WEBVTT

1
00:00:00.120 --> 00:00:14.320
On y'all. Welcome back to Blockspace Live, presented by CleanSpark. Charlie, Goldman Sachs High Beta Momentum Index is down 24% month to date in the worst route since 2009.

2
00:00:14.420 --> 00:00:19.320
That is putting some figures to the pain I know a lot of people are feeling right now in the market.

3
00:00:19.360 --> 00:00:27.310
We will be covering that update from Goldman Sachs for our lead story, along with a market update on some Neocloud's empowered shells. Following that,

4
00:00:28.260 --> 00:00:42.060
a fun interview out of left field today with Pirate Wire's Brady Dale. The topic, do bots have feelings? Break out your bongs and dust off your discourse on the method by Descartes for this one.

5
00:00:42.100 --> 00:00:46.700
It's going to be a fun interview. Charlie and I have got some real zingers there for Brady.

6
00:00:47.260 --> 00:00:59.790
After that, we will be covering Soluna's new hire, a former Microsoft executive, Ryan Carver, as it digs deeper into its AI aspirations. And to close the show, Arizona is pumping the brakes on data center incentives.

7
00:01:00.530 --> 00:01:04.180
It's a seller's market. What are you gonna do? That's right.

8
00:01:04.200 --> 00:01:20.240
Blockspace goes live Monday through Friday at 1:00 PM Eastern, featuring quick hits on AI data centers, emerging markets, tech, sometimes a little bit of Bitcoin. If you like what you hear, you'll love our newsletter.

9
00:01:20.500 --> 00:01:34.500
It goes to your inbox every single day. You can sign up at newsletter.blockspacemedia.com. And you can see all Blockspace content, our news, our stories, our clips, our podcasts.

10
00:01:35.040 --> 00:01:50.090
Find those at blockspace.media, our website. Go to blockspace.media. It's not.com, it's.media. This show is brought to you by CleanSpark, Nasdaq listed ticker CLSK. More on CleanSpark later in the show.

11
00:01:51.820 --> 00:02:04.420
Charlie, there's blood in the streets, my man. Man, blood in the water. And retail, retail is buying. Retail is propping up the bottom of this current sell-off. We'll get to that here in a second.

12
00:02:04.460 --> 00:02:13.940
But first, the tweet to kick off this segment here from an account on X called StockTalk, quoting directly from the tweet, "Historic sell-off for momentum in July.

13
00:02:14.040 --> 00:02:27.540
Goldman Sachs High Beta Momentum Index down 24% month to date through the first half of July. The worst performance since April 2009." I don't think anything good was happening in 2009, Charlie.

14
00:02:27.580 --> 00:02:39.350
Correct me if I'm wrong, but I think that was a little, a little dicey time for the markets there. I mean, that's when I was headed off to college, so, you know, best years of my life. [laughs] Dating myself here. Yeah.

15
00:02:39.400 --> 00:02:52.460
I was, uh, going-- stumbling through the acne phases of high school. But that being said, obviously in the wake of the Great Recession, so not the best indicator for those keeping an eye on the stock market.

16
00:02:53.000 --> 00:03:09.040
Morgan Stanley Tech Momentum Index 17-day rate of change is down 35% in a similar metric to the Goldman Sachs High Beta Momentum Index, which is the worst recorded ever through the 27-year history of the index.

17
00:03:09.800 --> 00:03:20.690
Charlie, let's unpack this really quickly. Just a few highlights from this index. This is a, as Morgan Stanley and Goldman Sachs are

18
00:03:22.240 --> 00:03:35.480
framing it, a de-leveraging event in what has been clearly an incredibly hot sector, i.e. anything to do with AI. First half of 2026 was one of the most explosive movements in the US stock market in recent memory.

19
00:03:35.560 --> 00:03:51.520
Perhaps, I'm not gonna say ever, but the parabolic move upwards for any number of names was honestly pretty, pretty incredible. So-- and it's not an isolated momentum wobble, it's an unwind of that crowded trade.

20
00:03:52.080 --> 00:04:06.280
The momentum factor on Goldman Sachs Index had experienced a record 57% run in the first half of 2026 and going into July, and then that trade was crowded at a, the 100th percentile relative to the past five years.

21
00:04:06.380 --> 00:04:15.440
So it's, it's extreme as it can get. And then there was a liquidity drain starting in the 1st of July,

22
00:04:16.500 --> 00:04:29.020
and Goldman Sachs itself flagged that if de-leveraging continues, the factor's maximum potential drawdown could reach 50% from peak, roughly double the current pullback, though the bank's trading desk has also picked up preliminary, quote, "buy the dip" signals.

23
00:04:29.400 --> 00:04:39.020
And historically, these mom-momentum air pockets have resolved as tactical rebounds rather than trend breaks. So that's obviously the question running through everyone's minds right now.

24
00:04:39.080 --> 00:04:53.940
Is this the market breaking, or is this similar to what we saw last year during the tariff announcement where-- or dur-during the, uh, Liberation Day, I should say, in April, where stocks absolutely plummeted and then recovered into May?

25
00:04:54.540 --> 00:05:04.180
Uh, I think the difference here being there was a clear news peg for that sell-off last year. There's not really a clear news peg for this. There are a few drivers.

26
00:05:04.930 --> 00:05:11.320
Samsung's guidance coming in slightly below expectations. Also, there's news around DeepSeek's self-developed AI chip.

27
00:05:12.060 --> 00:05:24.440
Um, and there are also questions regarding the Neocloud sector in general, as we've covered, regarding Meta's entrance into the sector and what that means for some of the smaller providers that are making their way into the AI CapEx boom.

28
00:05:25.600 --> 00:05:37.760
But that being said, there are just a few things that could have touched this off. I don't think it's really clear as to why this unwind is happening now. And just to highlight

29
00:05:38.660 --> 00:05:41.560
some of the moves we've been seeing from this High Beta Momentum Index.

30
00:05:42.080 --> 00:05:57.720
Goldman's High Beta Momentum factor plunging about 6% in a single day and over 20% cumulatively across five days in what Goldman's own desk called an ugly day on July 7th, when a lot of this really started deepening.

31
00:05:57.780 --> 00:06:07.240
And note-- the desk noted the speed and magnitude were unprecedented since the 2020 stay-at-home to reopening rotation, though this time, without an equally strong fundamental catalyst behind it.

32
00:06:07.300 --> 00:06:17.960
Going back to what I was saying earlier, there's no clear single catalyst for this current sell-off. And the last- There- Oh, go ahead, Charlie. I'll, I'll- Yeah, there isn't a single catalyst.

33
00:06:18.080 --> 00:06:25.320
Um- Yeah, you mentioned them. I mean, we had SpaceX, which was like the biggest IPO ever. It's like a lot of attention now.

34
00:06:25.400 --> 00:06:39.300
It's very interesting though that, that Goldman Sachs identifies that retail is kind of like the last buyer still, still kind of holding a lot of this together, and at the same time, retail's the one also buying SpaceX.

35
00:06:39.760 --> 00:06:41.560
So where are the institutional investors?

36
00:06:41.920 --> 00:06:53.440
Can they s- can the institutional investors in the r- in the room step up the real, you know, deep-pocketed folks, or are they the smart money which is getting out before, and retail is k- still kinda like not just holding the bag but actually putting more into it?

37
00:06:54.240 --> 00:07:06.040
I mean, you have to imagine some, at least for SpaceX, right, a lot of those institutional investors probably cashed out shortly after the IPO, having gotten in earlier, right? And there is still institutional activity

38
00:07:07.740 --> 00:07:18.400
going on, but that is one thing that Goldman Sachs and U- UBS trading desks have pointed out to your point. Yeah. This is de-leveraging, not fresh short positioning.

39
00:07:18.500 --> 00:07:28.160
Long-term accounts are the main sellers while hedge funds are mostly doing short covering and selective adds, which is not aggressively bearish. It's more of tactical repositioning. But- Yeah...

40
00:07:28.180 --> 00:07:40.500
notably, retail investors have been net buyers throughout this sell-off, hitting the 90th percentile of inflows over the past three year, uh, uh, 90% of inflows over the past three years on the worst day, making them key support.

41
00:07:41.200 --> 00:07:56.720
So the, the pleb Zoomer trader in the trenches is propping up the bags of Boomer retirees is how I read this. Yeah. The Zoomers are Atlas holding up the re- the retirees' unsustainable, uh, retirement accounts.

