
00:08:10.079 --> 00:08:13.120
<v Speaker 3>will be another jump and it could be a much

154
00:08:13.160 --> 00:08:15.120
<v Speaker 3>further jump. So yeah, we can debate it.

155
00:08:15.199 --> 00:08:16.720
<v Speaker 2>This is just the language part of them.

156
00:08:16.800 --> 00:08:20.480
<v Speaker 3>But also remember one thing, all these Silicon Valley CEOs,

157
00:08:21.319 --> 00:08:24.279
<v Speaker 3>they are raising billions and tens of billions and even

158
00:08:24.399 --> 00:08:28.040
<v Speaker 3>hundreds of billions on the back of this hype for AGI.

159
00:08:28.279 --> 00:08:30.800
<v Speaker 3>So there's an old story that says, an old saying

160
00:08:30.839 --> 00:08:33.960
<v Speaker 3>that says, you can never convince a man to believe

161
00:08:34.159 --> 00:08:37.039
<v Speaker 3>that which his paycheck depends on him not believing.

162
00:08:37.759 --> 00:08:41.320
<v Speaker 2>Right, All right, let's move on. Hacker new says clawjacked,

163
00:08:41.399 --> 00:08:46.120
<v Speaker 2>flaw lets, malicious sites hijack local open claw AI agents

164
00:08:46.679 --> 00:08:50.080
<v Speaker 2>via web socket. Man, here's what I know about web sockets.

165
00:08:50.559 --> 00:08:55.159
<v Speaker 2>There's like sockets that you connect to with a yeah yeah,

166
00:08:55.519 --> 00:08:57.919
<v Speaker 2>that you connect to at an IP address. But they're

167
00:08:57.960 --> 00:09:00.639
<v Speaker 2>all over port eighty so they go right through firewalls.

168
00:09:01.559 --> 00:09:02.960
<v Speaker 2>So what what is this all about?

169
00:09:03.320 --> 00:09:05.879
<v Speaker 4>So this this is interesting. Poor open claw has been

170
00:09:06.000 --> 00:09:10.960
<v Speaker 4>in the news constantly.

171
00:09:11.039 --> 00:09:12.200
<v Speaker 2>Yeah, I mean, that's great.

172
00:09:12.240 --> 00:09:15.679
<v Speaker 3>You misunderstood.

173
00:09:15.799 --> 00:09:18.840
<v Speaker 2>Wow, wow, he was a good guy. Yeah.

174
00:09:19.519 --> 00:09:23.879
<v Speaker 4>So what's happening is there's a local website. So when

175
00:09:23.879 --> 00:09:27.720
<v Speaker 4>you install open claw, there's a local management website where

176
00:09:27.799 --> 00:09:30.919
<v Speaker 4>obviously it's you can put a password on it. But

177
00:09:30.960 --> 00:09:32.840
<v Speaker 4>that's how you log in and you do things, right,

178
00:09:32.960 --> 00:09:35.480
<v Speaker 4>you you know, set up agents, trusted computers, all sorts

179
00:09:35.519 --> 00:09:38.200
<v Speaker 4>of other stuff, right. You know, that's your interface to

180
00:09:38.440 --> 00:09:44.799
<v Speaker 4>this locally running agentic AI. But it's only accessible on

181
00:09:44.840 --> 00:09:47.600
<v Speaker 4>what's called local host right, right, So for those of

182
00:09:47.600 --> 00:09:51.360
<v Speaker 4>you who aren't sort of networking, it didn't didn't grow

183
00:09:51.399 --> 00:09:54.960
<v Speaker 4>up in the networking world. Local host is your computer,

184
00:09:55.440 --> 00:09:57.759
<v Speaker 4>but it's a little bit more unique than that. The

185
00:09:57.799 --> 00:10:00.840
<v Speaker 4>address that resolves to local hosts. So every computer on

186
00:10:00.879 --> 00:10:04.120
<v Speaker 4>the Internet obviously has an IP address. In your home,

187
00:10:04.240 --> 00:10:07.559
<v Speaker 4>your computer has an IP dress that's called non routable. Yeah,

188
00:10:07.600 --> 00:10:08.200
<v Speaker 4>so you're not you.

189
00:10:08.200 --> 00:10:10.440
<v Speaker 2>Can't access it from outside your local network.

190
00:10:10.600 --> 00:10:13.320
<v Speaker 4>Yeah, I you know, my house might be one nine

191
00:10:13.360 --> 00:10:16.759
<v Speaker 4>six eight, you know, zero dot two, Go on, right,

192
00:10:16.840 --> 00:10:19.399
<v Speaker 4>I know, right, yeah, go ahead, find that IP address.

193
00:10:20.039 --> 00:10:22.200
<v Speaker 4>But so is ninety nine percent of the rest of

194
00:10:22.240 --> 00:10:25.799
<v Speaker 4>the planet. Right, Your home addresses are nonroutable. Their computers,

195
00:10:26.080 --> 00:10:27.679
<v Speaker 4>you know, can talk to each other that way, but

196
00:10:27.759 --> 00:10:29.639
<v Speaker 4>it's not like I'm connecting to you on that IP.

197
00:10:29.840 --> 00:10:32.240
<v Speaker 3>All routers on the Internet drop those p exactly.

198
00:10:32.600 --> 00:10:32.799
<v Speaker 5>Now.

199
00:10:32.840 --> 00:10:36.399
<v Speaker 4>On top of that, you have another address associated with

200
00:10:36.399 --> 00:10:38.840
<v Speaker 4>your computer called one two seven dot zero dot zero

201
00:10:38.840 --> 00:10:39.120
<v Speaker 4>dot one.

202
00:10:39.120 --> 00:10:40.039
<v Speaker 2>There's no place like home.

203
00:10:40.080 --> 00:10:43.960
<v Speaker 3>There's no place like one two seven zero zero one. Yes,

204
00:10:44.200 --> 00:10:47.679
<v Speaker 3>all right, it's a great nerd t shirt because that's

205
00:10:47.679 --> 00:10:48.200
<v Speaker 3>how right.

206
00:10:48.320 --> 00:10:49.039
<v Speaker 2>That's your home.

207
00:10:49.440 --> 00:10:53.000
<v Speaker 4>Yeah, it's your computer. If if anything, any software in

208
00:10:53.039 --> 00:10:56.879
<v Speaker 4>your computer has to talk to one two seven dot one,

209
00:10:57.000 --> 00:10:59.080
<v Speaker 4>but one two seven do zero dot zero dot one,

210
00:11:00.039 --> 00:11:04.399
<v Speaker 4>then it's talking to itself locally. Nobody sees that traffic.

211
00:11:04.759 --> 00:11:06.559
<v Speaker 4>It doesn't go out over the Internet, it doesn't go

212
00:11:06.559 --> 00:11:07.799
<v Speaker 4>out over your local network.

213
00:11:07.840 --> 00:11:08.759
<v Speaker 5>It just talks local.

214
00:11:08.919 --> 00:11:12.159
<v Speaker 3>So would that cause the average human to set up

215
00:11:12.200 --> 00:11:14.799
<v Speaker 3>really weak passwords since nobody can access it.

216
00:11:15.639 --> 00:11:17.960
<v Speaker 4>You should say that, Patrick, It's a great it's a

217
00:11:17.960 --> 00:11:21.240
<v Speaker 4>great point, Patrick. So let's say you set up your

218
00:11:21.279 --> 00:11:24.519
<v Speaker 4>open claw and you know it can only be accessed locally.

219
00:11:24.559 --> 00:11:26.759
<v Speaker 4>You are the only person who can open a browser

220
00:11:26.879 --> 00:11:30.360
<v Speaker 4>and view that configuration website. And let's say you set

221
00:11:30.360 --> 00:11:33.279
<v Speaker 4>the password to it to be password because you're like whatever,

222
00:11:33.519 --> 00:11:35.480
<v Speaker 4>or open claw, yeah, or open clause. It doesn't matter

223
00:11:35.519 --> 00:11:37.840
<v Speaker 4>because nobody can access it unless they're hands on keys

224
00:11:37.879 --> 00:11:38.519
<v Speaker 4>at your computer.

225
00:11:38.720 --> 00:11:39.399
<v Speaker 2>Not so fast.

226
00:11:40.600 --> 00:11:45.519
<v Speaker 4>Not so fast because communications happen over web sockets, right,

227
00:11:45.639 --> 00:11:51.720
<v Speaker 4>like Carl had said, And because JavaScript can open web sockets.

228
00:11:52.399 --> 00:11:56.879
<v Speaker 4>If you visit a malicious website, it will open a

229
00:11:57.000 --> 00:12:01.879
<v Speaker 4>web socket two one zero zero dot one for the

230
00:12:01.919 --> 00:12:04.840
<v Speaker 4>open claw web socket port, and it will brute force

231
00:12:04.879 --> 00:12:07.639
<v Speaker 4>your password and if it finds it, it will add

232
00:12:07.679 --> 00:12:10.879
<v Speaker 4>the hackers computers as a trusted device. Yeah, and this,

233
00:12:11.120 --> 00:12:12.759
<v Speaker 4>and then it will dump all the keys and do

234
00:12:12.799 --> 00:12:13.600
<v Speaker 4>all sorts of stuff.

