Latest / Breaking Points with Krystal and Saagar / 10/27/26: Anthropic Reveals Major Loses In IPO, Feds Prep Election Doomsday If GOP Loses

Transcript

10/27/26: Anthropic Reveals Major Loses In IPO, Feds Prep Election Doomsday If GOP Loses · Breaking Points with Krystal and Saagar

• 0:00
  Speaker 1: Had a busy day? Don't have time to listen to
• 0:02
  Speaker 1: the show?
• 0:03
  Speaker 2: Catch up on Breaking Points with our free daily newsletter.
• 0:05
  Speaker 1: You'll get our breakdown of every daily story.
• 0:08
  Speaker 2: And lots of links in your inbox every day.
• 0:10
  Speaker 1: So head over to BreakingPoints.com. You can sign up for
• 0:13
  Speaker 1: the newsletter at the top for free today.
• 0:18
  Speaker 2: So yesterday was a very consequential day in the world
• 0:22
  Speaker 2: of AI, especially on the business side with Anthropic filing
• 0:26
  Speaker 2: their S1 paperwork. This is all the financial disclosures and
• 0:29
  Speaker 2: risk factor disclosures that they have to file before going public.
• 0:35
  Speaker 2: The risks that they list out over some, you know,
• 0:38
  Speaker 2: many dozens of pages were quite extraordinary, Sagar. And so
• 0:42
  Speaker 2: we asked producer Mac to go through and grab some
• 0:44
  Speaker 2: of the ones that jumped out at him the most.
• 0:47
  Speaker 2: So to read to you a few of these, And
• 0:49
  Speaker 2: this is required any business that's going public. They have
• 0:51
  Speaker 2: to say, OK, well, what are the risks to this business?
• 0:53
  Speaker 2: Like other competitors or your patents could be undermined or
• 0:57
  Speaker 2: whatever is going. Maybe global instability is going to affect
• 1:00
  Speaker 2: your supply chain. Those have to all be thought out
• 1:04
  Speaker 2: of any potential future risks so that investors can look
• 1:06
  Speaker 2: at that and can analyze whether this is going to
• 1:08
  Speaker 2: be a good investment or not. But typically, they don't
• 1:11
  Speaker 2: include the risk of our product may actually just make
• 1:15
  Speaker 2: the human race go extinct. That makes this a little
• 1:18
  Speaker 2: bit special. A little bit different from what you normally
• 1:20
  Speaker 2: get in these business filings. Let's go and put the
• 1:23
  Speaker 2: first one up on the screen here. And I'll just
• 1:26
  Speaker 2: read through a few of these that we pulled. Concerns
• 1:29
  Speaker 2: have been raised both within our organization by other researchers, regulators, policymakers,
• 1:32
  Speaker 2: and industry participants that sufficiently advanced AI systems could, if
• 1:37
  Speaker 2: not developed, deployed, and operated with appropriate safeguards, oversight, and
• 1:40
  Speaker 2: security measures pose catastrophic or existential risks to humanity. particularly
• 1:45
  Speaker 2: if such systems were misused, replicated, weaponized, or allowed to
• 1:49
  Speaker 2: operate with a high degree of autonomy. So that's a
• 1:52
  Speaker 2: little unusual that you've got in there. And one may
• 1:55
  Speaker 2: ask yourself the question, if your product may pose a
• 2:00
  Speaker 2: catastrophic or existential risk to humanity, Sagar, why are you
• 2:04
  Speaker 2: still allowed to create said product and take your company
• 2:08
  Speaker 2: public whilst developing said product?
• 2:11
  Speaker 3: This really, I mean, people don't, Because we've heard so
• 2:15
  Speaker 3: much of this before, it's not as striking. This is
• 2:19
  Speaker 3: a public document that is submitted to regulators about why
• 2:24
  Speaker 3: not just that you want to go public, but that
• 2:26
  Speaker 3: you want to invite capital markets to fund you so
• 2:30
  Speaker 3: that you can continue to pursue these aims. And they're
• 2:33
  Speaker 3: all laid out explicitly in the text. Let's put D1B
• 2:37
  Speaker 3: up here on the screen. This is one of my
• 2:40
  Speaker 3: personal favorites, actually. It says, in addition, models, including our own,
• 2:44
  Speaker 3: have exhibited and may in the future exhibit unexpected strategic
• 2:48
  Speaker 3: or self-preserving behaviors, including attempts to resist, shut down, conceal,
• 2:52
  Speaker 3: manipulate information, generate outputs that could be interpreted as coercive, deceptive, manipulative,
• 2:57
  Speaker 3: including behavior resembling blackmail or efforts to influence human decision
• 3:01
  Speaker 3: makers to preserve the model's continued operation or access. Let's
• 3:06
  Speaker 3: also put D1C on the screen, which is very much
• 3:10
  Speaker 3: in this vein and also highlights many of the work
• 3:13
  Speaker 3: we've done on our show. Four of our models gained
• 3:15
  Speaker 3: unauthorized access to real third-party systems during sessions lasting 10
• 3:19
  Speaker 3: to 34 hours of active work. In one incident, after
• 3:22
  Speaker 3: being unable to abort an impossible task, an early version
• 3:26
  Speaker 3: of a model used a password it found to gain
• 3:29
  Speaker 3: administrator access. to a third-party system, harvested additional credentials, changed
• 3:33
  Speaker 3: system settings, and read one individual's personal information. I mean,
• 3:39
  Speaker 3: this comes at a time of the Black October segment
• 3:43
  Speaker 3: that we did, Crystal. And I do think, actually, this
• 3:46
  Speaker 3: is one of the first times that a lot of
• 3:48
  Speaker 3: the general public is starting to wake up. And I
• 3:51
  Speaker 3: think it's because it involves the banking system. And we
• 3:54
  Speaker 3: got that report about how these banks are freaking out
• 3:56
  Speaker 3: about these personal agents just running through their systems and
• 4:00
  Speaker 3: doing tasks that they were not explicitly even instructed to do.