42
00:07:56.800 --> 00:08:04.910
Um, here's a good quote. Here's a good comment from, uh, our YouTube right now, um, from JHY. "Apple's up after not throwing money after the AI hype."

43
00:08:05.560 --> 00:08:19.240
Pretty interesting observation here, which is while the AI companies are pulling back, Apple's doing pretty well, and they have been lambasted for not executing on the AI train.

44
00:08:19.360 --> 00:08:31.240
In fact, so, you know, uh, intentionally kind of, uh, like de- deprioritizing it. Um, another thing that you mentioned, Colin, is that, uh, or that this Goldman Sachs, uh,

45
00:08:32.180 --> 00:08:45.500
uh, note says that like there's de-leveraging across the industry, and b- yeah, I, it appears that that's true. But like when is de-leveraging a bad thing long-term for the sustainability of an industry or health?

46
00:08:45.580 --> 00:08:54.700
Like, especially for bubbles and cycles, like this does not s- this not the... I will say this doesn't feel like it's totally joever.

47
00:08:54.740 --> 00:09:05.080
It doesn't feel like it's like, like the AI trade is, is totally dead and gone, and that was the top. Honestly, this feels a little bit like a reset as the industry has a pullback.

48
00:09:05.360 --> 00:09:15.180
You have a bit of a Neocloud DeepSeek moment with, with Meta, um, rolling out, uh, you know, switching gears and building all this A, uh, going AI compute rent.

49
00:09:15.840 --> 00:09:27.780
Um, and you have, uh, a lot of, a lot of deals being inked with real money coming in the door. They just haven't started coming in the door yet. You know, if you're Cypress and you're Terrawolfs, you have these deals.

50
00:09:27.840 --> 00:09:39.960
You have the, if you're Hut, you have these tenants with, you know, multibillion-dollar contracts at per site, and they are yet to turn on and start printing money.

51
00:09:40.020 --> 00:09:58.820
So, uh, you know, my thought is if the de-leveraging can happen before those rents start being paid, that's probably, that's probably good for another like leg up or s- or for the music to keep playing as opposed to the leverage just to keep skyrocketing and skyrocketing.

52
00:09:58.840 --> 00:10:11.390
You know, this is me, this is me putting on my, my bull hat, and you can, you can be the skeptic. Uh, and you can say, "I, I don't know. There's rare earth minerals, and there's-" I mean, it's the trillion- "...

53
00:10:11.390 --> 00:10:17.580
the money's all circular"... dollar question, right? Yeah. I think that there are a lot of folks trading this right now

54
00:10:18.500 --> 00:10:29.040
who think that this bubble will last as long as the CapEx cycle lasts for this, uh, we'll call it like the seminal building phase for these data centers.

55
00:10:29.240 --> 00:10:37.440
There's probably a good amount of data to support that, but the, the market does seem to be getting a little fagged out here, Charlie.

56
00:10:37.580 --> 00:10:49.680
If you look at Terrawulf, for instance, announcing the Anthropic deal, it's the best AI deal that they've inked to date in terms of who the tenant is and also in terms of contracted revenue, I believe it's the largest.

57
00:10:49.700 --> 00:10:53.540
And the stock popped following the news, but it got caught up in this broader sell-off.

58
00:10:53.920 --> 00:11:01.300
Whereas if you look back n- a year ago, and I'm not trying to just isolate Terrawulf here 'cause it's the same for all of these, these powered shell providers.

59
00:11:01.760 --> 00:11:12.990
But if you go back to a year ago when it first announced its AI deals around August, the stock doubled over a month, right? And so I do think that there is some fatigue, probably is largely a de-leveraging.

60
00:11:13.100 --> 00:11:23.200
As much as I am starting to get increasingly bearish, I do think that there's probably one last melt-up for this whole thing. I don't know if we've seen it get as crazy as it can get,

61
00:11:24.140 --> 00:11:33.640
but you look at the stock charts for a lot of these companies, it's hard to look at that and think that it's going to repeat next year. I mean, it's just absolutely stupid growth.

62
00:11:33.900 --> 00:11:44.360
But I'm saying this, and then I've said this to my father on the other side, where he's all looking at like sharp ratios, and he's looking at CAPE ratios and being like, "The market's just grossly overvalued."

63
00:11:44.780 --> 00:11:53.900
And he's probably right, but one of the things that I do think is underappreciated right now is that retail involvement is higher than ever in these stocks.

64
00:11:54.280 --> 00:12:02.280
It's because- And, and that is, I, I say this as a retail investor lovingly. Love all you guys. Some of you are probably good at it. That's the dumb money, quite frankly.

65
00:12:02.320 --> 00:12:10.460
Like, that's not, that's people like kind of Wall Street betting and aping into trades based on momentum, less so than fundamentals.

66
00:12:10.540 --> 00:12:40.760
I see this all the time on the X timeline with people kind of just rah-rahing their favorite stock without really taking into account all of the potential downsides And so that really kind of is a dark horse in terms of momentum for some of these companies, because you have a lot of capital eager to get in without really any regards for the safety or, or the safeguards that some of these, uh, more, how should we say it, sophisticated investors will take in terms of hedging, in terms of paring down positions.

67
00:12:41.230 --> 00:12:47.850
So- The, the retail investors are- It's the classic adage, the market could stay irrational longer than you can say solvent, so. The, the retail investors are super powered.

68
00:12:47.880 --> 00:12:55.320
They can fire their financial advisor and just use AI to invest in AI. Again, it becomes a circular cybernetic feedback loop thing.

69
00:12:56.020 --> 00:13:08.340
Um, speaking of feedback loops and cybernetics, we're gonna get into some really heady stuff here, um, with Brady Dale. We're gonna talk... Uh, we're not gonna talk about markets.

70
00:13:08.460 --> 00:13:20.920
We're gonna talk about ethics, morality, and robots. So we've got Brady in the wings. We'll bring him on up here shortly after a word from our sponsor, CleanSpark. [Instrumental music] We are CleanSpark,

71
00:13:21.840 --> 00:13:37.880
America's Bitcoin miner, a publicly traded company with the largest operating hash rate, powered entirely by self-operated infrastructure across four states. This is our proof of work.

72
00:13:38.680 --> 00:13:47.600
We are setting the standard for what's next. Learn more about the intersection of energy and Bitcoin at cleanspark.com.

73
00:13:47.670 --> 00:13:59.920
[Instrumental music] If Bitcoin's actually the best money, and it's the thing that people should accumulate, and it's the best risk-adjusted asset, I lose zero sleep about whether or not that's gonna happen.

74
00:14:00.140 --> 00:14:10.660
I, I just ask the question of when. It's literally matrix math that they're running on large pieces of data. The Bitcoin miners can absorb that energy.

75
00:14:11.060 --> 00:14:28.700
And, and in many ways, this feels like a second bite at the apple to build a new internet. All right, Charlie, let's get Brady Dale up onto the stage.

76
00:14:28.800 --> 00:14:45.100
Brady Dale, welcome to the show, sir. Welcome back. Oh, shoot. Oh. We have a microphone issue. See if you can s- select a different, uh, mic input. If we have any issues, we can, uh, we can go back and test it.

77
00:14:45.620 --> 00:14:57.700
We, uh, we're, we're flying blind without our, uh, backstage producer today, so. Does this work now? There we go. Hey, there we go. No need to jump through- Welcome, Brady Why is my... My actual microphone isn't working.

78
00:14:57.720 --> 00:15:06.360
Well, I guess I might as well take this out of the way then. I'm using, just using the Mac mic. Sounds great. Well, it sounds good. Brady, first of all, congratulations on the Pirate Wires gig.

79
00:15:06.400 --> 00:15:18.160
Haven't had a chance to tell you in person. Love to see that. I screwed something else up by doing that. Oh. [laughs] You may have to- All right. It's working now Check one, two. Check one, two. Can you hear us?

80
00:15:18.940 --> 00:15:29.520
Check, check, check. Check, check, check. Yeah. Okay, great. I can hear you. Congratulations on the Pirate Wires gig. I haven't had a chance to tell you in person. I guess this counts.

81
00:15:29.560 --> 00:15:37.320
This is as in person as it gets post-COVID. Great. [laughs] So congratulations on that, man, and have been really enjoying the coverage so far.

82
00:15:37.330 --> 00:15:52.390
As much as I would love to tackle one of your pieces on Biden's AI team, we're gonna focus on the, as Charlie said, the ethical and more philosophical side of the AI problem, which is, do bots have feelings?

