Even less of a done deal—back on the Paramount/Warner beat.

See here and here for the rest of the thread.

the thing is all state AGs have to do to deep-six this merger is delay it until OpenAI can't pay Oracle, Oracle's stock dives, Larry's collateral goes poof, etc.

[image or embed]

— Christopher Mims (@mims.bsky.social) July 20, 2026 at 11:09 AM

(Christopher Mims is a longtime tech columnist for The Wall Street Journal, just so you know I'm not quoting some rando here.)

In case you missed it, here's the story from the Los Angeles Times.

[Emphasis added]  

Paramount-Warner Bros. deal on hold after court ruling - Los Angeles Times
Meg James


Hollywood’s biggest deal in decades is on hold.

On Monday, a federal judge temporarily blocked Paramount Skydance’s efforts to complete its purchase of Warner Bros. Discovery, ruling that the proposed $111-billion merger “raises serious questions” about whether the combination violates U.S. antitrust law.

District Judge Araceli Martínez-Olguín, based in Oakland, granted a request for a temporary restraining order from a coalition of 12 state attorneys general, led by California Atty. Gen. Rob Bonta, to freeze the deal while the court delves more closely into its impact on markets.

The order pauses the deal for 14 days. Martínez-Olguín’s ruling sets up a showdown for Aug. 3, when she considers a motion for a preliminary injunction — which, if granted, could tie up the deal for months in advance of a trial.


...

“The judge basically said, ‘Look, let’s not race to the finish line here,’” Eric Talley, a Columbia Law School professor, said in an interview. “At the end of the day, maybe this thing gets signed off on, but I think the AGs are going to be given a fair chance to bring their claims forward.”

...

“Plaintiffs present compelling evidence that the combined firm resulting from the transaction will possess substantial market share in the wide-release theatrical distribution market,” Martínez-Olguín wrote in her 10-page order.

If allowed to merge, Paramount-Warner Bros. would control about 27% of the market of films that are initially released into more than 3,000 theaters.

“On this combined firm market share alone, the Court is persuaded that it can presume the proposed merger is likely to violate antitrust laws,” the judge wrote.

The ruling doesn’t signal that the states will win but, Talley said: “This is an important mark in the road that suggests that, in the eyes of the judge, at least one of their allegations has the seeds of a valid case.”

 

Here are some notes, roughly in order of importance.

No matter how things turn out, I should have at least a few months before I have to cancel HBO Max. I will make it through their Janus Films collection.

As mentioned before, along with SoftBank, Oracle is the major corporation most vulnerable to even a partial collapse of the AI bubble. The company has gone all in on data centers, is drowning in debt, has seen its credit rating slashed, and has an incredibly volatile stock, which is currently down more than 50% from its recent high on June 1.

Larry Ellison is contractually committed to pony up $47 billion when the deal closes.

Even the very rich are seldom liquid enough to easily cough up that kind of cash. The standard solution is to take out a loan with stock as collateral. Unfortunately, Ellison already has considerable debt, probably structured in such a way that another serious downturn in the stock could trigger a round of margin calls, which might force him to sell a large chunk of Oracle stock, which would further depress the price, which could trigger additional margin calls. You see where this is going.

In other words, Ellison's ability to close the deal depends on Oracle holding most of its value.

Then there's the ticking fee. Part of the merger agreement stipulates that if the deal is not closed by September 30, Paramount has to pay $650 million a quarter to Warner stockholders. Admittedly, this doesn't seem like a whole lot compared to the billions we've been talking about, but Paramount is already deeply in debt and, as the saying goes, "A half a billion here, a half a billion there, pretty soon you're talking about real money."

And then there's this:

A longer delay would have huge financial costs for Paramount. Starting Oct. 1, Paramount has to pay Warner shareholders a "ticking consideration" of roughly $650 million for every 90 days the deal is set back. If the deal is not consummated by June 4, 2027, Paramount will have to pay Warner $7 billion.