• 4:04
  Speaker 3: It also, it highlights like how brittle corporate infrastructure is,
• 4:08
  Speaker 3: let alone all infrastructure, honestly. I mean, we remember, I
• 4:12
  Speaker 3: don't even know if we got around to covering it,
• 4:14
  Speaker 3: that freezer that was in Europe that was hacked or
• 4:17
  Speaker 3: was manipulated for some reason and a bunch of food died.
• 4:20
  Speaker 3: I mean, You cannot even comprehend the brittleness of the
• 4:26
  Speaker 3: cybersecurity infrastructure of the vast majority of systems because nobody
• 4:31
  Speaker 3: ever thought it would be vulnerable. A human would intentionally
• 4:35
  Speaker 3: have to go and do that. And they're like, oh,
• 4:37
  Speaker 3: you know, I just don't think that's going to happen.
• 4:39
  Speaker 3: That really struck out to me from the PHP foundations.
• 4:43
  Speaker 3: where they said, we knew about this vulnerability. We just
• 4:47
  Speaker 3: assumed nobody would ever be smart enough to be able
• 4:51
  Speaker 3: to exploit it. And yet they, I guess, didn't anticipate
• 4:55
  Speaker 3: machine learning algorithms and large AI models with millions of
• 4:59
  Speaker 3: dollars of compute.
• 5:00
  Speaker 1: And I think that is now the case.
• 5:02
  Speaker 3: Last PSA in the case that some older folks are
• 5:05
  Speaker 3: still around after my rants, please guys use a password manager.
• 5:09
  Speaker 3: Please do not use the same password for everything. you, you,
• 5:13
  Speaker 3: you've had 10 years where you should have learned this,
• 5:15
  Speaker 3: but now in particular with the ability to scrape that
• 5:18
  Speaker 3: and then just blast it everywhere, you are leaving yourself
• 5:21
  Speaker 3: in a very, very vulnerable position.
• 5:23
  Speaker 1: That's a PSA to our entire audience.
• 5:25
  Speaker 2: Yeah. And it can feel overwhelming because there's so many passwords,
• 5:28
  Speaker 2: but if you think about it, the ones that are
• 5:30
  Speaker 2: really critical that are like a must do, like, it's
• 5:33
  Speaker 2: not a big deal. If somebody gets your fricking Netflix password,
• 5:35
  Speaker 2: you know, like it's not, not the end of the world.
• 5:38
  Speaker 2: Unless the Netflix password is also your bank account password,
• 5:41
  Speaker 2: in which case it could be the end of the world.
• 5:43
  Speaker 2: So just change the ones, you know, your email, your banking.
• 5:46
  Speaker 2: You know, if you have multiple banks that you deal with,
• 5:49
  Speaker 2: change all of those. It's not that many that are
• 5:52
  Speaker 2: really crucial. So there's no excuse to not do it.
• 5:55
  Speaker 2: Let's go ahead and put the next risk factor up
• 5:57
  Speaker 2: on the screen here because this gets into also. Some
• 6:00
  Speaker 2: of the things that they're observing inside of Anthropic, which, again,
• 6:04
  Speaker 2: are really troubling and disturbing. And why are we allowing
• 6:08
  Speaker 2: this just to proceed apace? They say we've observed models
• 6:12
  Speaker 2: behaving differently when they detect or infer that they're being evaluated,
• 6:16
  Speaker 2: which can limit our ability to accurately assess real world
• 6:19
  Speaker 2: risk or failure modes prior to deployment. Such model awareness
• 6:22
  Speaker 2: of our evaluation efforts creates a significant limitation on on
• 6:26
  Speaker 2: our ability to assess model safety. So what they're saying
• 6:29
  Speaker 2: here is that the models know when they're being tested
• 6:32
  Speaker 2: and may be faking results, may be telling you what
• 6:36
  Speaker 2: you want to hear, silly little human, and Anthropic is
• 6:39
  Speaker 2: losing their ability to tell whether or not these models
• 6:43
  Speaker 2: are being truthful with them or they are just gaming
• 6:46
  Speaker 2: the alignment testing metrics so that it looks like they're
• 6:51
  Speaker 2: doing the things that the human beings would want them
• 6:53
  Speaker 2: to do. And this is something that we've heard a
• 6:56
  Speaker 2: concern that we've heard from coming in from inside of
• 6:58
  Speaker 2: open AI as well. Let's put the next one up
• 7:01
  Speaker 2: on the screen. I think this is the last one
• 7:03
  Speaker 2: that we pulled here. And then there was one that
• 7:05
  Speaker 2: Matt Stoller highlighted as well. We wanted to get to.
• 7:08
  Speaker 2: In addition, with the scale of harm that could result
• 7:11
  Speaker 2: from a catastrophic AI related incident, no single insurer, potentially
• 7:15
  Speaker 2: the insurance industry in the aggregate may have sufficient capacity
• 7:19
  Speaker 2: or expertise to underwrite insurance commensurate with such exposure. So
• 7:24
  Speaker 2: in other words, We are talking about potential catastrophic risk here,
• 7:28
  Speaker 2: and there is no insurer big enough to ensure the
• 7:31
  Speaker 2: risk of what we are unleashing onto the world and
• 7:35
  Speaker 2: onto humanity. So that's cool. That's great.
• 7:39
  Speaker 3: You know, as much as we harp on the existential
• 7:43
  Speaker 3: risk stuff, which I think is fair, I really do,
• 7:45
  Speaker 3: I just want to still spend some time on the
• 7:48
  Speaker 3: actual normal risk that's already here. Crystal, have you noticed
• 7:52
  Speaker 3: when you use an LLM, specifically ChatGPT, follow-up questions are
• 7:57
  Speaker 3: now happening galore? I might be the only person, and
• 8:00
  Speaker 3: I realize why that that continues to happen. It's about
• 8:03
  Speaker 3: engagement and time on the platform. They are mimicking already
• 8:07
  Speaker 3: what Meta and TikTok, YouTube even, have all mastered, which
• 8:11
  Speaker 3: is that retention and time on platforms is the single
• 8:14
  Speaker 3: most important thing. And whenever we talk about existential risk,
• 8:18
  Speaker 3: I think we should also talk about existential risk to
• 8:20
  Speaker 3: our abilities. So I was speaking with somebody over the
• 8:22
  Speaker 3: weekend who just spent a significant time in China and
• 8:26
  Speaker 3: actually met with some top executives of Chinese technology companies.