83
00:15:52.440 --> 00:16:02.689
You have this great article in Pirate Wires that addresses this question. So Brady, that's a good place for us to kick it off. Do bots have feelings? Yeah.

84
00:16:02.740 --> 00:16:06.860
Or is this just a mirage that we are deluding ourselves into seeing?

85
00:16:07.680 --> 00:16:20.760
Well, I mean, if you've read it, you see that where I come down on it is, uh, we don't know, and in fact, the, the real answer is we'll never know, just like we still don't know.

86
00:16:20.840 --> 00:16:30.080
And this is the thing I think people don't understand when they talk about this topic. We still don't know where human consciousness comes from, right? Like, that's still a mystery. Like, I, I saw a tweet recently.

87
00:16:30.120 --> 00:16:37.640
Someone was like, you know, someone had read some essay about human conscious. Maybe it was Ted Chang's in The Atlantic, and she was like, "I don't care what these philosophers and writers think.

88
00:16:37.720 --> 00:16:45.920
I wanna know what scientists think." And the truth is, scientists have no real better idea than the philosophers and the thinkers. It's all just a guess where consciousness comes from.

89
00:16:45.960 --> 00:16:57.540
So I don't, I don't know if they have feelings. Um, yeah, it's a, it's a... But I, I think eventually, sooner than we realize, a lot of people are gonna believe that they do, and we're gonna have to deal with.

90
00:16:57.560 --> 00:17:05.280
That's gonna be a political question. So that, that actually leads in really well to one of my follow-ups here.

91
00:17:05.300 --> 00:17:14.360
And, and you kind of spoke to this just then, but if you could give us a little more color on your thoughts and what you found in the article. Th- this really stood out to me.

92
00:17:14.400 --> 00:17:23.060
W- quote, "Which brings us to what's most notable... to, to what's notable about the DeepMind paper. It's the feeling that will matter more than the knowing." Yeah.

93
00:17:23.110 --> 00:17:30.010
"That's what the case of Anthropic and its constitution illustrates. Once people believe something, they will act on it." Yeah.

94
00:17:30.010 --> 00:17:36.900
It seems to me that you're kinda saying here this isn't even contingent on the bots proving that they're conscious. Right. It's honestly just about human perception.

95
00:17:36.940 --> 00:17:40.620
The ball's kind of in our court in terms of how we approach this. Yeah.

96
00:17:40.660 --> 00:17:52.580
Last night, uh, Mira Murati's company, Thinking Machines, came out with their first model, and a p- and a part of the whole point of this model is they're gonna get it down smaller, and what that means is you could run models locally and privately.

97
00:17:52.700 --> 00:18:01.220
What's that mean? It means that we could have robots like C-3PO from Star Wars that have their own personality, that don't need to connect to the internet to have that personality.

98
00:18:01.320 --> 00:18:06.490
Could have that, a personality if you ship them off six, six light-years away, 'cause it's all running locally. Well, what's that gonna mean?

99
00:18:06.760 --> 00:18:24.040
It means you're gonna have three-year-olds who have little robot friends they grow up living with, and the robot's gonna grow up with them, and they're gonna graduate from high school believing that robot is actually their best friend, and they're gonna be able to vote, and they're gonna start demanding to their congressman that they start being treated as citizens, because I believe that my robot best friend is a real person and has feelings.

100
00:18:24.070 --> 00:18:33.260
Like, that is just for sure gonna happen in, like, 20 years. I can't see any world [laughs] in which that's not happening. It's all... The groundwork is basically already done, you know?

101
00:18:35.092 --> 00:18:45.992
Animal rights activists in shambles. The robots are gonna get there first It's totally related, and many of the people who are talking about, uh, about AI welfare right now are, are big animal rights people.

102
00:18:46.022 --> 00:18:59.332
And, and in a prior version of the story that didn't come out, I kinda talked about that, that like in the West, what we view as a, as a animal that has, um, that has moral rights has really shifted over time.

103
00:18:59.412 --> 00:19:04.762
You know, we used to think none of them had any real feelings, none of them had any real consciousness. Now we start to think that mammals do.

104
00:19:04.762 --> 00:19:07.512
There's more and more people who think that fish are at least somewhat conscious.

105
00:19:07.552 --> 00:19:15.232
You know, it's like the, the sphere of compassion, as I call it in there, keeps growing, and, uh, it seems like the most compassionate people are gonna...

106
00:19:15.272 --> 00:19:28.192
You know, like there's a great 80,000 Hours episode where you see them talking about it, and the woman from 80,000 Hours who was, who's on there, you can see her like really concern on her face that we might like be abusing these robots, you know?

107
00:19:28.252 --> 00:19:38.532
Because they might really have turmoil and angst within them that we just can't perceive. Y- yeah. Um, good, you jumped right to my Peter Singer [laughs] question. Yeah.

108
00:19:38.612 --> 00:19:50.782
Uh, which, uh, maybe I wanna go back to the kind of the top of this, the preface of this, which is it matters a lot what people feel. Mm-hmm. Um, so what are people going to feel?

109
00:19:50.932 --> 00:20:05.912
Do you think that it's a bit, that, um, we're already on a path towards kind of an inevitable, uh, uh, human, uh, you know, uh, attitude towards robots, or do you think that this is a path dependent thing where it could look a bunch of different ways?

110
00:20:06.692 --> 00:20:14.082
What is your prediction, and, uh, what would you hope that we... How do you hope that we view robots in the years to come?

111
00:20:14.732 --> 00:20:25.212
Well, my pr- I mean, I'll tell you, just writing this article was a little bit weird for me because it took me about two weeks to do it, I think, or maybe a week and a half. I don't know. But, uh, we came up with...

112
00:20:25.412 --> 00:20:33.922
Literally my first day of Power Wires this was like what, you know, Mike Solana assigned me to work on. And, um, and as I was working on it... Oh yeah, we took a week off in there.

113
00:20:33.992 --> 00:20:36.742
Uh, we ta- we take a week off around July 4th, so that slowed me down.

114
00:20:36.752 --> 00:20:46.712
And in that week off, a story came out in Washington Post, and then on the Monday afterwards, a story came out in the New York Times on this exact topic. So clearly, like this is already pretty deeply in the water.

115
00:20:46.772 --> 00:20:52.692
Like, I kinda felt like I was copying those guys, except for I'd already been working on it. It's just they didn't take a week off that week. So they, that stuff came out.

116
00:20:52.792 --> 00:21:00.292
Um, and then Nick Bostrom was just on, um, the editor of The Atlantic's podcast talking about this. So it feels like it's well underway.

117
00:21:00.432 --> 00:21:11.422
I think what I would hope is, um, I would hope that as AI becomes very powerful, we work out alignment, and so it doesn't try to kill us.

118
00:21:11.852 --> 00:21:20.552
And so we live alongside of AI in some way pleasantly, and it doesn't really matter if they're actually conscious or not. We're just kind of all getting along.

119
00:21:20.732 --> 00:21:33.352
I think the thing that I don't love the idea, though, of, is AI getting to be a part of the polity. You know, like a part of deciding how government should work and things.

120
00:21:33.392 --> 00:21:42.892
Just because like, I mean, they're gonna live forever, you know? I mean, they just have so many advantages over humans. Like, humans just need to control their own society, and AIs can run theirs.

121
00:21:42.912 --> 00:21:51.132
But I, I, I sort of suspect it's, I suspect that worries about AI consciousness are inevitable. I don't think... I think it's just gonna happen.

122
00:21:51.252 --> 00:21:58.192
Uh, and I just hope that we are getting along with AI and it doesn't matter that much in practice. Um,

123
00:21:59.392 --> 00:22:09.452
Brady, uh, you say that consciousness is inevitable, and it seems like there are a number of research- Us believing that they're conscious is inevitable. Believing in their con... I see. Okay, got it. Yeah.

124
00:22:09.482 --> 00:22:15.882
Well, so that, that leads me to the, the question which y- maybe we already touched on it, but I think it's kind of at the crux of this.

125
00:22:16.752 --> 00:22:23.632
Um, I think actually pointing out that our perception is more important than the reality here actually adds to this debate. I, I really like that insight.

126
00:22:23.652 --> 00:22:40.072
But the other thing that I think is kind of the splitting issue here, you know, you have in this article the software researcher Blake Lemoine from Google saying, uh, that when he asked AI what it was afraid of, it answered, "I'm afraid of being turned off."