 

The best-case scenario for the Ellisons is an expensive and painful merger process. The middle case is that the merger does not go through. The worst possible case is that a collapse of Oracle's valuation will force the father-and-son team to start selling off assets like TikTok and Paramount.

I'm not going to assign any kind of probabilities here, but given the tone of the conversation around the merger as recently as a couple of weeks ago, this is a remarkable turn of events.

"The media is important for many, many reasons, but one of the biggest ones is that scrutiny is what keeps capital in check, for the benefit of humanity and at times the companies themselves."

 Ed Zitron has a new epic post out, massive post out at Where's Your Ed at? To be honest, it's probably a bit much unless you've already fallen deep into the AI bubble rabbit hole. New comers may want to start with his Better Offline podcast where he regularly talks with some of the sharpest AI bubble skeptics  Cal Newport, Paul Kedrosky, and Carl Brown. He has also has become a fixture on shows like On the Media (an interview from last year. OTM was characteristically early to the story) and CNBC's Squawk Box looking for a counterweight to the stock pumpers. 

This section in particular caught my attention. 

 

From The OpenAI Bubble: 

The double-edge sword of a mythology-inflated bubble is that it’s much harder to sustain when said mythology dies. The AI bubble was able to grow to such a horrendous size because the markets and the media were willing to accept basically anything that Sam Altman or the greater AI industry said. 

By waving away any economic problems as growing pains and dismiss those who would scrutinize it as haters or cynics, reporters and analysts provided investors with the justification to invest again and again in these companies without them ever having to make a real business, which means that, well…they don’t have real businesses, which is a problem when you need to actually pay somebody money that wasn’t given to you by a venture capitalist.

This will leave the AI industry short-changed in its most-desperate times. 

The media is important for many, many reasons, but one of the biggest ones is that scrutiny is what keeps capital in check, for the benefit of humanity and at times the companies themselves. By choosing to pull their punches, ignore glaring economic problems and accept every projection with blind faith, the media empowers grifting and suffocates good businesses as a result, encouraging bad behavior and helping them raise unbelievable amounts of money at ridiculous valuations without worrying about having to make a good business. In some cases, the media even encourages them to do so, saying that “all startups lose money at first” instead of thinking about things for a fucking second.

When companies know they won’t face that scrutiny, they engineer themselves as such, putting off ever finding a real business model in favor of whatever will make them buzzy enough to get coverage and raise funding as a result. In a vacuum of skepticism, bubbles inflate, monsters get rich, and regular people always get left holding the bag. As a result, if companies ever bother to become a real business, they only do so at the very last minute, endangering anyone who has backed them and every counterparty in the event they’re incorrect.

 

If things get really bad, Ellison can always sell that island back to Hawaii

“.. hard to overstate how critical $ORCL is to the entire AI narrative. No big company has levered itself (figuratively and financially) more .. the stock's horrendous trading performance underscores heightened anxiety about how it can profitably build all the infrastructure ..” - Vital Knowledge

[image or embed]

— Carl Quintanilla (@carlquintanilla.bsky.social) July 16, 2026 at 7:59 AM

 Picking up our Paramount/Warners thread.

 Ed Zitron has an epic length post on the pivotal role and precarious state of OpenAI. Among other things, he points emphatically and in great detail a fact that lots of smart people have been saying quietly, that if that company collapses it is likely to take down one of tech's biggest and most established players* along with one of the world's largest fortunes. 

Oracle is currently spending over $340 billion to build out over 7.1GW of data center capacity for OpenAI, as part of its $300 billion, five-year-long cloud compute contract that began, at least in theory, on June 1, 2026 at the beginning of its Fiscal Year 2027, though much of the capacity is yet to be built. To fund the buildout, Oracle has had to raise over $50 billion via stock sales and debt, spent $55.7 billion in its last fiscal year, and expects to spend at least $90 billion more in FY2027.