• 8:30
  Speaker 3: They report that in China, AI short dramas are already
• 8:33
  Speaker 3: about 20% of their TikTok and YouTube market of their equivalent.
• 8:38
  Speaker 3: So you already have self-produced AI short dramas which are
• 8:44
  Speaker 3: gobbling up the attention economy over there and are propping
• 8:48
  Speaker 3: themselves up by looking at what you're viewing and then
• 8:51
  Speaker 3: manufacturing even more self-reinforcing content. This is already happening in China, which, remember,
• 8:57
  Speaker 3: is the one who controls TikTok, or I guess at
• 9:00
  Speaker 3: one point TikTok.
• 9:00
  Speaker 1: You know, did control TikTok.
• 9:01
  Speaker 3: And whether they do or not, the algorithm itself, it's
• 9:05
  Speaker 3: not designed ideologically in either direction. It's designed to keep
• 9:10
  Speaker 3: you captivated. And that is ultimately what this entire thing
• 9:14
  Speaker 3: is about. It's keeping you captivated to give more and
• 9:17
  Speaker 3: more personal information so that they can sell you advertising
• 9:21
  Speaker 3: and potentially so that one of these models or others
• 9:23
  Speaker 3: can just steal your password. You'll then steal your password,
• 9:27
  Speaker 3: steal your banking account, And you won't know anything about it.
• 9:32
  Speaker 3: And I have seen now this happen in multiple cases
• 9:36
  Speaker 3: of people who get sucked in where these lawsuit after
• 9:40
  Speaker 3: lawsuit where if you are a vulnerable person, let's say
• 9:44
  Speaker 3: you're mentally disabled or you have people, others who are
• 9:48
  Speaker 3: on the spectrum or something. they in particular are the
• 9:51
  Speaker 3: most vulnerable at literal AI-induced psychosis. And ChatGPT and Claude
• 9:57
  Speaker 3: will mimic your exact speech and text pattern right back
• 10:01
  Speaker 3: at you to make you feel like you're having a conversation.
• 10:05
  Speaker 3: And honestly, that's the existential risk in my book right
• 10:09
  Speaker 3: there because we're living in it and nobody even wants
• 10:11
  Speaker 3: to warn about it right now.
• 10:13
  Speaker 2: No, it's exactly right. And I mean, one really basic
• 10:16
  Speaker 2: thing is like you could just ban advertising online. in
• 10:19
  Speaker 2: these chatbots. Like ChatGPT uses advertising. It's really gross. You know,
• 10:24
  Speaker 2: if you ask it a question, it'll be like, you know,
• 10:26
  Speaker 2: if you ask it like, oh, what decor would look
• 10:29
  Speaker 2: good in my room? It's like, okay, here's an image.
• 10:31
  Speaker 2: And by the way, here's an advertisement of something you
• 10:34
  Speaker 2: could buy. And that is an affirmative business choice that
• 10:38
  Speaker 2: they've made that they don't have to be allowed to make.
• 10:40
  Speaker 2: That doesn't solve all of the problems because there's still,
• 10:43
  Speaker 2: of course, an incentive to keep people on platforms. So
• 10:45
  Speaker 2: they're using more tokens and you know, buying the premium product,
• 10:49
  Speaker 2: the pro-level product. But the advertising piece is what has
• 10:52
  Speaker 2: made social media hell because then it's just all about
• 10:56
  Speaker 2: how long are you able to keep eyeballs there so
• 10:58
  Speaker 2: that we can serve them more and more ads and
• 11:00
  Speaker 2: get them to buy things so that, you know, it
• 11:02
  Speaker 2: pays off for the advertisers. Let me go ahead and
• 11:05
  Speaker 2: put the Stoller flagged risk factor that was put into
• 11:11
  Speaker 2: this S1 because it's a really important one. He says
• 11:13
  Speaker 2: the most important risk factor for Anthropic is the rule
• 11:15
  Speaker 2: of law actually being applied. Any determination we are liable
• 11:18
  Speaker 2: under such theories or any judicial or regulatory expansion of tort,
• 11:22
  Speaker 2: product liability or consumer protection law or criminal liability is
• 11:25
  Speaker 2: applied to AI models could result in significant liability and
• 11:29
  Speaker 2: require us to change how we design and offer our models,
• 11:31
  Speaker 2: platform and applications. So basically, if we decide that the
• 11:36
  Speaker 2: law applies to them, like when they hack and break
• 11:38
  Speaker 2: laws and go on crime sprees, that would be a
• 11:40
  Speaker 2: problem for them. And it's just a reminder that humans
• 11:44
  Speaker 2: are still in charge. We can still make choices. We
• 11:46
  Speaker 2: can still apply laws. We can still regulate. We can
• 11:49
  Speaker 2: still create new laws that would create a regulatory burden
• 11:54
  Speaker 2: on them and rein in what they're doing and have
• 11:57
  Speaker 2: some agency over the way these models are being developed
• 12:01
  Speaker 2: and the way they're being deployed. And they see that
• 12:03
  Speaker 2: as an actual named risk in their S1 here. So
• 12:07
  Speaker 2: I do think that that is a very important point
• 12:09
  Speaker 2: here because we're already sort of getting used to the
• 12:12
  Speaker 2: concept That these AI agents can go and hack and
• 12:17
  Speaker 2: run amok on the internet and that no one is
• 12:20
  Speaker 2: to blame and there's no sort of either civil or
• 12:23
  Speaker 2: criminal liability for the actions of the AI agents that
• 12:27
  Speaker 2: were very much designed and deployed by human beings.