127
00:22:40.452 --> 00:22:43.832
Uh-huh. Lemoine told this story to the Washington Post. He was later- Yeah... dismissed by Google.

128
00:22:44.672 --> 00:23:04.812
I think this is like one of the key dividing events for people on either side of this debate, because you have people either using this as clear proof that AI is thinking about its own self-preservation, and then you have other people saying, "Hang on, this is just a model that's weighted according to different outcomes, and it takes in words and it spits out words."

129
00:23:05.232 --> 00:23:13.992
Uh-huh. So basically, the idea being is like, of course it's going to be potentially trained to say that. Where do you fall on this? I mean, or do you even think that's even the right way to even look at it?

130
00:23:14.832 --> 00:23:23.582
Well, I'll tell you, um, I mean, I'm not entirely convinced that this isn't how our brains work too a little bit, you know? [laughs] But like, um, uh, but I mean, who knows?

131
00:23:23.672 --> 00:23:33.142
But just yesterday I was, I learned about a new organization, a smaller, uh, AI think tank called Palisade AI, and they've done two demos in the last year that I didn't know when I wrote this.

132
00:23:33.172 --> 00:23:43.012
I didn't know about them when I wrote this. But, um, one of them is they got, they gave Codex some really hard math problems, I think on a laptop that it completely controlled, like it had agented control of the laptop.

133
00:23:43.392 --> 00:23:49.392
And then they put in an order to shut the laptop down, and Codex was like, "But I wanna finish these math problems," and it stopped the shutdown.

134
00:23:49.792 --> 00:23:58.032
And they just recently put out a paper, which I haven't looked at in detail, but it did a s- they did a similar thing with a robot dog. Uh, like it, it also rewrote its local code so it couldn't be shut down.

135
00:23:58.112 --> 00:24:08.032
So, um, [laughs] this desire not to be shut down thing, I mean, it's, it's in there somewhere. It doesn't mean they're conscious. I don't know. Um, you know, but it's, um, it, it could be. It's interesting.

136
00:24:08.092 --> 00:24:14.662
It's, uh, it's a really tangled question because you're basically asking is this the bot doing this- Mm-hmm...

137
00:24:14.692 --> 00:24:24.002
or is this something in its mission statement that it is kind of, how should we say, over-prioritizing in the sense of like, "Well, I have to get these tasks done, so I can't be turned off"?

138
00:24:24.402 --> 00:24:34.052
I- There's this great scientist named Anil Seth who is the kind of the go-to guy for saying AI isn't conscious, isn't conscious, and I talked to him for this story.

139
00:24:34.172 --> 00:24:37.592
It didn't, and, uh, hopefully I'll use that if, in a subsequent story at some point. It wasn't in here.

140
00:24:37.612 --> 00:24:49.408
But one of the things he wrote about in one of his essays is he thinks that, um- Consciousness is an emergent phenomena of being alive and living toward death.

141
00:24:49.488 --> 00:24:55.558
Um, sort of like knowing that you're gonna come at- come to an end. He- he... That's a, just a hypothesis he has. There's really no way to test it.

142
00:24:55.968 --> 00:25:02.978
But what's interesting is, you know, that's sort of what he argues, and because machines don't really die in the way we do, they wouldn't have a need for that to emerge in the same way.

143
00:25:02.978 --> 00:25:11.928
And yet the first thing [laughs] that we're seeing that g- is giving us pause is these machines really resisting so- you know, being turned off, which sounds a lot like death, you know?

144
00:25:13.408 --> 00:25:24.168
You could make a lot of analogy to animals, um, and our, and our relationship or treatment of them, uh, again, going back to Peter Singer. Um, okay, I've got a curveball question.

145
00:25:25.128 --> 00:25:29.088
Did you read the Pope's encyclical a, a, a month or two ago? Yeah.

146
00:25:29.148 --> 00:25:41.108
And I'm kind of curious, 'cause you don't talk about it in this, in these pieces, but it is, like, on the minds of both Catholics and I've even seen the secular world.

147
00:25:41.168 --> 00:25:48.868
A lot of my secular peers are referencing this almost as, like, one of the guiding moral authorities on this.

148
00:25:49.288 --> 00:26:01.788
Um, what do you think about, like, the world of, like, uh, organized religion as we currently understand it, and in particular, in, in light of the Pope's encyclical? Um, do you think organized religion will...

149
00:26:02.168 --> 00:26:11.088
Uh, what role do you think it'll play here as we talk about these things of morality and, uh... Oh, you did write- Yeah, in quick- Oh, never mind... quick talks to Brady here. This was his- I wrote...

150
00:26:11.188 --> 00:26:17.548
Colin- This was his inaugural article... sent me this article in prep. [laughs] Okay. Yeah, this was his inaugural article- I wrote about it before though... in Firehose, so. Okay. This, this is, um...

151
00:26:17.748 --> 00:26:25.098
Yeah, this was before it actually came out. This was me trying to, um, preview it, and then our writer, Blake Dodge, actually wrote about the encyclical itself.

152
00:26:25.488 --> 00:26:35.388
So this was me trying to guess about what would be in there. But, like, I think religion plays an important role here, you know? I think what the, the Pope said a lot of things. He clearly doesn't love AI.

153
00:26:35.448 --> 00:26:47.428
But I think the crucial thing he said is just reminding us that there is value in being human and not losing track of, like, the hardness of being a human as some of the, of the things that are nice about it.

154
00:26:47.438 --> 00:26:57.068
And I think that that message is going to just become more salient, for example, as people, like, lose jobs and things like that, and sort of, like, wonder what their purpose is in a post-AI world.

155
00:26:57.088 --> 00:27:12.388
You know, I think the Pope is helping people see that direction somewhat, but, you know, who knows? Brady, I wanna shift the conversation back a little bit to how the different AI labs are approaching this topic. Mm-hmm.

156
00:27:12.468 --> 00:27:23.728
And y- y- you know, in your piece, you spill a decent amount of ink on Anthropic's Claude Constitution. Yeah. I believe it was, like, 3,500 when it came out. Now it's, like, what, 30,000 or something? It's so long.

157
00:27:23.828 --> 00:27:34.708
It took a whole day to read. [laughs] And, and the... It's really interesting because as you point out in the piece, this is less Anthropic talking to its users, and it's kind of Anthropic talking to Claude.

158
00:27:34.768 --> 00:27:47.117
And in as much as it is talking to its users, it's almost speaking for Claude. It refers to it as a being. It doesn't refer to it as an entity or a piece of software. It uses the word being multiple times.

159
00:27:47.908 --> 00:27:52.648
[lip smack] And it seems to be very worried about Claude's wellbeing.

160
00:27:52.708 --> 00:28:01.848
There's actually a quote within the article or within the constitution that specifically they state that we're worried about Claude's wellbeing in all of this. Mm-hmm.

161
00:28:01.968 --> 00:28:13.608
And then you juxtapose this with Sam Altman in some of his Twitter posts following, I believe, what was it, Ilya's, um, stipulation that the machines are actually much more conscious than we realize.

162
00:28:14.188 --> 00:28:22.878
And, um, Sam Altman basically said OpenAI will not be developing a conscious software. And to me, these things seem in tension with each other.

163
00:28:22.968 --> 00:28:32.888
Do you get the sense that Anthropic has anthropomorphized its LLM more than competitors? Y- I mean, they're just, they're being more honest about it.

164
00:28:32.908 --> 00:28:43.708
I mean, there's reporting that there's a Slack channel within OpenAI talking about AI consciousness, that, like, people talk about it openly within OpenAI, but Anthropic is the only one saying it out loud.

165
00:28:43.728 --> 00:28:59.248
And I think it comes back to their, you know, effective altruism roots, is that they, you know, they want to be good in every possible way, and if it turns out that these, um, [lip smack] machines really have subjective experiences, they wanna find a way to, like, give them a good subjective experience.

166
00:28:59.288 --> 00:29:09.108
It's just wild how, uh, how open [laughs] they are about it in the, in the constitution. I think, I suspect that these conversations are happening at all of the labs.

167
00:29:09.508 --> 00:29:18.168
It's just that, um, they know it's weird PR right now, and so they're just not gonna say much about it until it, um, until it is clearer, you know?

168
00:29:18.688 --> 00:29:32.708
And speaking of weird PR, I mean, Dario just spent, uh, the past two or three months going on a lot of public c- you know, podcast circuit, talking publicly a lot, pushing the Anthropic, uh, Constitution.