235
00:12:13.639 --> 00:12:15.960
<v Speaker 3>So so in this way, the web socket's acting like

236
00:12:16.000 --> 00:12:22.759
<v Speaker 3>a shell exactly, exactly good good, I mean, play the music,

237
00:12:23.000 --> 00:12:24.200
<v Speaker 3>because for the play of the.

238
00:12:24.200 --> 00:12:26.440
<v Speaker 2>Music, guys, all right, go ahead, brand and play the music.

239
00:12:27.039 --> 00:12:33.759
<v Speaker 4>It's hey, if you need a shell that's not a shell,

240
00:12:33.960 --> 00:12:34.840
<v Speaker 4>use web sockets.

241
00:12:36.360 --> 00:12:38.480
<v Speaker 3>Shell class have shells.

242
00:12:38.480 --> 00:12:42.120
<v Speaker 4>Exactly that's where it should be. So it's an interesting attack.

243
00:12:42.159 --> 00:12:45.080
<v Speaker 4>We've seen attacks of our web sockets before, but it's

244
00:12:45.080 --> 00:12:49.000
<v Speaker 4>that weird. I'm safe because nobody can access it unless

245
00:12:49.000 --> 00:12:51.519
<v Speaker 4>they're at my computer, but realize you're also on your

246
00:12:51.559 --> 00:12:54.639
<v Speaker 4>computer doing things, and malicious code can then access it

247
00:12:54.679 --> 00:12:55.000
<v Speaker 4>down way.

248
00:12:55.000 --> 00:12:57.120
<v Speaker 2>So the pass where it's going to be like password

249
00:12:57.159 --> 00:12:59.679
<v Speaker 2>with an act or something, nobody will know.

250
00:13:00.519 --> 00:13:06.120
<v Speaker 4>Set your passwords to security this week. Nobody will gets

251
00:13:06.159 --> 00:13:11.159
<v Speaker 4>that right, all right, So that was an interesting thing.

252
00:13:12.240 --> 00:13:14.759
<v Speaker 2>I don't think there's anything that anybody can do about

253
00:13:14.799 --> 00:13:16.279
<v Speaker 2>it except don't run open claw.

254
00:13:16.360 --> 00:13:18.480
<v Speaker 4>Come on, I mean yeah, And if you're going to

255
00:13:18.559 --> 00:13:21.200
<v Speaker 4>run a better password and isolate it from your network

256
00:13:21.240 --> 00:13:23.679
<v Speaker 4>and don't run it on a computer you're actually sitting at.

257
00:13:23.960 --> 00:13:26.200
<v Speaker 3>If you get to run anything, you need to understand

258
00:13:26.200 --> 00:13:29.679
<v Speaker 3>the risks. Yeah's And the problem is most people.

259
00:13:29.399 --> 00:13:30.000
<v Speaker 5>Don't, right.

260
00:13:30.320 --> 00:13:33.759
<v Speaker 4>I think Amazon just announced that you can actually deploy

261
00:13:33.840 --> 00:13:37.000
<v Speaker 4>open claw in their cloud, oh boy, which would actually

262
00:13:37.039 --> 00:13:39.120
<v Speaker 4>make it super isolated and that sort of stuff. So

263
00:13:40.000 --> 00:13:42.799
<v Speaker 4>although here's what I'm thinking, that actually might be a

264
00:13:42.840 --> 00:13:45.720
<v Speaker 4>really great attack path to have a whole take off

265
00:13:46.120 --> 00:13:51.879
<v Speaker 4>army of Claus sitting up in Amazon. Anyways, Okay, Danne

266
00:13:51.919 --> 00:13:56.159
<v Speaker 4>dream on. So this one is a PSA. Most Wi

267
00:13:56.200 --> 00:14:00.600
<v Speaker 4>Fi routers are vulnerable to an air snitch attack. Here's

268
00:14:00.639 --> 00:14:02.559
<v Speaker 4>what to do. First of all, what's the air snitch?

269
00:14:03.120 --> 00:14:06.600
<v Speaker 4>So air snitch is what they're dubbing the attack because

270
00:14:06.600 --> 00:14:10.080
<v Speaker 4>it can be man in the middle. And we've talked

271
00:14:10.080 --> 00:14:12.679
<v Speaker 4>about these tactics before. So mad in the middle is

272
00:14:12.720 --> 00:14:15.639
<v Speaker 4>like Carl goes to his bank on the Internet and

273
00:14:15.679 --> 00:14:18.279
<v Speaker 4>he's doing banking. Now, if I can get in between

274
00:14:18.360 --> 00:14:21.559
<v Speaker 4>that traffic and in this case, let's assume there's no

275
00:14:21.759 --> 00:14:23.879
<v Speaker 4>secure socket layers or anything like that. I can to

276
00:14:23.919 --> 00:14:25.320
<v Speaker 4>see the traffic that's man in the middle, and I

277
00:14:25.360 --> 00:14:28.360
<v Speaker 4>can manipulate it. I could send back to Carl that

278
00:14:28.399 --> 00:14:31.159
<v Speaker 4>there's no money in his accounts. I could send to

279
00:14:31.240 --> 00:14:33.159
<v Speaker 4>the bank that I want to transfer money whatever. Right,

280
00:14:33.200 --> 00:14:35.200
<v Speaker 4>because I'm in the middle, I get to change that traffic.

281
00:14:35.960 --> 00:14:39.600
<v Speaker 4>So this air snitch is actually man in the middle

282
00:14:39.679 --> 00:14:43.159
<v Speaker 4>on Wi Fi. Now, there are some caveats because that

283
00:14:43.200 --> 00:14:46.200
<v Speaker 4>sounds really bad where they're like, listen, there are no protections.

284
00:14:46.480 --> 00:14:50.600
<v Speaker 4>You know you're in trouble. Blah blah blah. The way

285
00:14:50.679 --> 00:14:54.360
<v Speaker 4>Wi Fi typically works is you have a password, right

286
00:14:54.759 --> 00:14:56.679
<v Speaker 4>that goes along with the Wi Fi, and it has

287
00:14:56.720 --> 00:14:59.399
<v Speaker 4>to be at least eight characters. We always recommend longer

288
00:14:59.399 --> 00:15:00.960
<v Speaker 4>than a characters, but it has to be at least as

289
00:15:01.039 --> 00:15:02.960
<v Speaker 4>long as possible. As long as eight might as well

290
00:15:03.039 --> 00:15:05.639
<v Speaker 4>be three. But I will tell you, yeah, I know, right,

291
00:15:05.720 --> 00:15:09.960
<v Speaker 4>I will tell you, like open networks. Go to Starbucks,

292
00:15:09.960 --> 00:15:12.600
<v Speaker 4>it's an open network, right, Go to the airport. Go

293
00:15:12.639 --> 00:15:15.240
<v Speaker 4>to the airport, is sit on a delta flight. If

294
00:15:15.240 --> 00:15:16.639
<v Speaker 4>they're all may go.

295
00:15:16.600 --> 00:15:18.360
<v Speaker 2>To Chili's and get Chili's Wi Fi.

296
00:15:18.600 --> 00:15:21.720
<v Speaker 4>Right, there's still networks. So there's no password there, which means.

297
00:15:21.639 --> 00:15:23.080
<v Speaker 2>So that means there's no encryption.

298
00:15:23.320 --> 00:15:25.799
<v Speaker 3>Yeah, the barrier to entry is love.

299
00:15:25.600 --> 00:15:27.159
<v Speaker 2>Because the password is used for encryption.

300
00:15:27.399 --> 00:15:32.720
<v Speaker 4>Yeah, right, exactly. So what happens is, let's say it's

301
00:15:32.759 --> 00:15:35.720
<v Speaker 4>not strong, Patrick, Carl and I we're on a you know,

302
00:15:35.720 --> 00:15:37.519
<v Speaker 4>we're on a delta flight and we're using the Wi Fi.

303
00:15:38.399 --> 00:15:41.000
<v Speaker 4>And let's say I, being the nefarious hacker, want to

304
00:15:41.000 --> 00:15:43.559
<v Speaker 4>see who else is on the network, right, And I'm

305
00:15:43.559 --> 00:15:45.039
<v Speaker 4>going to go and scan that network. And what I'm

306
00:15:45.039 --> 00:15:47.879
<v Speaker 4>going to find is I see nobody except for me

307
00:15:48.600 --> 00:15:49.320
<v Speaker 4>in the gateway.

308
00:15:49.720 --> 00:15:50.200
<v Speaker 5>That's it.