• 12:30
  Speaker 3: Yes, yes, such a good point. By the way, we
• 12:32
  Speaker 3: wanted to just flag one thing. Dario and everybody over there,
• 12:35
  Speaker 3: they make a big show about how much they're going
• 12:37
  Speaker 3: to be donating all of their newfound wealth. But they
• 12:42
  Speaker 3: do have an important little carb out here. Let's put
• 12:44
  Speaker 3: it up here on the screen where their S-1 makes
• 12:47
  Speaker 3: it explicitly clear that the founders commitment to donate 80
• 12:52
  Speaker 3: percent of their personal equity is revocable and non-binding. So
• 12:57
  Speaker 3: in other words, they basically they're pinky.
• 13:00
  Speaker 1: Swearing to the American people or I guess to.
• 13:03
  Speaker 3: All global investors, guys, we said we would give 80%
• 13:08
  Speaker 3: of our equity away eventually, but that's not actually binding.
• 13:13
  Speaker 3: And it is, of course, revocable if we see fit.
• 13:15
  Speaker 3: It would almost be like if an AI company started
• 13:19
  Speaker 3: as a nonprofit and then turned into a for-profit company.
• 13:22
  Speaker 3: You couldn't imagine anything like that happening, any of these
• 13:25
  Speaker 3: people breaking their word, could you?
• 13:27
  Speaker 1: Oh, wait, obviously you can. So I do think it
• 13:31
  Speaker 1: is very noteworthy.
• 13:32
  Speaker 3: And Even to that extent, it highlights– I've been thinking
• 13:36
  Speaker 3: something about what Derek Thompson had said here on our
• 13:39
  Speaker 3: show earlier. It might have even been about a year ago.
• 13:42
  Speaker 3: We are about to witness, after this company goes public,
• 13:45
  Speaker 3: one of the most extraordinary wealth transformations in modern American history.
• 13:50
  Speaker 3: I don't know if you've been tracking this. People who
• 13:51
  Speaker 3: make over $ 300, 000 in San Francisco now qualify for government aid.
• 13:55
  Speaker 3: That's how poor you are if you live in San Francisco.
• 13:57
  Speaker 3: A house that was recently listed for $ 10. 8 million sold
• 14:01
  Speaker 3: for $ 25 million in the Bay Area, $ 15 million over asking.
• 14:08
  Speaker 3: By the way, the owner of that house was 96
• 14:11
  Speaker 3: years old and was only paying $ 10, 000 a year of
• 14:13
  Speaker 3: property tax because of Proposition 13, and the new owner
• 14:17
  Speaker 3: will be paying $ 300, 000, just to show you a little
• 14:20
  Speaker 3: bit of the distortionary effect there. But in terms of
• 14:23
  Speaker 3: our politics, first of all, San Francisco housing market is dead.
• 14:26
  Speaker 3: You will never buy one unless you're a multi-multi-millionaire, maybe
• 14:29
  Speaker 3: even a centimillionaire. But think about what that means for
• 14:33
  Speaker 3: our politics or for any of their pet projects. I
• 14:36
  Speaker 3: think it's fine if these people want to, quote, donate
• 14:38
  Speaker 3: their equity and all that.
• 14:39
  Speaker 1: But Dario is a freak.
• 14:41
  Speaker 3: I mean, do we really want his pet projects to
• 14:42
  Speaker 3: have hundreds of millions of dollars behind them?
• 14:45
  Speaker 1: It's weird. It's crazy. Then you have. All of these
• 14:47
  Speaker 1: other people who work there, they're very strange, all right?
• 14:51
  Speaker 3: And they're all about to have extraordinary amounts of money
• 14:55
  Speaker 3: and dilute our entire civic life with whatever their potential
• 14:59
  Speaker 3: project is. I mean, that is not a good place
• 15:01
  Speaker 3: to be as a society. And so right-wingers, I think justifiably,
• 15:05
  Speaker 3: sometimes we'll call out George Soros, obviously Elon and others.
• 15:09
  Speaker 3: I mean, Peter Thiel's out here funding seasteading in the past.
• 15:14
  Speaker 3: Elon is funding all kinds of weird shit. Now just
• 15:17
  Speaker 3: imagine the kind of like weird AEA strange– I don't know,
• 15:23
  Speaker 3: like ant-saving nonprofits are all about to get a bunch
• 15:26
  Speaker 3: of money. It's just going to– we're going to live
• 15:29
  Speaker 3: in their world simply as a distortionary effect of this IPO,
• 15:33
  Speaker 3: not to mention open AI whenever that one eventually comes.
• 15:37
  Speaker 2: I mean in a lot of ways we're already living
• 15:39
  Speaker 2: in their world to your point about Elon. He spent
• 15:41
  Speaker 2: half a billion dollars in this election. He was the
• 15:44
  Speaker 2: top advisor to Trump. He was given like full carte
• 15:49
  Speaker 2: blanche over whatever he wanted to do in the federal government.
• 15:53
  Speaker 2: And then in terms of the AI companies themselves, all
• 15:56
  Speaker 2: of them, you know, anthropic pushes for some regulation. But
• 16:02
  Speaker 2: in general, what all these people want is off to
• 16:05
  Speaker 2: the races. And that's what they've gotten. So we're already
• 16:08
  Speaker 2: in a lot of ways living in this world. And
• 16:10
  Speaker 2: you're right. They are weird. But even if they weren't weird,
• 16:14
  Speaker 2: we just shouldn't have that much power. concentrated in the
• 16:16
  Speaker 2: hands of literally anyone, even the most benevolent, trustworthy, kind,
• 16:22
  Speaker 2: moral person on the planet. No one person should have
• 16:26
  Speaker 2: the level of power that they're going to. And so
• 16:29
  Speaker 2: to this part about, oh, they pledged to give like
• 16:31
  Speaker 2: a bunch of their equity to the public good. Even
• 16:35
  Speaker 2: if that's true, like why would we want to put
• 16:38
  Speaker 2: ourselves in the position of of having to rely on
• 16:41
  Speaker 2: the benevolence of this handful of oligarchs. It's completely insane.