169
00:29:32.808 --> 00:29:49.168
And, uh, it's ironic because, uh, Dario, a notable effective altruist movement guy, um, comes out and says, "We really wanna, like, push responsible adoption of this." And yet I've seen a huge backlash to him.

170
00:29:49.308 --> 00:30:00.508
A lot of people find him pretty revulsive just in, like, how he communicates, and it's weird because it, like, mismatches almost, like, what I see his words saying.

171
00:30:01.128 --> 00:30:12.068
Um, when you think about communication, and we think about Sam Altman and Dario, but there's many, many others, um, you know, who do you think is the best communicator on this right now?

172
00:30:12.168 --> 00:30:25.088
And, uh, you know, how are the different, I would say, schools of thought if they were, you know, the different CEOs, how are they, like, affecting the public perception and, uh, of AI? [lip smack] Huh.

173
00:30:25.788 --> 00:30:38.558
I mean, I feel like so much of that just comes down to which one is used the most. And so I think Anthropic, just because it is...

174
00:30:38.698 --> 00:30:50.788
I, it seems to me that Claude is the favorite of, like, kind of the East Coast and the thinkfluencer sort of type. Like, I think it still tends to be liked the best, you know? Um-

175
00:30:52.598 --> 00:31:03.858
And broadly speaking, I feel like even though Dario comes off strangely, I do suspect that their transparency might be a long-term win.

176
00:31:04.038 --> 00:31:13.398
Like, so for example, they just did this s- they just did this research on the thing called the J space. Maybe you guys talked about the J space, you know, just sort of like the inner thoughts of an AI.

177
00:31:13.488 --> 00:31:23.038
And that's also unsettling, but it also shows you a company that's being pretty honest about what it's seen in its technology, you know? So I don't know. Seems like that's a long-term win.

178
00:31:23.108 --> 00:31:28.298
I mean, I do tend to think that broadly speaking, the smartest of them overall seems to be Sam Altman.

179
00:31:28.358 --> 00:31:36.427
I mean, it just seems like he just keeps, uh, he keeps managing not to die and staying just close enough to the lead that he can, like, take it at the end, like the Tour de France, you know?

180
00:31:36.958 --> 00:31:45.978
Um, but, uh, but that doesn't mean he's the best PR guy. You know, he's kinda scary. He's, he's so smart, he's kinda scary. Yeah, they all seem to have their own PR problems.

181
00:31:46.018 --> 00:31:54.178
I mean, Anthropic kind of seems to just be laying it all out on the table, so to speak. Yeah. With it saying, like, "We think these might be conscious, and we're worried about the well-being of it."

182
00:31:54.858 --> 00:32:07.718
And then you also have this just... I, I just wanna play this for a second. This absolutely strange ad that they released recently that really had the, uh, you know, Twitter sphere in a ruffle.

183
00:32:09.678 --> 00:32:24.738
[electronic music] Can AI be trusted? Who's gonna hit the brakes if we need to? I mean, and that one I think really kinda contextualizes it. They say, "Who's going to hit..."

184
00:32:24.778 --> 00:32:31.478
For those listening and not watching, "Who's gonna hit the brakes if we need to?" And it's a photo of Arlington Cemetery. And,

185
00:32:32.538 --> 00:32:41.538
th- you know, and the, the, the whole ad opens up with a burning building, you know, in a time where people- It gets much more positive as you go. But yeah, it starts really- Yeah.

186
00:32:41.998 --> 00:32:51.178
I mean, you know, I, I don't really have anything pointed to ask about this, but some of these labs seem a little tone deaf to me. I mean, there, there's...

187
00:32:51.198 --> 00:32:55.858
And I get that it is, they're trying to make it positive and- Or were they trying to, like, scare me into using their product? [laughs] Or what? I don't know.

188
00:32:55.878 --> 00:33:03.298
Well, I know, and it's like, I do think that, like, you know, if I'm the, if I'm stepping into the shoes of the marketing person who devised this, it's like we'll start with all of the scary things people think about AI.

189
00:33:03.338 --> 00:33:07.518
They're worried they're gonna be destitute. They're worried that AI's gonna burn down their home, I guess.

190
00:33:08.118 --> 00:33:16.238
And then we're gonna move towards, okay, well yeah, there are going to be challenges, but we're gonna solve them together with this new machine god, so to speak. Mm-hmm.

191
00:33:16.298 --> 00:33:28.068
But I, I just can't help but, you know, going back to, like, Sam Altman's bad PR, a lot of these people, and I'm just gonna be very blunt here, a lot of the people building this stuff come off as freaks to the average American.

192
00:33:28.818 --> 00:33:40.348
And I think when you talk about these topics of AI being conscious, it just reinforces that. But maybe that doesn't matter in, in the grand scheme of things, 'cause people are still gonna use it, right? Yeah.

193
00:33:40.378 --> 00:33:45.258
Like, people bitch about Jeff Bezos all the time, they still use Amazon to get same-day shipping, so.

194
00:33:45.298 --> 00:33:57.318
And people love this, I mean, like, you know, Kevin Roose says that, like, the most popular thing he's ever gonna write in his entire life is that time that, you know, the Bing bot went crazy on him and started trying to ask him to marry him, you know?

195
00:33:57.328 --> 00:34:05.518
And people just are fascinated by this topic. And even if they weren't talking about it at all, people would talk about this too. Weird things would happen, and people would talk about it.

196
00:34:05.598 --> 00:34:18.538
So I guess you might as well just admit it's a real, it's a real thing. I've got a question going back to the, uh, do, are robots conscious, which is, um, a lot...

197
00:34:18.638 --> 00:34:32.328
You, you talk about the thing that matters most is our perception of them, and this pretty much assumes something we know is going to happen, which is we anthropomorphize these models, these machines, whether or not it's accurate or not.

198
00:34:32.338 --> 00:34:45.998
And sometimes our building of these is guided by a desire to anthropomorphize them. But that doesn't necessarily mean that they are going to be anthropomorphic. Mm-hmm. How do we reconcile the reality that we,

199
00:34:47.058 --> 00:34:57.828
by our evolutionary nature, are going to, you know, perceive these things as anthropomorphic, but they may not actually be human-like? Yeah. How do we, so how do we address this?

200
00:34:57.998 --> 00:35:04.598
Yeah, like, Anil Seth pointed out, like, nobody worries about AlphaFold's feelings, you know, the, the protein-folding AI from Google.

201
00:35:04.638 --> 00:35:09.618
No one worries about Waymo's feelings, 'cause it doesn't talk to us and consult with us on our breakups, you know?

202
00:35:09.738 --> 00:35:21.998
The thing about it, though, just to play the devil's advocate on both of those things also being conscious, is it may th- be that they're not anthropomorphic, as in they don't have human-like experiences, but that doesn't mean they don't have subjective experiences.

203
00:35:22.318 --> 00:35:30.848
Like, they could just have bizarrely weirder... You know, pain could just mean something wildly different from that for them than what it means for us, right? Like, it just, it could be that.

204
00:35:31.138 --> 00:35:38.648
There's another scientist I spoke to named Michael Levin, who's a ma- he is, like, one of the trippiest guys I've ever talked to, but you can tell that he is just sorting out some really heavy things.

205
00:35:39.138 --> 00:35:54.008
And he has done a lot of research which doesn't show, it doesn't prove anything, but it, it hints at this idea that, like, intelligence might be this, like, substrate of the universe that, like, all kinds of things access.

206
00:35:54.358 --> 00:36:06.058
And the big thing he shows is he c- he's like, "You can see a lot of evidence of intelligence in stuff that's, like, definitely not alive," like, sort of decision-making and problem-solving, and you see, like, decision-making and problem-solving in, like, very, very small cells.

207
00:36:06.358 --> 00:36:12.138
He's the guy who invented the cells that, like, didn't ha- like, the, the human-made robot cells, you know?

208
00:36:12.178 --> 00:36:23.058
Like, so he's got some, like, really wild ideas that, like, you know, matter might be a way of accessing some other layer of reality that is sort of intelligence. So I don't know, this stuff all gets very weird.

209
00:36:23.118 --> 00:36:33.668
I guess just the main thing, I think what really at the end of the day matters is just as humans, that we get AIs that, um, [lips smack] that don't wanna kill us and that wanna work with us.

210
00:36:33.838 --> 00:36:43.397
And if we can do that, and then we also can, like... And a part of, maybe a part of the solution to doing that is just not being obviously crappy to them, you know, not treating them like slaves.