As a result of that, S&P Global downgraded Oracle’s credit rating to BBB/A-2, the literal lowest level before it’ll become junk-grade, meaning that one more downgrade (though it would have to be from two ratings agencies) from here would risk Oracle becoming a “fallen angel,” with investment funds (that can’t hold junk grade debt) having to jettison its debt from indexes, as happened to Ford in March 2020, leading to over $35 billion in debt being dumped and its borrowing costs skyrocketing to between 8.5% and 9.625% when it raised in April 2020. For some context, Ford reported an average interest rate of 5.2% on its long term debt in its 2019 annual report.   

You’ll never guess why S&P Global downgraded Oracle! And, once again, the emphasis is theirs:

OpenAI remains a key credit risk. We estimate that OpenAI makes up roughly half of the $638 billion in RPO. OpenAI’s ability to meet its contractual obligations and raise external financing will be contingent upon AI tailwinds continuing and its models being market leaders. If OpenAI were unable to pay Oracle, we believe Oracle could be left with massive data center leases that it might be unable to exit or have to re-lease to new tenants under less-favorable terms. As a proxy for OpenAI’s future prospects, we’re tracking OpenAI’s financial commitments to data center operators and chip makers to gauge its overall financial exposure and its market share among enterprise and consumers.

That’s a load-bearing if, brother! 

...

As a reminder, the only way that OpenAI will be able to afford to pay its $300 billion cloud compute contract with Oracle will be if it continues to hit revenue projections (per The Information) that have it making $113 billion in 2028, $184 billion in 2029, and $284 billion in 2030, a year when it will magically become profitable, and no, I don’t know how that happens:

 ...

Generative AI is the only reason that Wall Street started liking Oracle again as its other business plateaued, even as it burned billions of dollars on capital expenditures and cut its gross margins by a little under 15% since 2022, with the vast majority of that value coming from its revenue from OpenAI and what’s actually active at Stargate Abilene. 

... 

As I said in my piece about how OpenAI Kills Oracle: 

[The collapse of Stargate and OpenAI would be] a very bad thing for Larry Ellison, who holds around 40% of Oracle’s shares and receives a dividend of around $2.3 billion a year as a result, especially as he’s backed the $111 billion Paramount-Warner Brothers Discovery merger deal with $45.7 billion of that as an equity commitment from the Ellison Trust (the Ellison family investment arm which holds his Oracle shares) with Larry himself guaranteeing the amount, with $24 billion of those funds likely coming from the Middle East. 

This leaves the Ellison family with around $12 billion left to fund the deal. Depending on how liquid the trust is, it could foreseeably fund that in cash, but if Ellison is a little light, he might have to take out further margin loans on his Oracle stock. 

Yes, I used the word “further.” Ellison has already pledged 346 million shares of his Oracle stock — or around $61.5 billion — “to secure certain personal indebtedness, including various lines of credit,” meaning “many big, beautiful loans against his Oracle shares.” which IFR estimated back in September (when Oracle’s stock price was much higher) could allow him to secure as much as $21.4 billion in debt at a (they say “conservative”) loan-to-value ratio of 20%, and that’s assuming the banks weren’t particularly generous.

One of the consistent themes of this piece is that much of the “value” of AI is hot air — by which I mean whatever people are willing to pay for a stock that’s continually inflated by specious media-driven hype. 

...

Things could get much darker if Oracle plunges below $50, as at that point the encumbrances of his various enterprises and his own margin loans could become too much to avoid having to liquidate Oracle stock. If that happens, it creates a vicious cycle that will potentially involve selling off Paramount, dumping further Oracle shares, or even trying to engineer a firesale for the company.

Sidenote: While it’s true that Oracle’s software is economically important — its database systems and ERP platforms power a bunch of big businesses and government organizations — I don’t believe that any external intervention (whether that be an external investor chucking it some cash, or some form of bailout) would be able to stop the pain that’s coming to it. 

Simply put, the bets it made are too big — and, economically important Oracle’s software might be, there’s no reason that said software couldn’t continue development under the stead of another company.

All of this was entirely avoidable if he had never met Sam Altman, and never gave in to the temptation of the AI trade.