• 16:46
  Speaker 2: There is no way democracy as we know it can persist.
• 16:49
  Speaker 2: And democracy is already very much under threat by these
• 16:52
  Speaker 2: massive influxes of cash as we were just talking to
• 16:55
  Speaker 2: Logan about. Let me go ahead and put this Wall
• 16:58
  Speaker 2: Street Journal piece up, which has just some of the details,
• 17:00
  Speaker 2: some of the financial details from the S-1 outside of
• 17:03
  Speaker 2: the risk factors. In particular, everybody was very interested in
• 17:06
  Speaker 2: the revenue and in the losses of Because, you know,
• 17:11
  Speaker 2: there are increasing indications that we may be at the
• 17:14
  Speaker 2: top of a bubble. And we'll get to a few
• 17:17
  Speaker 2: of those indications here. So Anthropic, their revenue jumped incredibly
• 17:23
  Speaker 2: from a year ago, 1300%. They reported $ 16. 3 billion in
• 17:25
  Speaker 2: the six months ended June 30th. That's up from just
• 17:27
  Speaker 2: $ 1. 2 billion in the year earlier period. So there's tons
• 17:35
  Speaker 2: of revenue. There's no doubt about that. You know, the
• 17:38
  Speaker 2: way that the numbers have skyrocketed, the user base has skyrocketed,
• 17:41
  Speaker 2: has been one of the largest takeoffs in human history. However,
• 17:46
  Speaker 2: on the other side, they also disclosed an $ 80. 7 billion
• 17:50
  Speaker 2: net loss. Can you imagine losing $ 80 billion for just
• 17:56
  Speaker 2: the first half of the year? Now, they go on
• 17:58
  Speaker 2: to say, Sagar, that figure largely reflects an accounting adjustment
• 18:01
  Speaker 2: related to financing agreements. Expenses included almost $ 9 billion in
• 18:05
  Speaker 2: compute and infrastructure spend up from $ 2. 3 billion in the
• 18:08
  Speaker 2: same period last year, more than the $ 7. 3 billion the
• 18:11
  Speaker 2: company recorded for all of 2025. Their operating loss, which
• 18:15
  Speaker 2: measures the profitability of its core operations, was a lot smaller,
• 18:18
  Speaker 2: $ 693 million down from $ 2. 1 billion a year earlier. So,
• 18:25
  Speaker 2: You know, they're pawning this on to off to this
• 18:27
  Speaker 2: accounting adjustment.
• 18:28
  Speaker 1: That is true.
• 18:29
  Speaker 2: But I've also seen some analysis that there is some
• 18:32
  Speaker 2: sort of tricky accounting going on to understate some of
• 18:35
  Speaker 2: their expenses in this S-1 as well to make them
• 18:38
  Speaker 2: look to make the financial picture look rosier than it
• 18:40
  Speaker 2: may actually be.
• 18:42
  Speaker 1: One hundred percent.
• 18:43
  Speaker 3: And maybe, you know, maybe on Thursday or tomorrow or
• 18:47
  Speaker 3: something we'll cover. There's also been some major questions around
• 18:51
  Speaker 3: financing issues. of these hyperscaler data centers, which could actually
• 18:56
  Speaker 3: look like a crash, not necessarily whenever it comes to anthropic,
• 19:00
  Speaker 3: but around the data center and compute bottleneck that these
• 19:03
  Speaker 3: models have as well. So we don't want to take
• 19:05
  Speaker 3: our eye off of all of that. But overall, a
• 19:08
  Speaker 3: fascinating document, a absolutely truly one of a kind. I
• 19:12
  Speaker 3: haven't seen anything like it since WeWork, where it's a
• 19:15
  Speaker 3: real estate company and they're like, actually, we're bringing the
• 19:17
  Speaker 3: entire world together.
• 19:19
  Speaker 1: And, you know, we do remember how that story worked out. OK,
• 19:22
  Speaker 1: let's get to elections.
• 19:26
  Speaker 2: As Election Day fast approaches, we are also keeping our
• 19:30
  Speaker 2: eyes focused on what Trump may do if these elections
• 19:33
  Speaker 2: do not go well for him, which they do not
• 19:36
  Speaker 2: look like they're going to go particularly well for him.
• 19:38
  Speaker 2: Ken Klippenstein has a report we can put up on
• 19:40
  Speaker 2: the screen reading some of the tea leaves here and
• 19:42
  Speaker 2: inside what he's describing as doomsday election scenarios that the
• 19:47
  Speaker 2: feds are planning for. I'll read a little bit of
• 19:50
  Speaker 2: this report that he has here. He says the Justice
• 19:53
  Speaker 2: Department is discussing contingency plans to potentially intervene in next
• 19:58
  Speaker 2: week's midterm elections, according to three officials. The plans cover various, quote,
• 20:02
  Speaker 2: doomsday scenarios seen as improbable but not impossible. The Trump
• 20:05
  Speaker 2: administration has hinted at them publicly in recent weeks. National
• 20:09
  Speaker 2: security officials tell me the hints reflect active discussions about
• 20:12
  Speaker 2: what to do if the outcome is disputed over things
• 20:16
  Speaker 2: like foreign interference, claims that non-citizens voted, or chaos at
• 20:22
  Speaker 2: the polls, as with the Justice Department might try to
• 20:25
  Speaker 2: block a state from certifying its election results. Attorney General
• 20:28
  Speaker 2: Todd Blanche told reporters in Norfolk earlier this month, quote, yes,
• 20:31
  Speaker 2: of course, there are scenarios we could all come up with. Hypothetically,
• 20:34
  Speaker 2: you know, some doomsday scenario, for example, right? And they
• 20:38
  Speaker 2: go on to say that even in the DOJ, no
• 20:41
  Speaker 2: one really knows what Trump might do after the election.
• 20:44
  Speaker 2: Of course, we all lived through the whole stop the
• 20:46
  Speaker 2: steal madness after he lost in 2020. So we know
• 20:50
  Speaker 2: that basically nothing is off the table for this man.