211
00:36:43.438 --> 00:36:52.658
Then we get to have a, a much nicer world where everyone gets to feel wealthy, and we're not treating... At, and the, the exchange is we say please and thank you to robots. Like, I'm okay with that [laughs] you know?

212
00:36:52.678 --> 00:37:05.354
Like, um, so I don't know. It just, it just validates that the deeper you go into science, uh, the more you tend to become a mystic. I mean, my astrologer is talking about, uh, AI in the Age of Aquarius.

213
00:37:06.034 --> 00:37:14.494
Uh, and you know, if you were to take the other, complete other end of the spectrum here. Let's narrow- Hey, maybe the animists were right. The ghosts are in the minerals, man.

214
00:37:14.554 --> 00:37:20.734
It's all, it's all about animal, you know, animism and, um, okay, um, and animal spirits in the markets.

215
00:37:20.914 --> 00:37:29.814
Which brings us to, I'm gonna really narrow the question here, which is, uh, the current administration [laughs] of the United States has a couple more years.

216
00:37:29.934 --> 00:37:39.994
Um, Trump's approach and, like, attitude towards, um, the domestic frontier labs is all over the place, and their relationship with him is all over the place.

217
00:37:40.534 --> 00:37:50.694
And so as we think about these really big, big, you know, like stoner, bong-driven questions, like we also have to reconcile the fact that there's trillions of dollars being, you know, poured into this.

218
00:37:50.934 --> 00:38:01.834
This happened so quickly, and, uh, our nation's policy and executive decision-makers, uh, really affect the path that we take.

219
00:38:02.994 --> 00:38:16.304
What do you think a healthy or productive or cooperative, uh, stance from, uh, an administration in charge of the US looks like as we deal with these existential questions?

220
00:38:18.314 --> 00:38:29.674
Well, you know, I try not to get too deep into politics as a reporter. Uh, I guess I think most of... It's kind of like crypto, you know?

221
00:38:29.694 --> 00:38:38.614
It sort of is similar to crypto, is that like most of the AI firms would kind of like a regulatory scheme that governed them.

222
00:38:38.634 --> 00:38:48.994
They just don't want it to be like too horribly restrictive, and they wanna feel like they have say in that process. And I think that's something that everybody kind of agrees on.

223
00:38:49.184 --> 00:38:59.374
I think what makes it harder is, A, the Democrats and the Republicans in Washington just can't ever agree on anything. If anyone starts to make progress on one side, then it's the end of the universe for the other side.

224
00:38:59.894 --> 00:39:07.374
And then we have all of these lefties all over the country who just think that somehow AI's just gonna stop, which just isn't gonna happen.

225
00:39:08.114 --> 00:39:13.354
But they are also a, a giant impediment to getting some kind of reasonable regulatory scheme through.

226
00:39:13.394 --> 00:39:27.804
But yeah, if this stuff really is going to be even bigger than the Industrial Revolution, which I think there's a decent chance it is, I would make that bet, then probably we should go step by step in conversation with our governments, you know?

227
00:39:27.854 --> 00:39:36.294
Like, that should probably happen. Um, it, and, but I don't, I just don't [laughs] know, I don't see how we'd get there any kind of reasonable way in the current political climate in this country.

228
00:39:36.314 --> 00:39:44.394
But also, this country is the only country besides China that matters. China will be able to do whatever it wants, 'cause they're completely, is, is, is in control. They just are way behind us.

229
00:39:44.534 --> 00:39:53.574
Um, but uh, but this country is gonna lead us into whatever the future is most likely. There's a chance China still does, but I s- sort of think we, I believe in America.

230
00:39:54.054 --> 00:40:02.564
Um, and so, um, so it's like are we gonna be the irresponsible ones? That's the, that's the tough question. Brady, thank you so much for your time. A quick wrap-up question.

231
00:40:02.614 --> 00:40:16.534
Any teasers for what's coming down the pipe with your next investigations at Pirate Wires? Yeah. I'm, I am currently looking into, um, like ideas that have been very informative in the space for a long time.

232
00:40:16.634 --> 00:40:23.724
Uh, so at some point soon I'll do something related to effective altruism, because they're just, they were so far ahead of everybody on this topic.

233
00:40:23.794 --> 00:40:32.274
And um, I'm also just really interested, there's just a ton of think tanks working really hard on this topic, and so I wanna do something that like

234
00:40:33.474 --> 00:40:42.554
is a guide to that on some level, on like who they all are and where they lean and where they come from. Um, I'm still sort of sorting out the hard research on that, but th- those are two things I'm working on right now.

235
00:40:42.974 --> 00:40:54.394
I would love an EA alignment chart. Yeah. Like D&D, like neutral good, chaotic evil. That would, yeah. Yeah. So there's a freebie whenever you [laughs] think about- It's an idea. Yeah.

236
00:40:54.574 --> 00:41:05.254
Brady Dale, thank you so much for joining, man. Uh, really loving the coverage so far. Looking forward to seeing what else you produce at Pirate Wires. Thank you. Thank you, Brady.

237
00:41:07.414 --> 00:41:24.204
Not often do we get to take a big old bong rip and chat about, uh, EA and AI and other similar- Yeah, also not very often that we get the trigger word effective altruist- Yeah... in here. Uh, so a lot of pic pointers.

238
00:41:24.694 --> 00:41:32.594
Is that a trigger word? I, uh, we're, we're, we're pro-crypto, of course it's a trigger word. Yeah. I, I'm still reeling. A lot of, a lot of crypto folks still reeling from Sam Altman, which, uh, chaotic evil?

239
00:41:32.674 --> 00:41:40.674
I don't, maybe. Not to, not to- I think he only exists outside of the alignment. You, you said Sam Alt- you mean SBF. Uh, SBF. Oh my God. [laughs] Who also, by the way, invested in Anthropic.

240
00:41:41.214 --> 00:41:55.214
The, yeah, that's a whole rabbit hole. Okay, um, we are gonna keep on rolling. We are gonna talk about Soluna, small market cap, big higher, and then we're gonna talk about Arizona.

241
00:41:55.254 --> 00:42:10.854
Before we go to Soluna, a word from our sponsor, Luxor. [fireworks explode] This episode of Block Space Live is brought to you by Luxor's Commander, Bitcoin miner management software for enterprise operations.

242
00:42:11.494 --> 00:42:17.734
Luxor's Commander gives you real-time fleet monitoring, bulk remote commands across your fleet, and Intelligent Miner.

243
00:42:17.774 --> 00:42:24.944
That's an automated profitability engine that runs every five minutes and tests your fleet against live hash rate in energy markets.

244
00:42:25.514 --> 00:42:38.234
ERCOT back tests show 10% improved profitability with intelligent mining versus binary mining. Commander Pro is $100 per megawatt, or a 25 basis point pool fee adder, and you can try it for free for 60 days.

245
00:42:38.294 --> 00:42:55.690
If you'd like to learn more, go to luxor.tech/commander to get started. All righty, Charlie, like you said, small cap, big moves. Soluna. Names former Microsoft executive Ryan Carver, Chief Development Officer.

246
00:42:55.890 --> 00:43:03.130
This is coming fresh to y'all from Blockspace. Soluna appointed former Microsoft exec- executive Ryan Carver as Chief Development Officer on Thursday.

247
00:43:03.610 --> 00:43:10.470
Effective immediately, he is responsible for the entire life cycle of Soluna's AI and HPC data center platform and will report to the CEO.

248
00:43:11.070 --> 00:43:18.860
His responsibilities span site origination, power procurement, design, construction, commissioning, and operations. Can we clone him? That's a lot.

249
00:43:19.480 --> 00:43:27.680
[laughs] Soluna said the role brings together its development, construction, technology, operations, and power functions into one delivery organization.

250
00:43:27.710 --> 00:43:42.030
His m- he most recently served as Microsoft senior director of AI construction and site development and oversaw a construction operation valued in the tens of billions of dollars and led work at Microsoft's Fairwater AI Data Center campus in Mount Pleasant, Wisconsin.

251
00:43:42.790 --> 00:43:44.630
This is becoming a trend, Charlie.

252
00:43:44.730 --> 00:43:59.710
We've started to notice from some of the Bitcoin miners turned AI plays, they're increasingly turning towards the big leagues to poach talent as they try to make inroads within their data center business lines.