309
00:15:51.000 --> 00:15:53.600
<v Speaker 4>And the reason being is most public networks like that

310
00:15:53.679 --> 00:15:57.960
<v Speaker 4>do what's called client isolation, where I can't talk directly

311
00:15:57.960 --> 00:16:00.440
<v Speaker 4>to Patrick, I can't talk directly to Carl, and Carl

312
00:16:00.519 --> 00:16:03.159
<v Speaker 4>can't talk to Patrick. Client isolation allows me only talk

313
00:16:03.200 --> 00:16:06.159
<v Speaker 4>to gateway like v Lance exactly. V Lance is a

314
00:16:06.159 --> 00:16:11.639
<v Speaker 4>great example of client isolation. What they found with air

315
00:16:11.759 --> 00:16:17.120
<v Speaker 4>snitch is that level of isolation, if it's on, can

316
00:16:17.279 --> 00:16:22.600
<v Speaker 4>be bypassed where I can manipulate the traffic to tell

317
00:16:22.840 --> 00:16:28.200
<v Speaker 4>the router the access point I'm Carl's gateway send all

318
00:16:28.279 --> 00:16:31.159
<v Speaker 4>of his traffic to me. So what happens is Carl,

319
00:16:31.480 --> 00:16:34.279
<v Speaker 4>when he gets on that network, he gets isolated. How

320
00:16:34.279 --> 00:16:36.240
<v Speaker 4>does he get isolated? It's actually a little bit more

321
00:16:36.240 --> 00:16:39.279
<v Speaker 4>complicated than that. There's a session key between Carl and

322
00:16:39.320 --> 00:16:42.960
<v Speaker 4>the access point. Then nobody can decrypt, right And there's

323
00:16:42.960 --> 00:16:45.559
<v Speaker 4>a unique session key from Patrick to that access point,

324
00:16:45.559 --> 00:16:47.759
<v Speaker 4>and there's unique session key from me to that access point,

325
00:16:47.840 --> 00:16:50.080
<v Speaker 4>So I can't decrypt that. But if I tell the

326
00:16:50.120 --> 00:16:52.279
<v Speaker 4>access point Carl has to go to me to get

327
00:16:52.279 --> 00:16:55.600
<v Speaker 4>to the Internet. What happens is the access point routes

328
00:16:55.799 --> 00:17:00.519
<v Speaker 4>gets the traffic from Carl, decrypts it, re encrypts it

329
00:17:00.559 --> 00:17:03.399
<v Speaker 4>with my key, and sends it to me. Then if

330
00:17:03.399 --> 00:17:05.519
<v Speaker 4>I can do whatever I want with it, and I

331
00:17:05.559 --> 00:17:07.519
<v Speaker 4>can then send it to the Internet. So it's a

332
00:17:07.519 --> 00:17:10.079
<v Speaker 4>pretty novel attack when you look at it that way,

333
00:17:10.119 --> 00:17:13.359
<v Speaker 4>where they found out which layers of Wi Fi they

334
00:17:13.359 --> 00:17:17.559
<v Speaker 4>can manipulate, whether it's ARP spoofing or whatever that bypasses.

335
00:17:17.559 --> 00:17:21.279
<v Speaker 3>If Carl's using a VPN, he'll still be protected right absolutely.

336
00:17:21.319 --> 00:17:23.480
<v Speaker 4>And that's a great point. Like if you're encrypting your

337
00:17:23.519 --> 00:17:26.599
<v Speaker 4>traffic another so you have that encrypted traffic to your

338
00:17:26.640 --> 00:17:28.880
<v Speaker 4>access point, but the access point can decrypt it. If

339
00:17:28.920 --> 00:17:32.119
<v Speaker 4>I'm using a VPN, I already have an encrypted session

340
00:17:32.119 --> 00:17:34.720
<v Speaker 4>to a remote server, right, So at that point you're

341
00:17:34.759 --> 00:17:35.799
<v Speaker 4>not going to get mad in the middle.

342
00:17:35.920 --> 00:17:38.960
<v Speaker 2>And here's the story. You should probably always use a

343
00:17:39.039 --> 00:17:42.160
<v Speaker 2>VPN when you're using a Wi Fi endpoint that doesn't

344
00:17:42.200 --> 00:17:43.440
<v Speaker 2>have a password.

345
00:17:43.039 --> 00:17:44.119
<v Speaker 5>One hundred percent.

346
00:17:44.240 --> 00:17:45.960
<v Speaker 4>Absolutely, Yeah, attacks like this.

347
00:17:46.240 --> 00:17:48.480
<v Speaker 2>But here's the problem though. You have to connect to

348
00:17:48.559 --> 00:17:51.200
<v Speaker 2>the end point before you can enable the VPN, right.

349
00:17:51.799 --> 00:17:53.880
<v Speaker 4>So you do, yeah, and you're going to connect out

350
00:17:53.880 --> 00:17:55.839
<v Speaker 4>to that. But what but if I connect to that

351
00:17:55.960 --> 00:17:59.000
<v Speaker 4>endpoint access point. Let's say I'm on the Delta flight

352
00:17:59.039 --> 00:18:03.559
<v Speaker 4>again and I go to VPN, my certificate is not

353
00:18:03.680 --> 00:18:06.200
<v Speaker 4>going to allow me to connect to the remote host

354
00:18:06.240 --> 00:18:08.480
<v Speaker 4>if it's man in the middle. So what's gonna happen

355
00:18:08.519 --> 00:18:10.839
<v Speaker 4>is because they unable to connect, unable to connect, unable

356
00:18:10.839 --> 00:18:13.359
<v Speaker 4>to connect. And if that's the case, shut your laptop

357
00:18:13.400 --> 00:18:14.440
<v Speaker 4>down and wait until you land.

358
00:18:14.559 --> 00:18:17.400
<v Speaker 3>You've detected. Yeah, it's a telltale sign that you've got

359
00:18:17.440 --> 00:18:18.039
<v Speaker 3>somebody yep.

360
00:18:18.119 --> 00:18:20.480
<v Speaker 4>Yeah, if you're a VPM won't connect, it's they could

361
00:18:20.480 --> 00:18:22.400
<v Speaker 4>be blocking it. But either way, I would just shut

362
00:18:22.440 --> 00:18:22.680
<v Speaker 4>it down.

363
00:18:22.720 --> 00:18:24.839
<v Speaker 3>It's how I It's how I found out find out

364
00:18:24.839 --> 00:18:26.559
<v Speaker 3>if Dwayne's checked into the hotel yet.

365
00:18:26.799 --> 00:18:30.960
<v Speaker 4>Exactly if you can VPN in, you know, I'm not

366
00:18:31.000 --> 00:18:32.000
<v Speaker 4>at the hotel.

367
00:18:32.839 --> 00:18:34.960
<v Speaker 2>And if you get attacks now, man, I'm in the lobby.

368
00:18:34.960 --> 00:18:40.359
<v Speaker 2>Don't believe that. Don't believe it's.

369
00:18:38.920 --> 00:18:41.799
<v Speaker 4>What's weird is this is a very novel, unique attack

370
00:18:42.000 --> 00:18:47.440
<v Speaker 4>for bypassing client isolation. I have been on plenty of

371
00:18:47.480 --> 00:18:49.799
<v Speaker 4>networks I've been traveling over the last couple of weeks.

372
00:18:49.799 --> 00:18:51.480
<v Speaker 4>I've been on plenty of networks over the last couple

373
00:18:51.480 --> 00:18:54.000
<v Speaker 4>of weeks that did not even have client isolation. I

374
00:18:54.000 --> 00:18:56.640
<v Speaker 4>could see all of the other clients on that on

375
00:18:56.680 --> 00:19:00.559
<v Speaker 4>that Wi Fi. So, yeah, this is a WPA three attack.

376
00:19:00.799 --> 00:19:05.680
<v Speaker 3>So WPA is WPA, WPA two, WPA three, WPA three

377
00:19:05.799 --> 00:19:08.599
<v Speaker 3>is regarded as the best. It's still the best, Yeah

378
00:19:08.640 --> 00:19:11.119
<v Speaker 3>it is. I don't think there's a WPA four defined yet.

379
00:19:11.480 --> 00:19:11.920
<v Speaker 2>I don't know.

380
00:19:12.519 --> 00:19:15.200
<v Speaker 4>No, there's there are, Like there are tons of RFCs

381
00:19:15.200 --> 00:19:18.000
<v Speaker 4>for brand new you know whatever is. But your router

382
00:19:18.079 --> 00:19:20.279
<v Speaker 4>at home is going to be supporting WA three, right, But.

383
00:19:20.599 --> 00:19:23.279
<v Speaker 3>There's no WPA four to four to support yet that

384
00:19:23.279 --> 00:19:26.000
<v Speaker 3>that is widely out there, is there? I don't think

385
00:19:26.079 --> 00:19:27.839
<v Speaker 3>I don't think so, it's been a while since I

386
00:19:27.839 --> 00:19:28.640
<v Speaker 3>can figured a router.

387
00:19:30.039 --> 00:19:31.680
<v Speaker 2>All right, you guys ready to move on?

388
00:19:31.920 --> 00:19:32.880
<v Speaker 5>Yeah, we're ready.

389
00:19:33.039 --> 00:19:38.039
<v Speaker 2>Hacker News says Europole led operation takes down Tycoon two

390
00:19:38.200 --> 00:19:43.119
<v Speaker 2>FA fishing as a service linked to sixty four thousand attacks.

391
00:19:43.359 --> 00:19:43.920
<v Speaker 3>That's awesome.

392
00:19:44.079 --> 00:19:45.480
<v Speaker 2>Yeah, that is good news.

393
00:19:45.519 --> 00:19:48.519
<v Speaker 4>Golf Clapp, well done, well done.