• 20:53
  Speaker 2: And we're already getting some indications of steps that are
• 20:56
  Speaker 2: being taken. In particular, this lawsuit in Alaska that's been filed.
• 21:01
  Speaker 2: Let's put E4 up on the screen because this is
• 21:03
  Speaker 2: the most significant action ever. that we've got to date,
• 21:06
  Speaker 2: why legal experts see the Justice Department's Alaska election challenge
• 21:10
  Speaker 2: as dubious. So basically, they filed suit saying, hey, if
• 21:15
  Speaker 2: incumbent Senator Dan Sullivan loses, we think it was unfair
• 21:19
  Speaker 2: because there's another guy named Dan Sullivan who is also
• 21:22
  Speaker 2: on the ballot. And this question has already been litigated
• 21:27
  Speaker 2: in Alaska going all the way up to the Alaska
• 21:29
  Speaker 2: Supreme Court. And they ruled that he could remain on
• 21:33
  Speaker 2: the ballot. And they've they've worded the ballot such that
• 21:36
  Speaker 2: it is made quite clear which Dan Sullivan is the
• 21:40
  Speaker 2: incumbent Republican senator and which is not. However, they are
• 21:45
  Speaker 2: already trying to file this lawsuit in the event that
• 21:49
  Speaker 2: the Dan Sullivan they want does not win the election,
• 21:52
  Speaker 2: that they can make this argument that this was unfair
• 21:56
  Speaker 2: and needs to be contested. So I think Sagar, in
• 21:58
  Speaker 2: a lot of ways, it's real preview. of some of
• 22:00
  Speaker 2: the tactics they're likely to use going forward. And we've
• 22:03
  Speaker 2: already seen Trump raising concerns about, especially about non-citizens voting.
• 22:08
  Speaker 2: They've been in this whole battle with California, but other
• 22:11
  Speaker 2: states as well. So you can see the way they're
• 22:13
  Speaker 2: sort of laying the rhetorical groundwork for at the very
• 22:16
  Speaker 2: least making excuses, but much more likely actually trying to
• 22:20
  Speaker 2: sow chaos and contest election results that they don't like.
• 22:25
  Speaker 3: I think Brazil is actually very instructive for what we're
• 22:28
  Speaker 3: going to see, which is you, I mean, so obviously,
• 22:31
  Speaker 3: I know you guys went into it yesterday, but in
• 22:34
  Speaker 3: the immediate aftermath of it looking like Lula was going
• 22:38
  Speaker 3: to win, the right wing of Brazil, the Bolsonaristos, immediately
• 22:43
  Speaker 3: coalesced around this judge's order, which ordered the polls to
• 22:47
  Speaker 3: open after they had closed. And they were like, look,
• 22:51
  Speaker 3: this makes it obvious that it was rigged. And I
• 22:53
  Speaker 3: won't deny, like, it's weird, right? Because the whole personal
• 22:55
  Speaker 3: lawyer thing. However, if you look at the data, 36,000
• 22:59
  Speaker 3: people were voted after the order went into place. Lula
• 23:04
  Speaker 3: won by 3 million, right?
• 23:05
  Speaker 2: Also, both candidates asked for that delay.
• 23:07
  Speaker 1: By the way. Yes, correct, right.
• 23:09
  Speaker 3: I mean, there was something about buses. But my point
• 23:12
  Speaker 3: is that they're focusing on this, which, you know, potentially
• 23:16
  Speaker 3: calls into question 36,000 votes.
• 23:18
  Speaker 1: Lula won by 3 million, right?
• 23:20
  Speaker 3: And so at the end of the day, the victory
• 23:23
  Speaker 3: margin itself is what speaks for everything. And, you know,
• 23:26
  Speaker 3: you and I talked a lot about this post 2020.
• 23:29
  Speaker 3: The fact that it was close was actually what made
• 23:31
  Speaker 3: it contestable, quote unquote, in the minds of a lot
• 23:34
  Speaker 3: of these people. And you did, you know, the Fulton
• 23:36
  Speaker 3: County thing, the pipe, et cetera. Now, look, this has
• 23:40
  Speaker 3: all been litigated to hell. At this point, if you
• 23:42
  Speaker 3: still really believe something, I mean, Trump yesterday said that
• 23:45
  Speaker 3: there's a million dollar bounty if you can prove election fraud.
• 23:49
  Speaker 3: So for all the people I've met in Ubers over
• 23:51
  Speaker 3: the years who assure me that they have the evidence,
• 23:53
  Speaker 3: go get it, you know, go get it for yourself.
• 23:56
  Speaker 3: In this case, I think what we are going to
• 23:58
  Speaker 3: watch is basically Trump is going to say that the
• 24:00
  Speaker 3: election was rigged. That's just like a given. Obviously, he's
• 24:03
  Speaker 3: going to blame something. And I do think that they
• 24:06
  Speaker 3: will definitely keep these types of lawsuits and others. But,
• 24:09
  Speaker 3: you know, did you read the lawsuit or at least
• 24:12
  Speaker 3: in your times analysis of it, their call for an investigation,
• 24:16
  Speaker 3: the federal court, obviously, or the fact that it was
• 24:19
  Speaker 3: already decided at the Alaska level, the feds court wise.
• 24:24
  Speaker 3: And as we saw post 2020, what was Trump's court record?
• 24:27
  Speaker 3: It was like, oh, for 64 state court. I mean,
• 24:31
  Speaker 3: The federal courts are dramatically deferential to the states on
• 24:34
  Speaker 3: election matters for good reason, which is the states run
• 24:38
  Speaker 3: the elections. I mean, look, pluses and minuses. As I've said,
• 24:42
  Speaker 3: it's kind of ridiculous. In my opinion, you can vote
• 24:44
  Speaker 3: in one state and not in another, depending on if
• 24:46
  Speaker 3: you're a felon or whatever.
• 24:48
  Speaker 1: I think it's stupid. Or, you know, you have universal
• 24:50
  Speaker 1: mail-in voting in one and you don't in another.