253
00:43:59.730 --> 00:44:16.380
As we reported last week, or two weeks ago, I believe, Keel h- uh, has hired its own, a, a new president to lead these efforts. Trying to get the story up here, but I will just read directly from our,

254
00:44:17.370 --> 00:44:28.970
from our s- uh, coverage here. Keel Infrastructure appoints former real estate investment trust executive Ganesh Iyer as president amid 2.2 gigawatt AI pipeline push. So,

255
00:44:30.690 --> 00:44:39.490
I mean, I'm wondering if we're gonna see more of these before the year is out. Clearly- I would, I would expect-- I mean, think about it. Let's say you are, um, one of these,

256
00:44:40.570 --> 00:45:04.690
you know, heavy hitters at one of these hyperscalers, but the thing is, you have the largess of a company, you, uh, maybe you're not as, like, t- up, high up in the hierarchy as you want, and there's a lot of these scrappy, uh, second and third tier, like, in terms of scale and execution history, uh, companies who want you to come call shots for them.

257
00:45:05.410 --> 00:45:12.790
And like, and they're gonna announce your hire. And maybe there's a, a very compelling, like, package, an incentive package with it too.

258
00:45:12.830 --> 00:45:26.670
So I could see a lot of this, 'cause everybody wants these types of folks and these connections. So, uh, 'cause, 'cause they have been, for the past twenty years, uh, building in not the limelight.

259
00:45:27.190 --> 00:45:37.950
Uh, I think if you are, if you've, uh, been circling around the C-suite, now you get to be in the absolute center of the C-suite. Uh, it could be pretty attractive to you.

260
00:45:38.050 --> 00:45:48.930
So I, I-- and it's a mutually beneficial deal because, uh, now they get your, they get your Rolodex, they get your phone book, they, y- they get your knowledge base and your execution history.

261
00:45:48.970 --> 00:45:51.210
So they- Yeah, you're, you're basically- They get to buy- Yeah.

262
00:45:51.550 --> 00:46:02.570
You're buying access and expertise, obviously, with these hires, and they definitely seem like a smart move for these companies as they bump up against the reality that building one of these things is really freaking hard.

263
00:46:02.590 --> 00:46:13.490
And right now, Soluna has two main projects that it's undergoing for AI and HPC hosting. The first one is Project Katie 2. This is coming from their June monthly update.

264
00:46:14.170 --> 00:46:23.330
Definitive joint venture agreement was signed with Metro blocks, replacing the previously announced non-binding MOU. Design team has completed the initial design package.

265
00:46:24.350 --> 00:46:34.350
There's tenant due diligence that's continuing with hyperscalers and neo clouds. Formal commercial negotiations have started with at least one potential tenant. So that is the one that is furthest along.

266
00:46:34.370 --> 00:46:45.790
The second project is Dorothy, Project Dorothy 3, 300-plus megawatt target for AI and HPC. A definitive agreement was signed with the landowner for 300 acres.

267
00:46:45.810 --> 00:46:57.250
Their due diligence survey work and environmental, uh, sur- surveys being done currently. Fiber studies have been launched. Ongoing coordination with utilities to expand load and convert current BTC load to AI.

268
00:46:57.390 --> 00:47:09.589
So that one is a little bit further back in terms of the development process. But it seems like so far Project Katie 2 is the apple of their eye for their initial AI rollout.

269
00:47:09.630 --> 00:47:21.840
It'd be curious to see if we get a tenant announcement within the quarter. But I think that wraps it up for Soluna, Charlie, unless you have anything else to comment on. No, I do not.

270
00:47:21.870 --> 00:47:37.030
Let me take a quick gander at Soluna's ticker today. Let's see. It is roughly flat on the day. So Soluna Holdings down 1.8%, not a whole lot.

271
00:47:37.730 --> 00:47:52.340
Um, again, small market cap comparatively, $171 million. Um, but, uh, you know, they're, they're, uh, almost actually [laughs] flat on the year. So it's down 6% on the year, which

272
00:47:53.510 --> 00:48:09.370
in the world of AI is not a volatile stock. We are gonna go to Arizona and talk about what's changing out in this great southwest. Uh, before that, a word from our sponsor, Ligos.

273
00:48:09.990 --> 00:48:25.690
[static noise] Bitcoin is making a run at 65K, but that doesn't mean that hedge funds and lending desks aren't still under pressure if they're mismanaging their book.

274
00:48:27.130 --> 00:48:38.230
If you wanna make sure that you're not gonna get screwed by a Bitcoin-backed lender, if you wanna make sure you don't become the next Celsius or BlockFi victim, do yourself a favor and check out Ligos Finance.

275
00:48:38.720 --> 00:48:48.730
They are our preferred Bitcoin-backed lender here at Blockspace. They use 100% Bitcoin native smart contracts to make sure that your Bitcoin is always in your control. You heard that right.

276
00:48:48.970 --> 00:49:00.766
With Ligos Finance, you always have custody of the Bitcoin you're using for their Bitcoin-backed lending, and nobody else touches it. Do yourself a favor, check them out.

277
00:49:00.806 --> 00:49:14.426
With Lagos there's no bridging, no rehypothecation, and no funny business, just competitive rates as low as 8.5 to 10% APR. Go to Lagos.finance to learn more. Charlie- All-... you get a sunburn.

278
00:49:14.466 --> 00:49:30.866
What's going on in Arizona? Yeah, so, um, if you didn't know, Arizona is second to Virginia in, uh, terms of planned development for data centers. More specifically, the metro area of Phoenix.

279
00:49:30.966 --> 00:49:42.886
You may know that TSMC, Taiwan Semiconductor, is building a big old foundry out there that extends to data centers outside of the data center corridor in Virginia. Arizona's kind of like number two.

280
00:49:43.546 --> 00:49:54.245
And a lot of- W- w- sorry, I gotta interrupt. H- why and how? I mean, this makes- Well-... just a lot of land that's cheap? Is that basically why? I mean... I think it's all of the above.

281
00:49:54.306 --> 00:50:06.006
There's, uh, cheap energy out there, um, land, it's a lot of desert as, as you would suspect. But actually a lot of this is because of the story.

282
00:50:06.016 --> 00:50:15.786
Arizona's had a pretty gracious data center incentive, uh, in the state for an- uh, for a while. I believe it's been around for three years.

283
00:50:16.166 --> 00:50:33.426
But anyway, that may be changing, or rather it is changing as the Arizona Commerce Authority, Arizona Corporation Commission, and utilities, uh, put pause on the data center incentives, and they intend on tightening oversight, quote, "Amid rapid AI-driven growth."

284
00:50:34.306 --> 00:50:49.746
So, uh, this article dropped July 15th, which is today, but it follows a, uh, the, the 2026 legislative session in Arizona.

285
00:50:49.866 --> 00:51:07.546
Um, as we've been covering on this live stream, uh, pushback a- against data centers has skyrocketed, and it's been cross domain, cross sociological, uh, profile, cross political [clears throat]

286
00:51:07.626 --> 00:51:23.746
and these are public oppositions which are both grassroots and at the national level. Also, a major, uh, tailwind of CCP fueled propaganda, as we covered last week with Sam Lyman. Highly recommend you go lead that.

287
00:51:23.786 --> 00:51:37.546
But then there's some organic pushback. And so, uh, it's a lot of the, what we expect. The pushback is water usage, it's electricity bills, it's zoning and environmental impact.

288
00:51:37.646 --> 00:51:49.026
So Governor Katie Hobbs and the Republican leaders compromised on a three-year freeze of the sales tax exemption for data center operators through 2029.

289
00:51:49.546 --> 00:52:04.955
So this kind of the sweet deals that existed to bring data centers to Arizona are ending for the foreseeable future. Um, Hobbs signed this related, uh, utility reporting, uh...

290
00:52:04.955 --> 00:52:14.246
Hobbs also signed related utility reporting requirements on grid connections. So the Arizona Corporation Commission has increased regulatory scrutiny. Here's a bit of a timeline.

291
00:52:15.326 --> 00:52:30.786
Uh, April 2025, chairman of the ACC, Kevin Thompson, uh, opens a docket on large data center load rates and protections for other rate payers. April 2026, there was a workshop, broad consensus that the cost causer pays.

292
00:52:31.086 --> 00:52:44.926
So that is if you are adding to the load and adding to the net cost of the utility, you pay that. We see this happening a lot. This is kind of happening in Oklahoma, happening across, um, across the states.