394
00:19:48.359 --> 00:19:51.880
<v Speaker 2>Well done. You So what's the story here? How did

395
00:19:51.880 --> 00:19:56.079
<v Speaker 2>they find it? And I mean sixty four thousand attacks that,

396
00:19:56.599 --> 00:20:01.000
<v Speaker 2>I mean, that's every attack could potentially harm you know,

397
00:20:01.119 --> 00:20:02.079
<v Speaker 2>millions of people.

398
00:20:02.240 --> 00:20:04.240
<v Speaker 4>Yeah, absolutely, So that's a lot.

399
00:20:04.039 --> 00:20:07.920
<v Speaker 2>Of a lot of bad stuff. Wow. Yeah.

400
00:20:07.960 --> 00:20:10.839
<v Speaker 4>And a lot of times the way that these are

401
00:20:10.960 --> 00:20:15.039
<v Speaker 4>detected is like you'll take something like Proofpoint where they're

402
00:20:15.039 --> 00:20:17.839
<v Speaker 4>seeing a lot of emails, they're seeing a lot of communications,

403
00:20:17.920 --> 00:20:21.000
<v Speaker 4>and they start to see a trend in people going

404
00:20:21.039 --> 00:20:24.079
<v Speaker 4>to hacker dot com for authentication or having links in

405
00:20:24.119 --> 00:20:24.640
<v Speaker 4>their emails.

406
00:20:24.640 --> 00:20:25.279
<v Speaker 5>And that's first off.

407
00:20:25.799 --> 00:20:27.920
<v Speaker 2>So a lot of this is for Joe's fishing as

408
00:20:27.960 --> 00:20:32.039
<v Speaker 2>a service dot com, right, Yeah.

409
00:20:31.200 --> 00:20:33.720
<v Speaker 4>So a lot of this is, you know, having the

410
00:20:33.799 --> 00:20:38.920
<v Speaker 4>proper defenses and monitoring in place that's looking for these

411
00:20:38.960 --> 00:20:41.599
<v Speaker 4>types of phishing attacks, looking for emails coming in with

412
00:20:41.640 --> 00:20:44.160
<v Speaker 4>all of the same type of URL going to the

413
00:20:44.200 --> 00:20:47.400
<v Speaker 4>same type of location, going to an untrusted site. Right,

414
00:20:47.440 --> 00:20:50.720
<v Speaker 4>there's all sorts of these metrics that are used inside

415
00:20:50.720 --> 00:20:53.440
<v Speaker 4>of phishing detection. And then what will happen is, you know,

416
00:20:53.480 --> 00:20:55.799
<v Speaker 4>they'll start gathering some of them and they'll get users

417
00:20:56.119 --> 00:20:58.720
<v Speaker 4>are probably clicking on them, but they'll they'll start picking

418
00:20:58.759 --> 00:21:00.359
<v Speaker 4>them all up, and then they put together there's some

419
00:21:00.400 --> 00:21:02.640
<v Speaker 4>sort of concerted effort to take them down, and a

420
00:21:02.640 --> 00:21:05.359
<v Speaker 4>lot of times that's organizing with local law enforcement in

421
00:21:05.400 --> 00:21:07.720
<v Speaker 4>different areas and that sort of stuff, or even ISPs

422
00:21:07.799 --> 00:21:08.839
<v Speaker 4>to take down these sites.

423
00:21:09.000 --> 00:21:09.240
<v Speaker 2>Yeah.

424
00:21:09.400 --> 00:21:13.680
<v Speaker 4>So hey Patrick, Yeah, in the article that they do

425
00:21:13.759 --> 00:21:16.119
<v Speaker 4>mention a very cool service which I will talk about.

426
00:21:16.200 --> 00:21:18.440
<v Speaker 2>No, please don't, but you'll have to read the article

427
00:21:18.480 --> 00:21:21.240
<v Speaker 2>if you want the read the article Career Criminal Advice.

428
00:21:21.359 --> 00:21:23.240
<v Speaker 4>But I can tell you if you're looking for some

429
00:21:23.480 --> 00:21:25.960
<v Speaker 4>high end access some really cool data. There's some companies

430
00:21:26.119 --> 00:21:28.440
<v Speaker 4>mentioned in this article that I know personally.

431
00:21:28.880 --> 00:21:31.480
<v Speaker 3>Yeah, wow, lose my number. Yeah.

432
00:21:31.519 --> 00:21:35.160
<v Speaker 2>So Microsoft was kind of involved in this too, right, Yeah.

433
00:21:35.200 --> 00:21:35.759
<v Speaker 5>Absolutely.

434
00:21:35.839 --> 00:21:38.480
<v Speaker 2>I mean their logo is right at the top of

435
00:21:38.519 --> 00:21:42.119
<v Speaker 2>this thing, right below europoll. So did they help out?

436
00:21:42.319 --> 00:21:43.359
<v Speaker 3>Yeah? Yeah, I think they did.

437
00:21:43.519 --> 00:21:45.839
<v Speaker 4>Yeah, yeah, I mean sixty two percent. According to this article,

438
00:21:45.880 --> 00:21:48.160
<v Speaker 4>sixty two percent of these fishing attempts were actually blocked

439
00:21:48.160 --> 00:21:51.039
<v Speaker 4>by Microsoft. Wow, it's funny. I talked to a lot

440
00:21:51.079 --> 00:21:53.240
<v Speaker 4>of people about phishing, of course, and they're like, what

441
00:21:53.279 --> 00:21:55.000
<v Speaker 4>should we do? And should I? You know, how do I?

442
00:21:55.079 --> 00:21:58.119
<v Speaker 4>And obviously we talk about like user education and training

443
00:21:58.119 --> 00:21:59.799
<v Speaker 4>and that sort of stuff like, yeah, get your users

444
00:21:59.839 --> 00:22:02.480
<v Speaker 4>to detect what's going on, right, But I will tell

445
00:22:02.480 --> 00:22:05.160
<v Speaker 4>you they're like, well do you trust like Microsoft's phishing

446
00:22:05.240 --> 00:22:07.680
<v Speaker 4>detection and blah blah blah. And I'm like, who do

447
00:22:07.759 --> 00:22:11.200
<v Speaker 4>you think other than Google sees more emails on the planet?

448
00:22:11.279 --> 00:22:11.599
<v Speaker 2>Yeah?

449
00:22:11.680 --> 00:22:14.680
<v Speaker 4>Right, is it? It's probably Microsoft?

450
00:22:14.839 --> 00:22:14.920
<v Speaker 5>Right?

451
00:22:15.000 --> 00:22:16.640
<v Speaker 2>Yeah, so they got all the corporate emails.

452
00:22:16.680 --> 00:22:19.440
<v Speaker 3>Yeah, they have like five hundred million customers.

453
00:22:19.599 --> 00:22:22.799
<v Speaker 4>Yeah, so I would, I would venture to say, and

454
00:22:22.839 --> 00:22:25.279
<v Speaker 4>this is this. Somebody can call me out on these statistics,

455
00:22:25.279 --> 00:22:28.279
<v Speaker 4>but between Google and Microsoft, you're probably looking at about

456
00:22:28.319 --> 00:22:30.720
<v Speaker 4>eighty to ninety percent of all emails sent on the planet.

457
00:22:30.799 --> 00:22:33.559
<v Speaker 4>I bet right, So, yeah, I would trust one of

458
00:22:33.599 --> 00:22:36.079
<v Speaker 4>those two to discover it's a phishing attack over you

459
00:22:36.079 --> 00:22:36.960
<v Speaker 4>know you?

460
00:22:37.480 --> 00:22:38.720
<v Speaker 2>Yeah, right, exactly?

461
00:22:39.119 --> 00:22:41.480
<v Speaker 4>No, even me, Honestly, I'll be honest with you No, I.

462
00:22:41.440 --> 00:22:45.119
<v Speaker 3>Mean the alarm goes offt never mind, is just the commenter?

463
00:22:46.279 --> 00:22:48.920
<v Speaker 4>No, I will tell you. Even me, like, our team

464
00:22:49.000 --> 00:22:51.799
<v Speaker 4>has looked at phishing emails of course all the time,

465
00:22:51.799 --> 00:22:53.400
<v Speaker 4>because we emulate them and use them and that sort

466
00:22:53.400 --> 00:22:55.599
<v Speaker 4>of stuff, and there are some that are so good

467
00:22:56.039 --> 00:22:58.240
<v Speaker 4>we're like, now I think this is real.

468
00:22:58.559 --> 00:23:00.839
<v Speaker 3>Think the hardest I think the to be the hardest

469
00:23:00.839 --> 00:23:03.119
<v Speaker 3>target is to be just so parent. Like I call

470
00:23:03.160 --> 00:23:04.920
<v Speaker 3>you regularly to say did you send this to me?

471
00:23:05.440 --> 00:23:05.799
<v Speaker 5>Yeah?

472
00:23:05.839 --> 00:23:07.920
<v Speaker 3>And sometimes it's things I've asked you to send to me.

473
00:23:10.720 --> 00:23:12.960
<v Speaker 4>I have had customers say, you know what, I passed

474
00:23:12.960 --> 00:23:15.440
<v Speaker 4>the fishing I passed the fishing tests all the time,

475
00:23:15.440 --> 00:23:15.839
<v Speaker 4>and I'm like.

476
00:23:15.759 --> 00:23:16.559
<v Speaker 2>Wow, how do you do that?