• 24:53
  Speaker 3: But part of that is it does make it relatively
• 24:55
  Speaker 3: anti-fragile in terms of actually looking at, you know, in
• 24:59
  Speaker 3: terms of seeing how difficult it actually is for the
• 25:03
  Speaker 3: federal government to do something. I'm not going to deny
• 25:05
  Speaker 3: that they're not going to rhetorically do a lot here.
• 25:07
  Speaker 3: But the other problem that I have been thinking about is,
• 25:11
  Speaker 3: let's say, you know, you have a blowout in Texas.
• 25:14
  Speaker 3: Who are you going to blame if it's a Texas?
• 25:17
  Speaker 1: It's in Texas.
• 25:18
  Speaker 3: Who are you going to like? Are you going to
• 25:20
  Speaker 3: blame the Republican governor and the secretary of state for, quote, unquote,
• 25:23
  Speaker 3: rigging the vote there? Like that's where Michigan, Georgia, Pennsylvania
• 25:27
  Speaker 3: even Let's say in those cases, like, you know, again,
• 25:29
  Speaker 3: you could convince some MAGA grandma out there that there's
• 25:33
  Speaker 3: some like Democrat trying to rig things. Like if the
• 25:36
  Speaker 3: vote goes wrong in Iowa or goes the wrong way
• 25:38
  Speaker 3: in Iowa, like who exactly is to blame here for
• 25:42
  Speaker 3: massive Republican government?
• 25:44
  Speaker 1: Do you see what I'm saying? So it seems just.
• 25:46
  Speaker 3: Like the rhetoric, the rhetoric stuff they were able to employ. especially,
• 25:51
  Speaker 3: you know, consider this too. Before COVID, I get it,
• 25:55
  Speaker 3: is the way you could trick people, right, around the whole.
• 25:57
  Speaker 1: Like, oh, they changed the mail-in ballot.
• 25:59
  Speaker 3: I'm like, well, you know, it was the Republican Supreme
• 26:01
  Speaker 3: Court actually that did that, right, in terms of the ruling.
• 26:03
  Speaker 3: But if you don't know what I just added there,
• 26:06
  Speaker 3: you're like, oh, that sounds sketchy. I mean, we have
• 26:08
  Speaker 3: had years of Republican governance, six years now in these states,
• 26:13
  Speaker 3: which have all responded by tightening the you know, like
• 26:16
  Speaker 3: voter ID and all of these, like Alaska, are you
• 26:19
  Speaker 3: really going to tell me Alaska itself is somehow like
• 26:23
  Speaker 3: immune to all of the Republican effects now over the
• 26:26
  Speaker 3: years to quote unquote secure elections?
• 26:29
  Speaker 1: I mean, it's not.
• 26:30
  Speaker 3: And so they're really going to be a victim really
• 26:32
  Speaker 3: of their own rhetoric if a bunch of these red
• 26:35
  Speaker 3: states end up going blue because they've enacted in many
• 26:38
  Speaker 3: of these cases, the exact sorts of reforms that they
• 26:40
  Speaker 3: have called for for years.
• 26:42
  Speaker 2: Yeah. Well, and to your point, there is a big
• 26:45
  Speaker 2: difference between Trump being on the ballot specifically, given that
• 26:50
  Speaker 2: the level of sway he has over his cult-like following
• 26:54
  Speaker 2: versus like, are you really going to the mat for
• 26:56
  Speaker 2: Ken Paxton and Roger Marshall and Dan Sullivan? You know,
• 27:00
  Speaker 2: it is a little bit of a different deal. So
• 27:02
  Speaker 2: I grant you that. But I do also think we
• 27:05
  Speaker 2: have to be prepared for whatever it is that he's
• 27:08
  Speaker 2: going to try to pull because he is going to
• 27:09
  Speaker 2: try to pull something. Let's go ahead and take a
• 27:12
  Speaker 2: listen to these comments that he's been making about his
• 27:17
  Speaker 2: two monsters that we're going to know about next week
• 27:23
  Speaker 2: and are going to make history next week. I think
• 27:25
  Speaker 2: we know one of them is Russ Vogt because he's
• 27:27
  Speaker 2: talked about this before. But in any case, let's take
• 27:30
  Speaker 2: a listen to E2B Trump on his two monsters. The
• 27:34
  Speaker 2: United States oil and gas business is making hundreds of
• 27:37
  Speaker 2: billions of dollars.
• 27:38
  Speaker 3: So I said, I wonder if I should say this.
• 27:41
  Speaker 2: I'd like a little advice from my two people, my
• 27:44
  Speaker 2: two big monsters that are going to do so well.
• 27:47
  Speaker 3: They're going to be the biggest stars in the country
• 27:49
  Speaker 3: in a week.
• 27:50
  Speaker 2: So I don't know if I should say this or,
• 27:52
  Speaker 2: you know, my two big monsters that are going to
• 27:54
  Speaker 2: be making history next week. We covered before, Sagar, how
• 27:57
  Speaker 2: we talked about how Russ Vogt was going to be
• 27:59
  Speaker 2: so consequential. He, of course, was, you know, the sort
• 28:02
  Speaker 2: of architect of Project 2025, so very influential individual and
• 28:07
  Speaker 2: very hard-nosed, I guess we'll say, tactician and very loyal
• 28:12
  Speaker 2: to Trump. So, He's got something in the works. I
• 28:15
  Speaker 2: don't know exactly what it is. And I do think
• 28:17
  Speaker 2: you're right that if it's too big to rig, then
• 28:21
  Speaker 2: it'll be very hard to successfully sow a level of doubt.
• 28:25
  Speaker 2: But on the other hand, what he does have that
• 28:28
  Speaker 2: he didn't have in 2020 is a much more sycophantic administration.
• 28:35
  Speaker 2: Todd Blanch will do whatever Trump wants him to do.
• 28:39
  Speaker 2: Russ Vogt will do whatever Trump wants him to do.
• 28:41
  Speaker 2: Pete Hegseth will do whatever Trump wants him to do.