 

* Caveat. Oracle will survive in some form: "While its many government contracts and national security significance make it unlikely that Oracle would be allowed to die, the collapse of its only growth segment will likely spell dark times for a company that’s already laid off 21,000 people as a means of funding its AI buildout." 

We haven't heard much about oil reserves lately. I'm sure everyting's fine.

Remember a little over three weeks ago when we ran a post about CNN's report on the oil reserves in Cushing, Oklahoma. We used this as an example of how highly important aspects of a major news story like the war in Iran, while not ignored, are covered only sporadically, while relative trivia runs constantly.

The CNN article was very good, but in the weeks that followed, other news organizations largely ignored it. Week after week, non-stories, often involving some unsupported and highly unlikely claim from the White House or yet another analysis of the quagmire, ran constantly at the top of the fold in the New York Times and its competitors.

Unless you were making a concerted effort to dig into the numbers, you could easily get the impression that an uptick in traffic through the Strait of Hormuz and other measures had brought the world's oil supply, if not back to where it had been, then at least well out of the danger zone.

You would be wrong.

From the Financial Times:

The International Energy Agency on Friday said its member countries had released almost three-quarters of the planned 400mn-barrel emergency stock release that was announced in March, meaning there are only a few more weeks to go before those supplies to the market dry up.

“We’ve burned through all of the buffers we had. Everything,” said one trader. “All of that’s now gone.”

...

    During the four-month closure before last month’s US-Iran agreement to reopen the strait, governments in the west and Asia pulled almost every lever available to them to ensure the supply crunch did not undermine the world economy.

Western powers released record volumes of strategic oil reserves, China cut its oil imports in half and made its state-backed companies pull fuel from inventories, while the White House let it be known the US could, in theory at least, intervene in futures markets if prices got out of hand.

The result was that Brent crude peaked at $126 a barrel in April, well below its all-time high, despite the IEA warning that the world was experiencing the worst supply disruption in history.

But traders said that if the renewed closure of the strait lasts for months, with some suspecting Iran wants to keep the pressure on Trump ahead of the November midterm elections, it is not clear this time where the oil to make up the shortfall would come from.

Amrita Sen, director of market intelligence at Energy Aspects, said that going into the war, the oil market had roughly 400mn barrels of excess inventories, not including strategic reserves controlled by governments.

“Now we have close to nothing,” she said. “Market complacency around Hormuz flows is being severely tested.”

 ...

    “Ultimately, the market was pricing an optimistic flow trajectory that now is clearly not on the table, at least . . . not until we get another round of diplomacy,” said Joel Hancock, a senior commodities analyst at Natixis Bank.
 

 

 

As previously mentioned, sometimes it's less about the facts and more about who's saying them.

If you've been following AI boom skeptics like Ed Zitron, you've heard this a dozen times in more detail and in far less dry language.

Over the past year or so, this message has been increasingly making its way into the discourse of mainstream economists and publications like the Financial Times, but it's still notable when a collection of the world's central bankers starts sounding just a bit panicked, albeit in their characteristically stuffy way. 

From the Bank of International Settlements (BIS)  annual report:  

In the near term, the ongoing AI investment boom raises questions about the sustainability of the current economic expansion. The five largest hyperscalers are set to spend over a trillion US dollars on AI-related capital expenditure from 2025 through 2026. These commitments are outpacing earnings and the free cash flow of these firms, leading some to issue debt to raise additional financing.

...

 Disappointment in returns could trigger a sudden pullback in financing and turn the capex boom into a protracted investment bust, with potential knock-on effects on financial conditions…should hyperscalers slow or halt the aggressive pace of capex deployment, many borrowers across the supply chain could struggle to replace lost revenue and service their debt.

The long-term commercial impact of large language models is a topic for another day. It should be noted that the BIS report is cautiously optimistic on this timeframe (I would put myself down in the "big, but not nearly big enough" category), but the concerns about the near-term future of the AI bubble as currently configured are definitely growing more widespread.