477
00:23:16.759 --> 00:23:19.279
<v Speaker 4>They're like every email that comes in my inbox, I say,

478
00:23:19.359 --> 00:23:21.799
<v Speaker 4>this is fishing, this is fishing, this is fishing, And

479
00:23:21.839 --> 00:23:24.680
<v Speaker 4>eventually they send them back going, no, that one's that

480
00:23:24.720 --> 00:23:27.359
<v Speaker 4>one's good. Nope, that one's good, and see, I'm good.

481
00:23:27.400 --> 00:23:29.480
<v Speaker 4>I never click on that. I'm like, I was like,

482
00:23:29.519 --> 00:23:31.880
<v Speaker 4>but what if I send you an email that at

483
00:23:31.880 --> 00:23:34.240
<v Speaker 4>the bottom says, if you think this is fishing, click here,

484
00:23:36.119 --> 00:23:40.279
<v Speaker 4>Yeah that'll probably yeah, no problem. If you think this

485
00:23:40.440 --> 00:23:48.119
<v Speaker 4>is click here come on, yes, credit card to verify

486
00:23:48.200 --> 00:23:50.599
<v Speaker 4>you're a human. Copy and paste this command.

487
00:23:54.359 --> 00:23:56.359
<v Speaker 2>All right, very good, we'll be right back after these

488
00:23:56.480 --> 00:24:03.000
<v Speaker 2>very important messages, and we're back. It's Security this week.

489
00:24:03.000 --> 00:24:06.519
<v Speaker 2>I'm Carl Franklin, that's Dwayne Laflott and Patrick Hines, and

490
00:24:06.720 --> 00:24:10.240
<v Speaker 2>we're halfway through the show today. We got a big

491
00:24:10.279 --> 00:24:13.119
<v Speaker 2>one coming up at the end. But let's go to

492
00:24:13.839 --> 00:24:17.559
<v Speaker 2>this one. State affiliated hackers set up for critical OT

493
00:24:17.880 --> 00:24:23.559
<v Speaker 2>attacks that operators may not detect. What is OT in this.

494
00:24:23.599 --> 00:24:25.359
<v Speaker 3>Case, operational technology?

495
00:24:25.519 --> 00:24:29.880
<v Speaker 4>Okay, yeah, so anything that runs a pump, power plants,

496
00:24:30.160 --> 00:24:34.519
<v Speaker 4>dams exactly, go, a water treatment plate, damn power plants, Yeah,

497
00:24:34.559 --> 00:24:37.599
<v Speaker 4>water treatment plants, nuclear plants, I mean, you name it.

498
00:24:37.640 --> 00:24:40.960
<v Speaker 4>That's all industrial control systems ROCK, what they call operation control.

499
00:24:41.359 --> 00:24:45.519
<v Speaker 2>So they're saying that the hackers are getting ready or

500
00:24:45.599 --> 00:24:50.599
<v Speaker 2>set up for critical infrastructure attacks that operators might not detect.

501
00:24:51.000 --> 00:24:54.079
<v Speaker 2>That's scary and fascinating at the same time.

502
00:24:54.319 --> 00:25:01.079
<v Speaker 4>So this is it's complex operation. We always recommend patching

503
00:25:01.119 --> 00:25:03.079
<v Speaker 4>your computers. We always recommend being on the latest version.

504
00:25:03.119 --> 00:25:05.079
<v Speaker 4>We always recommend being on the cutting edge because you're

505
00:25:05.559 --> 00:25:10.759
<v Speaker 4>less likely to be exploited if you have the latest patch. Right,

506
00:25:10.839 --> 00:25:13.720
<v Speaker 4>but you are more likely to blue screen your computer

507
00:25:13.799 --> 00:25:16.599
<v Speaker 4>and have to bring it back from death. Yes, And

508
00:25:16.640 --> 00:25:19.519
<v Speaker 4>where's the probably the one of two places you don't

509
00:25:19.519 --> 00:25:22.160
<v Speaker 4>want a computer to blue screen, Probably if it's running

510
00:25:22.200 --> 00:25:24.680
<v Speaker 4>a power grid or it's running an iron lung for

511
00:25:24.759 --> 00:25:27.480
<v Speaker 4>somebody in a hospital. Right, So those are the two

512
00:25:27.480 --> 00:25:32.160
<v Speaker 4>places we see computers with default logins. Windows XP never

513
00:25:32.240 --> 00:25:35.319
<v Speaker 4>been patched, right, because they're just we're gonna it's up

514
00:25:35.359 --> 00:25:37.640
<v Speaker 4>and running, it's doing what it needs to do. Just

515
00:25:37.680 --> 00:25:39.839
<v Speaker 4>don't touch it, Please, don't breathe near it, just leave

516
00:25:39.880 --> 00:25:43.079
<v Speaker 4>it alone. Right, And that's what we're seeing here is

517
00:25:43.079 --> 00:25:46.480
<v Speaker 4>this is automation of attacks of what they call edge systems,

518
00:25:47.039 --> 00:25:51.000
<v Speaker 4>which are either older technology, default user names and passwords,

519
00:25:51.000 --> 00:25:53.079
<v Speaker 4>et cetera. And once they get in, then they start

520
00:25:53.119 --> 00:25:55.119
<v Speaker 4>to infect the industrial control systems.

521
00:25:55.799 --> 00:25:59.880
<v Speaker 2>Yeah, so what can we do better?

522
00:26:00.480 --> 00:26:01.119
<v Speaker 4>We can do better?

523
00:26:01.240 --> 00:26:02.559
<v Speaker 2>Ye? Can we do better?

524
00:26:02.720 --> 00:26:02.920
<v Speaker 5>Yeah?

525
00:26:02.960 --> 00:26:05.480
<v Speaker 4>I mean, honestly, I will tell you there's a couple

526
00:26:05.480 --> 00:26:08.599
<v Speaker 4>of different things, especially when I'm dealing with operational technology

527
00:26:09.000 --> 00:26:13.400
<v Speaker 4>OT systems. One isolation that's key, Like if you can't

528
00:26:13.480 --> 00:26:16.240
<v Speaker 4>change the standard. If Windows XP is unfortunately the only

529
00:26:16.279 --> 00:26:21.079
<v Speaker 4>thing that can run these devices, the nuclear centrifugeon. Don't

530
00:26:21.079 --> 00:26:23.119
<v Speaker 4>put it on the internet, right.

531
00:26:23.759 --> 00:26:26.559
<v Speaker 2>Or at least have a strong VPN. Yeah, and rotate

532
00:26:26.599 --> 00:26:28.960
<v Speaker 2>the keys like you know, simple stuff.

533
00:26:29.160 --> 00:26:32.319
<v Speaker 4>Yeah, or or heads up, make it air gapped. I

534
00:26:32.359 --> 00:26:34.720
<v Speaker 4>know it's inconvenient for an engineer to have to actually

535
00:26:34.759 --> 00:26:37.960
<v Speaker 4>come into the building and do something, but it's probably

536
00:26:38.000 --> 00:26:40.079
<v Speaker 4>more secure than having it accessible from home.

537
00:26:40.160 --> 00:26:41.960
<v Speaker 3>So if it's not convenient, it's not secure.

538
00:26:42.000 --> 00:26:42.440
<v Speaker 2>That's right.

539
00:26:44.960 --> 00:26:46.160
<v Speaker 5>If it's convenient, switch.

540
00:26:47.319 --> 00:26:49.680
<v Speaker 3>If it's not, it's not inconvenient, it's not secure.

541
00:26:50.160 --> 00:26:50.319
<v Speaker 2>You know.

542
00:26:50.359 --> 00:26:58.839
<v Speaker 4>Willy Wonka there, good day, sir.

543
00:26:58.920 --> 00:26:59.960
<v Speaker 3>Roll it back reverse.

544
00:27:00.440 --> 00:27:01.519
<v Speaker 2>God. I love that movie.

545
00:27:01.599 --> 00:27:02.359
<v Speaker 5>It is so good.

546
00:27:03.160 --> 00:27:07.359
<v Speaker 4>So you know that honestly. And if you're ever interested

547
00:27:07.400 --> 00:27:11.000
<v Speaker 4>in really kind of deep diving into operational technology and

548
00:27:11.039 --> 00:27:12.759
<v Speaker 4>that sort of stuff, there's a there's a great guy

549
00:27:12.759 --> 00:27:17.920
<v Speaker 4>who runs a podcast called ot Defender. That's very great guy.

550
00:27:18.880 --> 00:27:21.920
<v Speaker 4>Awesome pen testing company, but they very much focus on

551
00:27:22.359 --> 00:27:24.240
<v Speaker 4>operational technology and industrial control.

552
00:27:24.319 --> 00:27:25.519
<v Speaker 5>So there's a lot of cool stuff there.

553
00:27:25.559 --> 00:27:30.000
<v Speaker 4>But in short, isolate computers that are controlling industrial control systems,

554
00:27:30.240 --> 00:27:33.640
<v Speaker 4>have strong passwords and then patch whatever you can.

555
00:27:33.839 --> 00:27:40.839
<v Speaker 2>Yeah, okay, so let's move on. CVE twenty twenty six,

556
00:27:41.000 --> 00:27:45.359
<v Speaker 2>DASH two seven nine four. This is engine x.