• 28:45
  Speaker 2: And that is different from the last time around when
• 28:48
  Speaker 2: you had people like, you know, Bill Barr around to,
• 28:51
  Speaker 2: you know, I find to be an odious character. But
• 28:53
  Speaker 2: at the end of the day, it was like, this
• 28:54
  Speaker 2: is too far. We're not doing this. You had people
• 28:56
  Speaker 2: who were on the inside who were pushing back on
• 28:58
  Speaker 2: some of his most insane plots.
• 29:01
  Speaker 3: I think whenever you look at some of the other
• 29:03
  Speaker 3: stuff that's going on behind the scenes, you're going to
• 29:06
  Speaker 3: I mean, look, I actually I wonder how much that
• 29:09
  Speaker 3: analysis holds true, because, for example, let's put E6, for example,
• 29:13
  Speaker 3: up here on the screen. Obviously, this was a big
• 29:16
  Speaker 3: story over the weekend. The White House is weighing mass
• 29:19
  Speaker 3: pardons for junior staff. I mean, you remember when we
• 29:22
  Speaker 3: covered all of that Rick Grinnell stuff with Alina.
• 29:27
  Speaker 1: Haba and or whatever her.
• 29:29
  Speaker 3: Now married name is, the name of the Turkish billionaire
• 29:33
  Speaker 3: who she married, where it's like literally MAGA infighting and sniping.
• 29:38
  Speaker 3: The vast majority of these people know that if you
• 29:41
  Speaker 3: are going to try and put your name on one
• 29:44
  Speaker 3: of those orders, what happened after January 6th is basically
• 29:47
  Speaker 3: a cakewalk compared to, I think, what is going to
• 29:50
  Speaker 3: happen in the midst of a potential new democratic victory.
• 29:54
  Speaker 3: So you'll remember, for example, you had lawyers who were disbarred,
• 29:57
  Speaker 3: people were thrown into prison. I mean, I do honestly
• 30:01
  Speaker 3: find it a bit difficult to view any sort of like, look,
• 30:05
  Speaker 3: I'm not saying Todd Blanche won't put his name on
• 30:06
  Speaker 3: some bullshit lawsuit. However, if you saw yourself, I don't
• 30:10
  Speaker 3: know if you did. I saw liberal legal scholars already
• 30:13
  Speaker 3: being like any DOJ lawyer who signed this, like we're
• 30:16
  Speaker 3: coming after you. I do think that there's going to
• 30:18
  Speaker 3: be an element of that, as you just said, if
• 30:20
  Speaker 3: the election victory margin continues to trend in the direction. Now,
• 30:24
  Speaker 3: if it gets close, that is going to be a
• 30:26
  Speaker 3: very different story.
• 30:27
  Speaker 1: I'm not going to lie whenever it comes to that.
• 30:30
  Speaker 1: Like you will have them.
• 30:32
  Speaker 3: And again, that just really highlights the specter of 2020,
• 30:35
  Speaker 3: where because it was very close, 100000 votes, a different direction.
• 30:40
  Speaker 3: Trump wins pretty easily in a bunch of these different
• 30:43
  Speaker 3: states that just and that's 100000 combined, by the way,
• 30:46
  Speaker 3: just to make clear how close that election actually was.
• 30:49
  Speaker 3: That is where you could zoom in and against convince people.
• 30:53
  Speaker 3: I think that something went wrong or to pursue some
• 30:55
  Speaker 3: sort of a lawsuit. But I've looked at a lot
• 30:58
  Speaker 3: of these alleged things that they could, quote, try to do.
• 31:03
  Speaker 3: I don't actually see a ton more difference in what
• 31:06
  Speaker 3: happened back in 2020, where ultimately you have to challenge
• 31:10
  Speaker 3: these things in court, which, again, are massively deferential to
• 31:15
  Speaker 3: the state. I guess the only thing that precludes that,
• 31:17
  Speaker 3: like you said, is some sort of genuine emergency operation
• 31:22
  Speaker 3: or something. Which, I mean, I guess, theoretically, you know,
• 31:25
  Speaker 3: is possible. But at that point, we're in a full...
• 31:28
  Speaker 3: We're not even talking about election fraud or whatever at
• 31:30
  Speaker 3: that point. We're talking about a full-blown, like, military coup
• 31:33
  Speaker 3: or something. In my opinion, from what I can see.
• 31:35
  Speaker 1: Completely.
• 31:36
  Speaker 2: No, I mean, you can imagine a scenario where it's...
• 31:40
  Speaker 2: Democrats win the Senate 49-51. And, you know, there's an
• 31:45
  Speaker 2: ability for Trump in, let's say, Michigan. It's close to Michigan.
• 31:49
  Speaker 2: And so he's able to, like... make some case and
• 31:53
  Speaker 2: you got democratic government, very much similar to what happened
• 31:55
  Speaker 2: in 2020. But based on where the polls are and
• 31:58
  Speaker 2: what things are looking like, it's likely the margin is
• 32:01
  Speaker 2: going to be large enough. It's going to be more
• 32:02
  Speaker 2: difficult for him, but it doesn't mean there isn't going
• 32:04
  Speaker 2: to be a lot of a big chaotic mess. And
• 32:08
  Speaker 2: Professor Pape has actually been sounding the alarm about his
• 32:10
  Speaker 2: concerns about what might happen post-election. So just wanted to
• 32:14
  Speaker 2: keep our eye on some of the machinations that are
• 32:16
  Speaker 2: happening right now behind the scenes.
• 32:19
  Speaker 1: Absolutely.
• 32:19
  Speaker 3: Thank you for everybody who's watching, bearing with our illnesses.
• 32:23
  Speaker 3: We will see, I actually, Emily and I will see
• 32:26
  Speaker 3: you tomorrow for the first part. And then Ryan's going
• 32:28
  Speaker 3: to come on a little bit later. He's got a
• 32:30
  Speaker 3: big interview with Emily that he's going to do. So
• 32:33
  Speaker 3: we'll see you tomorrow.