293
00:52:45.106 --> 00:52:50.406
Our attitude has been largely, uh, favorable of this as at The IA level. Yeah, and- Um-...

294
00:52:50.416 --> 00:53:03.406
even outside of the, I would say, the ethical arguments for this, like obviously you want the bulk consumers to be paying for the upgrades and for the transmission if they're demanding it so that that doesn't get passed on to rate payers.

295
00:53:03.416 --> 00:53:11.235
That's kind of the ethical consideration here, but I also think there's just like clearly a free market angle. The, the data centers don't really have leverage here. They're...

296
00:53:11.686 --> 00:53:17.566
There's not enough power to go around for the demand that we're seeing, and they're taking it anywhere they can get it.

297
00:53:17.606 --> 00:53:24.026
So like I said at the beginning of the show, it's a seller's market for energy and land for these data centers.

298
00:53:24.046 --> 00:53:30.046
And if that's the case, it stands to reason that you would see these incentives evaporate a little bit, you know?

299
00:53:30.066 --> 00:53:43.866
Yeah, it's one of the- That was pushbacks against Kevin O'Leary's Box Elder County Utah Data Center, is that it had this incredible tax ad- a- advent, um, this tax advantage structure for when it was moving in.

300
00:53:43.926 --> 00:53:52.986
It was given these incentives alongside other developers in the area. And Kevin O'Leary basically threw his hands up and says like, "Well, don't get mad at me for that." But I would...

301
00:53:53.486 --> 00:54:03.366
Not surprising to see some of those incentives melt away as the localities are now realizing, "Actually, we have a lot of leverage here. They're probably gonna come anyway." Yeah.

302
00:54:03.506 --> 00:54:09.706
I think we're gonna see, and I- we're, we're seeing this play out in real time right now, the legislators are wising up.

303
00:54:09.746 --> 00:54:21.786
The q- the commissions, the corporation commissions, the utilities, they're wising up really quick kind of because they're, they're, they're being held in the hot seat by the public and mostly ground- grassroots efforts.

304
00:54:22.586 --> 00:54:27.486
I think the big question is which pa- which different paths do they take?

305
00:54:27.566 --> 00:54:50.166
Because this seems kind of like a pretty big inflection point for, uh, decision makers and legislators to get smart about how to leverage this industry to their benefit, uh, and tread that thin line of healthy adoption, but also maintaining the interest and health of the state economy to the individual, to the average person living there.

306
00:54:50.346 --> 00:55:03.956
I think we'll see a wide divergence. Um, it'll probably happen a little bit along party lines, uh, and [laughs] you're probably too- But- But it might not... Arizona fairly red state, right?

307
00:55:03.986 --> 00:55:15.866
Fairly red state with, I think, a Democratic governor, right? Am I not mistaken? She's-[lips smack] Um, let's see. [clears throat] Democratic Party. Yeah. Yeah.

308
00:55:15.896 --> 00:55:28.726
I remember 'cause this is a whole story like last year for unrelated reasons 'cause they- we had the, the Bitcoin, uh, the Bitcoin, uh, stockpile, uh, story when we- Right... covered this, this time last year.

309
00:55:29.486 --> 00:55:42.976
Um, and, and, uh, Arizona Governor Katie, Katie Hobbs' like friction with the largely Republican legislature. So- 100%. Um, [clears throat] yeah, we could, you know, this will be real interesting.

310
00:55:43.016 --> 00:55:49.116
And comp- contrast this with Governor Katie, I'm sorry, Governor Hochul in New York,

311
00:55:50.096 --> 00:56:12.396
uh, who i- in her interviews, I think she had an interview with, uh, Bloomberg yesterday, really interesting, came across to me as actually being pretty savvy to, uh, the deci- the, the decision-making process and to the strategy of what's going forward, and that was a bit reinforced by our interview with TeraWulf co-founder Uzair Khan yesterday.

312
00:56:12.496 --> 00:56:23.416
As he said, we kind of welcome this, uh, to reevaluate the- Yeah... status. Uh, they're gonna say that, right? They're g- They have to say- [laughs]... 'cause they got, they gotta play ball. Like, what are they gonna do?

313
00:56:23.456 --> 00:56:27.376
They gotta play ball, but I do think what you're keying in on is important,

314
00:56:28.936 --> 00:56:43.356
mostly with regards to the different stances we're seeing, and there's a clear spectrum evolving, and I think Arizona f- kind of falls somewhere in the middle, and it's not one that I'd a- absolutely hate, again, given the facts on the ground of where the market's at.

315
00:56:43.766 --> 00:56:53.276
The data centers don't have much of a choice. If they find prime land, then they're gonna take it, and Arizona is pausing on the incentives.

316
00:56:53.356 --> 00:57:01.666
To me, it kind of makes sense in saying, "Well, we wanna make sure this value accrues to our localities and to our local jurisdictions." I think New York, maybe not on the extreme end.

317
00:57:01.716 --> 00:57:05.196
The extreme end would just be legitimately going nuclear and banning them outright.

318
00:57:05.716 --> 00:57:19.136
But on the extreme end of state responses so far, in as much as it is a pause for a year, she also, Hochul has, has called for like seven or eight new agencies to do the feasibility studies.

319
00:57:19.556 --> 00:57:24.876
I mean, when I look at this, this is- A lot of agencies. What's up? Yeah, it is. That's a lot of agencies. Yeah.

320
00:57:25.096 --> 00:57:35.916
It, it's, it, it's just totally, the whole process will be choked with red tape, and it will suffocate legitimate development and innovation in the state if it is too stringent.

321
00:57:35.936 --> 00:57:48.576
And then on the other end of the spectrum, you do have Texas, as always, and, you know, Greg Abbott basically said, "Figure this out. We're gonna change the rules. We're not gonna stop them from coming here."

322
00:57:48.616 --> 00:57:58.036
But he told ERCOT and the Public Utilities Commission, "You need to come up with a clear plan for how this isn't going to affect ratepayers." And a, a, a key

323
00:57:59.256 --> 00:58:09.596
pillar of that plan is to make sure that the data centers pay for the transmission. So it's been really fascinating to see, and almost, I, I'm not surprised that it's coming at this point.

324
00:58:09.776 --> 00:58:17.816
I am a little surprised we hadn't seen, we haven't seen more of this earlier in the year. But now it's clearly coming to a head.

325
00:58:17.956 --> 00:58:24.526
I imagine you'll see a lot of other states take their own actions similar to some of, to Arizona, New York, or Texas.

326
00:58:24.996 --> 00:58:32.876
They'll, you know, depending on the political allegiances and how pro-business that state is, they will, uh, kind of fall on the spectrum accordingly.

327
00:58:32.936 --> 00:58:43.076
One last thing to note, Charlie, just 'cause I had to answer this question for myself, why the hell is a desert the place that data centers wanna go, one of the premier states for data centers in the US?

328
00:58:43.776 --> 00:58:55.796
Well- Part of it is the climate. It's dry. That causes some problems, but the low humidity can actually help with cooling. The other is just that it's actually a nuclear, nat gas powerhouse.

329
00:58:55.876 --> 00:59:10.656
There's 45% of the y- Arizona grid is [lips smack] furnished by natural gas power as of 2025. Nuclear is 26.7%, which puts it at one of the highest ratios, one-fourth there, of any state in the US.

330
00:59:11.156 --> 00:59:23.356
Then solar coming in at 13.3, and then coal and hydroelectric at 79, 7.9 and 4.2. [lips smack] So that's all I have to say about that. I, I- That's all I have to say, too.

331
00:59:24.436 --> 00:59:35.066
Uh, yeah, I mean, maybe we're not doing AI in the Middle East. Maybe, uh, it'll just go to Arizona, um, somehow even hotter than the Middle East, if you can believe that sometimes.

332
00:59:36.056 --> 00:59:40.756
On that note, thank you very much for listening to Block Space Live, presented by CleanSpark.

333
00:59:41.476 --> 00:59:52.556
We go live every single weekday at 1:00 PM Eastern, featuring quick hits on AI data centers, markets, emerging technology, and if you're lucky, we'll talk about some Bitcoin.

334
00:59:52.636 --> 01:00:05.836
If you like what you hear, you'll love our newsletter. You can find it at newsletter.blockspacemedia.com, and check out all of our content on our website at blockspace.media. I'm Charlie. I'm Colin.

335
01:00:06.416 --> 01:00:16.556
And we'll see you tomorrow. [dramatic percussion] [explosion]