557
00:27:45.880 --> 00:27:47.039
<v Speaker 3>A newbie, but a goodie.

558
00:27:47.279 --> 00:27:50.759
<v Speaker 2>Yeah, it is a a newbie. Yeah, so this is well,

559
00:27:50.799 --> 00:27:53.359
<v Speaker 2>the story is a newbie engine XUI.

560
00:27:53.480 --> 00:27:56.640
<v Speaker 3>It's all about back is uh yeah.

561
00:27:56.319 --> 00:27:59.160
<v Speaker 2>It's a backup disclosure. So if you follow this as

562
00:27:59.160 --> 00:28:02.279
<v Speaker 2>a hacker wire and if you follow it the links

563
00:28:02.319 --> 00:28:05.000
<v Speaker 2>down at the bottom, the second one, the artificial intelligence

564
00:28:05.039 --> 00:28:09.319
<v Speaker 2>analysis is actually pretty good at describing the problem. A

565
00:28:09.359 --> 00:28:14.480
<v Speaker 2>critical unauthenticated sensitive data disclosure vulnerability and engine XUI a

566
00:28:14.519 --> 00:28:18.599
<v Speaker 2>web user interface for the engine x web server. This

567
00:28:18.680 --> 00:28:24.400
<v Speaker 2>flaw carries a CVSS score of nine point eight and

568
00:28:24.880 --> 00:28:28.119
<v Speaker 2>I just wonder what the other point two percent was

569
00:28:28.599 --> 00:28:34.359
<v Speaker 2>left off for. Right. An unauthenticated attacker can leverage this

570
00:28:34.480 --> 00:28:38.319
<v Speaker 2>vulnerability to download a full system backup and immediately decryptic,

571
00:28:38.359 --> 00:28:41.839
<v Speaker 2>gaining access to highly sensitive data. Probably the reason why

572
00:28:41.880 --> 00:28:45.920
<v Speaker 2>this is so high this score is because everybody who's

573
00:28:46.000 --> 00:28:51.799
<v Speaker 2>running Kubernetes or using containers uses engine x because it's

574
00:28:51.839 --> 00:28:54.839
<v Speaker 2>a lightweight Linux web server that they can run in

575
00:28:54.880 --> 00:28:57.440
<v Speaker 2>a container and that's the darling of.

576
00:29:00.079 --> 00:29:03.039
<v Speaker 3>Nanny's and Docker, and as a web server it needs

577
00:29:03.079 --> 00:29:04.119
<v Speaker 3>to be on the web.

578
00:29:04.119 --> 00:29:05.359
<v Speaker 5>Right, Yeah, absolutely.

579
00:29:05.640 --> 00:29:07.759
<v Speaker 3>I think the point too is that it's not delivered

580
00:29:07.759 --> 00:29:09.119
<v Speaker 3>to your inbox automatically.

581
00:29:09.759 --> 00:29:11.799
<v Speaker 4>Yeah, there's not a one one button click, I mean,

582
00:29:11.880 --> 00:29:15.240
<v Speaker 4>and this, you know, this looks like improper permissions on

583
00:29:15.279 --> 00:29:18.680
<v Speaker 4>certain directories as well. Right, So you're getting access to backups,

584
00:29:18.680 --> 00:29:20.319
<v Speaker 4>and we look for this all the time when we're

585
00:29:20.319 --> 00:29:23.559
<v Speaker 4>crawling websites, but in this particular case, there's a slash

586
00:29:23.599 --> 00:29:27.519
<v Speaker 4>API slash backup directory and there's a place for you

587
00:29:27.559 --> 00:29:31.680
<v Speaker 4>to then pull the keys for that backup. So so let's.

588
00:29:31.440 --> 00:29:34.680
<v Speaker 3>Talk about this just for a second. If Dwayne's team

589
00:29:34.839 --> 00:29:38.079
<v Speaker 3>is going after a server, they identify an IP address

590
00:29:38.160 --> 00:29:42.039
<v Speaker 3>of interest, they scan it with en map or some

591
00:29:42.119 --> 00:29:44.920
<v Speaker 3>other tool, it's almost always endmap, and they find that

592
00:29:45.039 --> 00:29:48.920
<v Speaker 3>engine x ui displays a banner from you know, they

593
00:29:48.960 --> 00:29:50.759
<v Speaker 3>get a banner and they know it's engine x ui.

594
00:29:50.920 --> 00:29:53.400
<v Speaker 3>Maybe they don't know the version, maybe they do. And

595
00:29:53.440 --> 00:29:55.559
<v Speaker 3>then what they do is they, oh, that's a web server.

596
00:29:55.680 --> 00:29:58.599
<v Speaker 3>Let's go and fuzz and see what directories we can

597
00:29:58.680 --> 00:30:04.279
<v Speaker 3>access by with unauthenticated And so that's what you're going

598
00:30:04.359 --> 00:30:07.400
<v Speaker 3>to get back. You're gonna they're gonna try API, they're

599
00:30:07.440 --> 00:30:10.640
<v Speaker 3>going to try you know, forms, they're gonna try images,

600
00:30:10.640 --> 00:30:13.480
<v Speaker 3>they're gonna try all those directories points and this one

601
00:30:13.519 --> 00:30:14.880
<v Speaker 3>would have come back open.

602
00:30:15.720 --> 00:30:16.119
<v Speaker 2>Not good.

603
00:30:16.240 --> 00:30:18.559
<v Speaker 4>Yeah, and then from there you just try and pull

604
00:30:18.640 --> 00:30:21.200
<v Speaker 4>all the information you possibly can out of that directory.

605
00:30:21.200 --> 00:30:23.920
<v Speaker 4>You try and get a files listing. There's all the

606
00:30:23.960 --> 00:30:27.960
<v Speaker 4>common things you look for, backup files, API keys, environment variables,

607
00:30:28.000 --> 00:30:29.599
<v Speaker 4>configuration files, you name it.

608
00:30:30.039 --> 00:30:30.960
<v Speaker 5>You try and pull it.

609
00:30:31.000 --> 00:30:35.599
<v Speaker 2>So yeah, not good. A patch is available, so go patch.

610
00:30:35.839 --> 00:30:38.359
<v Speaker 3>Yeah I cannot this is I wonder if this is

611
00:30:38.400 --> 00:30:39.400
<v Speaker 3>newly introduced.

612
00:30:40.240 --> 00:30:43.079
<v Speaker 4>Well, yeah, it's good question because.

613
00:30:42.920 --> 00:30:44.480
<v Speaker 3>This is something that should have been picked up if

614
00:30:44.480 --> 00:30:46.119
<v Speaker 3>they did any pen testing at all.

615
00:30:47.799 --> 00:30:50.759
<v Speaker 4>Use your upgrade to their new engine XS installation version

616
00:30:50.759 --> 00:30:53.240
<v Speaker 4>two dot three. I'm trying to see if.

617
00:30:54.119 --> 00:30:54.640
<v Speaker 5>They give it.

618
00:30:54.720 --> 00:30:58.599
<v Speaker 4>They give prior to version two dot three dot three,

619
00:30:58.799 --> 00:31:01.640
<v Speaker 4>so it sounds like, wow, it's probably been there a

620
00:31:01.680 --> 00:31:04.200
<v Speaker 4>while because they didn't say, you know, between two three

621
00:31:04.200 --> 00:31:05.519
<v Speaker 4>dot one and two three dots.

622
00:31:05.720 --> 00:31:06.559
<v Speaker 5>They just said prior to.

623
00:31:06.640 --> 00:31:08.920
<v Speaker 3>If I if I was told an AI found this,

624
00:31:09.000 --> 00:31:12.680
<v Speaker 3>I wouldn't be impressed, wow, because it's like okay, yeah,

625
00:31:12.720 --> 00:31:14.720
<v Speaker 3>I mean I one of our interns could find this.

626
00:31:15.079 --> 00:31:17.640
<v Speaker 4>Yeah, no, I see what you're saying. Yeah, I might

627
00:31:17.680 --> 00:31:19.880
<v Speaker 4>be impressed that it's been out so long and people

628
00:31:19.920 --> 00:31:23.279
<v Speaker 4>haven't seen it. Yeah, yeah, right, but you're right, it's

629
00:31:23.039 --> 00:31:25.240
<v Speaker 4>the level of complexity is low.

630
00:31:25.400 --> 00:31:29.680
<v Speaker 2>Yeah, all right. We moved on Atomic Edge reports reverse

631
00:31:29.759 --> 00:31:35.160
<v Speaker 2>engineering a phishing campaign with complex offfuscation. Now this is

632
00:31:35.200 --> 00:31:38.839
<v Speaker 2>a long article, but it kind of smells like Cliff

633
00:31:38.880 --> 00:31:42.480
<v Speaker 2>Stole story, like a cuckoo's egg about how this guy

634
00:31:42.599 --> 00:31:48.000
<v Speaker 2>or this company reverse engineered this fishing campaign and figured

635
00:31:48.000 --> 00:31:48.359
<v Speaker 2>it out.

636
00:31:49.359 --> 00:31:52.400
<v Speaker 4>So and it's it's interesting because I love some of

637
00:31:52.400 --> 00:31:55.680
<v Speaker 4>the tactics and techniques they use here for threat hunting,

638
00:31:55.720 --> 00:32:00.400
<v Speaker 4>for tracking down this particular campaign. And now I'll tell

639
00:32:00.400 --> 00:32:04.200
<v Speaker 4>you it takes time. It is funny, like I get it.

640
00:32:04.279 --> 00:32:05.960
<v Speaker 4>I get it from a lot of customers where they're like,

641
00:32:06.480 --> 00:32:08.440
<v Speaker 4>you know, ah, here's the email they sent me and

642
00:32:08.440 --> 00:32:11.839
<v Speaker 4>that's how they got in. Go find them. And I'm like, okay,

643
00:32:12.279 --> 00:32:17.000
<v Speaker 4>I mean could we Yes, it may represent week's worth

644
00:32:17.039 --> 00:32:18.559
<v Speaker 4>of work and deoffuscation.

645
00:32:18.839 --> 00:32:20.359
<v Speaker 2>It's gonna cost a lot of money.

646
00:32:20.400 --> 00:32:23.079
<v Speaker 4>I peas and like, and what are we gonna do

647
00:32:23.119 --> 00:32:24.039
<v Speaker 4>when we find them.

648
00:32:24.000 --> 00:32:25.440
<v Speaker 3>And the logs have to be in place?

649
00:32:25.720 --> 00:32:28.119
<v Speaker 4>Yeah, and we're gonna go ooh, you you shouldn't have

650
00:32:28.160 --> 00:32:31.559
<v Speaker 4>done it because they're probably outside the country anyways, and

651
00:32:31.559 --> 00:32:33.200
<v Speaker 4>we're not going to get them to get extra dided

652
00:32:33.279 --> 00:32:36.200
<v Speaker 4>like just lock down the systems better. But it is

653
00:32:36.319 --> 00:32:40.200
<v Speaker 4>cool to watch somebody go through this process and see

654
00:32:40.359 --> 00:32:44.000
<v Speaker 4>just how hard it is, just how much obfuscation goes

655
00:32:44.079 --> 00:32:47.799
<v Speaker 4>into changing the code arounds that people can't read it. So,

656
00:32:47.920 --> 00:32:49.759
<v Speaker 4>you know, because a lot of people say, well, why

657
00:32:49.759 --> 00:32:52.680
<v Speaker 4>can't anti virus find this, Why can't spam filters find this,

658
00:32:52.720 --> 00:32:55.519
<v Speaker 4>why can't Microsoft find this? Read through this article and

659
00:32:55.559 --> 00:32:59.640
<v Speaker 4>see how hard it is to de off use skate

660
00:32:59.680 --> 00:33:02.160
<v Speaker 4>and understand what's going on in this code, and then

661
00:33:02.240 --> 00:33:04.799
<v Speaker 4>say why you can't do that on masks and you'll

662
00:33:04.839 --> 00:33:05.240
<v Speaker 4>know why.

663
00:33:05.599 --> 00:33:09.279
<v Speaker 2>But it also just says volumes about these people that

664
00:33:09.440 --> 00:33:11.920
<v Speaker 2>figured it out, just how smart they are.

665
00:33:12.000 --> 00:33:13.880
<v Speaker 5>Oh yeah, you know, Yeah, it's cool stuff.

666
00:33:13.640 --> 00:33:15.599
<v Speaker 2>And it's good that we have good guys working on

667
00:33:15.640 --> 00:33:16.279
<v Speaker 2>our side.

668
00:33:16.960 --> 00:33:17.400
<v Speaker 5>Agreed.

669
00:33:18.400 --> 00:33:20.960
<v Speaker 2>Yeah, so that's a happy story. Now the not so

670
00:33:21.039 --> 00:33:27.319
<v Speaker 2>happy story, Oh boy, here we go. AI chatbots use

671
00:33:27.480 --> 00:33:32.000
<v Speaker 2>tactical nuclear weapons in ninety five percent of AI war

672
00:33:32.079 --> 00:33:36.799
<v Speaker 2>games launched. Why they always went launch strategic strikes three times?

673
00:33:37.640 --> 00:33:41.279
<v Speaker 2>Researcher put GPT five point two, Claude son At four,

674
00:33:41.720 --> 00:33:44.599
<v Speaker 2>and Gemini three against each other, with at least one

675
00:33:44.680 --> 00:33:48.319
<v Speaker 2>model using a tactical nuke in twenty out of twenty

676
00:33:48.359 --> 00:33:49.079
<v Speaker 2>one matches.

677
00:33:49.240 --> 00:33:51.799
<v Speaker 3>I'm surprised it wasn't twenty two out of twenty one matches.

678
00:33:52.160 --> 00:33:56.359
<v Speaker 4>Right, Yeah, this is an interesting article.

679
00:33:56.400 --> 00:33:58.599
<v Speaker 2>And then you listen to the news and they say that,

680
00:33:59.000 --> 00:34:01.519
<v Speaker 2>you know, the current minuetra is saying they're using AI

681
00:34:01.599 --> 00:34:03.519
<v Speaker 2>to help fight the war with Iran.

682
00:34:03.839 --> 00:34:08.519
<v Speaker 3>Yeah, well, they're prioritizing targets, they're they're coordinating. There's definitely

683
00:34:08.679 --> 00:34:12.480
<v Speaker 3>legitimate uses for as in the military realm, and then

684
00:34:12.559 --> 00:34:15.679
<v Speaker 3>there's also gray areas, and then there's just really really

685
00:34:15.719 --> 00:34:19.639
<v Speaker 3>bad ideas. This is a really really bad idea. Everyone

686
00:34:20.199 --> 00:34:23.559
<v Speaker 3>with half a brain, and I'm not saying everybody does

687
00:34:24.400 --> 00:34:27.840
<v Speaker 3>knows you don't allow ai' I mean there's just too

688
00:34:27.840 --> 00:34:31.679
<v Speaker 3>many movies. Yeah, sure, wargames, which is where the title

689
00:34:31.679 --> 00:34:35.599
<v Speaker 3>comes from, Terminator. I don't need to go any further.

690
00:34:35.639 --> 00:34:37.639
<v Speaker 3>Those two are so.

691
00:34:37.760 --> 00:34:40.639
<v Speaker 2>You know, somebody like Richard Campbell would say the current

692
00:34:40.920 --> 00:34:44.400
<v Speaker 2>fear factor around AI is because of Hollywood, But I

693
00:34:44.440 --> 00:34:46.679
<v Speaker 2>will tell you it's like, you know, that's the reason

694
00:34:46.679 --> 00:34:50.199
<v Speaker 2>that we mistrust it. Yeah, then I read articles like this, Yeah.

695
00:34:50.079 --> 00:34:53.880
<v Speaker 4>But I will tell you sometimes Hollywood actually drives change,

696
00:34:54.119 --> 00:34:57.599
<v Speaker 4>and they get it. So take, for example, Wargames nineteen

697
00:34:57.639 --> 00:35:00.519
<v Speaker 4>eighty three War Games. You're Ronald Rake and you're sitting

698
00:35:00.559 --> 00:35:02.760
<v Speaker 4>at Camp David. You're you're sitting down with your wife

699
00:35:02.800 --> 00:35:07.239
<v Speaker 4>and your box of popcorn, and you put Wargames on

700
00:35:07.679 --> 00:35:11.559
<v Speaker 4>the big movie theater there. And this actually happened, and

701
00:35:11.599 --> 00:35:14.440
<v Speaker 4>he watched it, and he was so freaked out that

702
00:35:14.519 --> 00:35:16.719
<v Speaker 4>when he went back to DC, he talked to his

703
00:35:16.880 --> 00:35:20.119
<v Speaker 4>entire cabinet and said, dear God, I saw this movie.

704
00:35:20.519 --> 00:35:23.320
<v Speaker 4>And he explained the plot, and they went, Now, that

705
00:35:23.400 --> 00:35:27.559
<v Speaker 4>can't happen, No way, Yeah, that actually might happen. And

706
00:35:27.599 --> 00:35:30.440
<v Speaker 4>they went and they they actually came out with some

707
00:35:30.480 --> 00:35:32.800
<v Speaker 4>of the computer interviews, fraud acts from nineteen eighty four

708
00:35:32.840 --> 00:35:36.320
<v Speaker 4>and that sort of stuff. So like it can drive change,

709
00:35:36.360 --> 00:35:39.400
<v Speaker 4>the more Hollywood kind of hypes something up and says,

710
00:35:39.480 --> 00:35:41.280
<v Speaker 4>these might be the possibilities.

711
00:35:41.840 --> 00:35:44.119
<v Speaker 3>Right, he wrote a strongly worded letter.

712
00:35:45.920 --> 00:35:50.480
<v Speaker 4>Now, but imagine like we all approached pairing AI with

713
00:35:50.840 --> 00:35:56.159
<v Speaker 4>robots with no restrictions as a bad thing. Why because

714
00:35:56.199 --> 00:35:59.079
<v Speaker 4>in our society we've seen movies where people did it

715
00:35:59.119 --> 00:36:02.400
<v Speaker 4>and it destroyed the humans. Right, So if none of

716
00:36:02.400 --> 00:36:05.559
<v Speaker 4>those existed, we might go, yeah, well, what could go wrong?

717
00:36:05.599 --> 00:36:09.760
<v Speaker 4>It's fine, it's just code and really dangerous machines. Let's

718
00:36:09.760 --> 00:36:10.920
<v Speaker 4>put them together and see what happened.

719
00:36:10.960 --> 00:36:12.880
<v Speaker 2>And the real fact is, I mean, let's not get

720
00:36:12.920 --> 00:36:15.199
<v Speaker 2>you a political here, otherwise Patrick will keep us here

721
00:36:15.199 --> 00:36:18.880
<v Speaker 2>for a half hour. But the reality is nuclear weapons.

722
00:36:18.960 --> 00:36:22.119
<v Speaker 2>A nuclear weapon was used to stop World War two. Great,

723
00:36:22.400 --> 00:36:26.159
<v Speaker 2>it can't be used again. It can't. I mean, the

724
00:36:26.239 --> 00:36:29.960
<v Speaker 2>thread of it is useful. Yeah, once a nuke goes off,

725
00:36:30.400 --> 00:36:35.679
<v Speaker 2>game over man. Right, Wait, that was from Aliens, now

726
00:36:35.800 --> 00:36:36.880
<v Speaker 2>that you're mixing.

727
00:36:36.679 --> 00:36:41.239
<v Speaker 6>All as I thought that was apocalypse now, all right, Well, okay,

728
00:36:41.000 --> 00:36:45.119
<v Speaker 6>so I think it's fundamentally stupid to try to put

729
00:36:45.199 --> 00:36:47.199
<v Speaker 6>AI in charge of nuclear There's no need.

730
00:36:47.280 --> 00:36:48.880
<v Speaker 3>No, we talked about this before.

731
00:36:48.920 --> 00:36:49.719
<v Speaker 2>There's dangerous.

732
00:36:49.760 --> 00:36:55.559
<v Speaker 3>There's this fallacy that we need instantaneous decision making, right,

733
00:36:55.639 --> 00:36:59.079
<v Speaker 3>and we need quick decision making, but not so quick

734
00:36:59.119 --> 00:37:02.400
<v Speaker 3>that we can't have a human in the loop. And

735
00:37:02.480 --> 00:37:06.280
<v Speaker 3>so I don't believe the US or any other major

736
00:37:06.320 --> 00:37:10.159
<v Speaker 3>power that has nukes is going to let AI make

737
00:37:10.199 --> 00:37:10.760
<v Speaker 3>the decision.

738
00:37:10.800 --> 00:37:10.920
<v Speaker 5>Now.

739
00:37:11.000 --> 00:37:11.960
<v Speaker 2>No, I don't think so either.

740
00:37:12.280 --> 00:37:14.719
<v Speaker 3>I also don't think I'm going to have a you know,

741
00:37:14.760 --> 00:37:17.960
<v Speaker 3>a doorbell camera that's going to decide whether it uses

742
00:37:17.960 --> 00:37:21.119
<v Speaker 3>a machine gun or turns on the speaker to talk

743
00:37:21.159 --> 00:37:23.440
<v Speaker 3>to the person at the door based on a you know,

744
00:37:23.519 --> 00:37:28.480
<v Speaker 3>facial recognition. There's just fundamentally stupid applications and this is

745
00:37:28.519 --> 00:37:32.000
<v Speaker 3>one of them. For the process. Now there's the gray

746
00:37:32.039 --> 00:37:34.400
<v Speaker 3>areas are actually probably much more interesting to talk about

747
00:37:34.440 --> 00:37:37.639
<v Speaker 3>because they're much more likely to happen, like having a

748
00:37:37.679 --> 00:37:43.119
<v Speaker 3>fighter identify a target and kill it without any any intervention.

749
00:37:43.719 --> 00:37:45.599
<v Speaker 3>But I think you're going to have fighters that are said,

750
00:37:45.679 --> 00:37:48.119
<v Speaker 3>go to this area and kill anything you see that

751
00:37:48.199 --> 00:37:51.159
<v Speaker 3>matches this description. And we're already doing that with switchblade

752
00:37:51.239 --> 00:37:54.320
<v Speaker 3>drones and some of the other technologies out there. That's

753
00:37:54.360 --> 00:37:57.320
<v Speaker 3>where more of a gray area. But but this whole

754
00:37:57.440 --> 00:38:00.599
<v Speaker 3>like off the chain, AI, I think, I think this

755
00:38:00.679 --> 00:38:04.559
<v Speaker 3>is this is illustrative of why this is a stupid idea.

756
00:38:04.280 --> 00:38:06.360
<v Speaker 2>Right sure, all right, So should we worry?

757
00:38:06.400 --> 00:38:08.800
<v Speaker 3>Patrick, I'm not that worried about it, because I think

758
00:38:08.880 --> 00:38:11.559
<v Speaker 3>I think as stupid as we are, we're smarter than this.

759
00:38:11.800 --> 00:38:14.000
<v Speaker 2>Well me worry.

760
00:38:14.719 --> 00:38:16.840
<v Speaker 3>There's plenty of things to worry about it. This isn't

761
00:38:16.960 --> 00:38:19.760
<v Speaker 3>This isn't one of them. Unless yeah, because honestly that

762
00:38:19.840 --> 00:38:22.280
<v Speaker 3>we have if the people who are in charge of

763
00:38:22.320 --> 00:38:24.480
<v Speaker 3>firing nukes are this stupid, they're going to fire nukes

764
00:38:24.519 --> 00:38:25.280
<v Speaker 3>on their own anyways.

765
00:38:25.480 --> 00:38:28.159
<v Speaker 2>Well, the fact is is that, you know, it illustrates

766
00:38:28.199 --> 00:38:32.400
<v Speaker 2>that AI doesn't think, right, it didn't. It doesn't think

767
00:38:32.519 --> 00:38:35.000
<v Speaker 2>through you know, it only thinks to.

768
00:38:35.119 --> 00:38:37.679
<v Speaker 3>Well back to my point, it's not really conscious.

769
00:38:37.599 --> 00:38:41.480
<v Speaker 2>Well of course, but it doesn't think through the whole process.

770
00:38:41.719 --> 00:38:43.119
<v Speaker 2>What's the endgame after that?

771
00:38:43.360 --> 00:38:43.599
<v Speaker 5>Right?

772
00:38:43.719 --> 00:38:46.400
<v Speaker 2>You know that that's not They only were given the

773
00:38:46.440 --> 00:38:50.079
<v Speaker 2>goal of defeating the enemy. Okay, but then you know

774
00:38:50.119 --> 00:38:53.400
<v Speaker 2>their life is one prompt exactly exactly. Yeah.

775
00:38:53.440 --> 00:38:55.599
<v Speaker 4>Well, and it is interesting you say that, Carl, because

776
00:38:55.599 --> 00:38:57.239
<v Speaker 4>they will fight for whatever the goal is. But if

777
00:38:57.239 --> 00:39:00.960
<v Speaker 4>the goal is like win, defeat the enemy and minimize

778
00:39:01.039 --> 00:39:04.039
<v Speaker 4>human collateral, you know, for the rest of time, right

779
00:39:04.119 --> 00:39:06.880
<v Speaker 4>then right, maybe maybe the outcome is different.

780
00:39:07.159 --> 00:39:10.559
<v Speaker 2>Yeah, we got to do it more smart, smart, wicked.

781
00:39:10.239 --> 00:39:12.679
<v Speaker 3>Smat Someday we may have an AI that is is

782
00:39:12.800 --> 00:39:15.280
<v Speaker 3>more compassionate than us and smarter than us. But I

783
00:39:15.559 --> 00:39:17.280
<v Speaker 3>think that's decades and decades away.

784
00:39:17.440 --> 00:39:19.480
<v Speaker 2>Just got to have it programmed by the Dalai. Lama

785
00:39:19.639 --> 00:39:24.400
<v Speaker 2>there you go, you know, Lama said my death, I

786
00:39:24.559 --> 00:39:27.639
<v Speaker 2>rushed the consciousness so I got that comb for me

787
00:39:29.719 --> 00:39:32.920
<v Speaker 2>and on that note, and his name was Carl in

788
00:39:32.960 --> 00:39:37.920
<v Speaker 2>that movie. It's okay, that's it, Thank you very much. Hey,

789
00:39:38.320 --> 00:39:40.599
<v Speaker 2>check out our Discord server RT. Twain.

790
00:39:40.880 --> 00:39:42.519
<v Speaker 4>Heck yeah, discords where it's at.

791
00:39:42.679 --> 00:39:44.480
<v Speaker 2>Discord is where it's at. We get a lot of

792
00:39:44.519 --> 00:39:45.599
<v Speaker 2>ideas from there too.

793
00:39:45.519 --> 00:39:48.440
<v Speaker 5>So Discord dot security this week dot com.

794
00:39:48.559 --> 00:39:50.639
<v Speaker 2>Right, So that's all that happened last week, and we'll

795
00:39:50.639 --> 00:39:52.400
<v Speaker 2>talk to you next week on security this week.

796
00:39:52.480 --> 00:40:04.239
<v Speaker 7>Bye bye bye, buddy.

797
00:40:00.079 --> 00:40:07.519
<v Speaker 1>Oh her