A fun little video from science writer and YouTuber Kyle Hill

... going through a recent study from JPL that takes a serious look at the unserious idea of terraforming Mars. At the risk of a spoiler, the conclusion is not encouraging.

While everyone reading this probably already knew how this was going to come out, it's worth walking through the numbers to get some sense of the full absurdity of what Elon Musk and his fellow travelers have been spouting on this.

One particularly silly point that the video doesn't really dig into is Musk's fondness for framing Martian colonization in terms of a Plan B, a place that humanity can go if Earth becomes uninhabitable. The standard hypothetical here would be an asteroid strike or some other disaster. The problem with that idea is that, despite all sorts of cataclysmic events, Earth has been more habitable than current-day Mars for 3 billion or so years, and it's difficult to come up with a scenario that would change that.

Though it's a rather silly exercise, if you were really concerned with humanity surviving the apocalypse, you could find all sorts of safe harbors to hide out in, either deep in the Earth or under the oceans. It would neither be that challenging nor that expensive, though it does have the downside of actually being doable.

Of course, the push for Martian colonies from the techno-optimists, like the enthusiasm for humanoid robots or hyperloops, has nothing to do with engineering, or economics, or any other practical concerns. These things are all inspired by a bizarre fascination with juvenile post-war science fiction combined with generally fascist-leaning libertarianism and the constant need for grifting.

Terraforming Mars is an Industrial Nightmare 

A not-so-done deal

 

Pick up where we left off in June. 

As is so often the case, this story in the New York Times is less of interest for the facts, which have been reported earlier and in greater depth by other news organizations such as NPR, Reuters, CNBC, Variety, The Hollywood Reporter, and others, than it is for what it tells us about the standard narrative.

 States Prepare Lawsuit to Block Paramount’s Merger With Warner Bros.

David McCabe, Benjamin Mullin, Lauren Hirsch

 

July 12, 2026

A group of states are preparing to file a lawsuit to block Paramount’s acquisition of Warner Bros. Discovery as soon as this week, according to four people briefed on the plans, a legal challenge that would create a major obstacle for one of the biggest media mergers in history.

A draft of the lawsuit currently circulating argues that the $111 billion deal would harm competition in the market for so-called tent pole films, the expensive blockbusters that make up a large portion of studio revenues, among other claims, two of the people said.

California has taken the lead on the lawsuit, and states including New York, Washington and Connecticut have said they will join the effort, according to three of the people, as well as another person familiar with the states’ plans. All of them spoke on condition of anonymity to discuss a sensitive legal matter before it was public.

[I don't want to get too sidetracked into bitching about the Gray Lady, but how can it take three credited reporters to write an account so superficial it doesn't even mention the state which has most aggressively pursued this story?] 

Once the lawsuit is finalized, the states could decide to delay filing it or scrap it completely. Reuters earlier reported states could sue as soon as this week.

A spokeswoman for Paramount said in a statement that the company was prepared to address “legitimate antitrust issues,” adding that its merger with Warner Bros. Discovery “raises no such concerns.”

“We are confident the facts and the law support this transaction, and we will continue to defend it vigorously,” she added.

Paramount has said it plans to close the deal in the third quarter of the year. As part of its deal with Warner Bros. Discovery, Paramount has said it would pay the company’s shareholders about $650 million in cash for each quarter the deal doesn’t close, starting in October.

...

 

Internationally, the company has already secured approvals from more than 20 countries and regions, including China and Australia. Some international regulators must still approve the deal, including Britain. In June, a British official said her government was leaning toward examining the acquisition.

...

In filings last month related to a separate brought by streaming subscribers seeking to block the deal, Paramount executives said in sworn declarations that they planned to release at least 30 movies in theaters annually and keep new releases in theaters for at least 45 days before putting them on streaming platforms.

[A bit more bitching. It's true that  Paramount executives said this but literally ten minutes of research would reveal that virtually no one outside of the company believes this is possible. This huge surge in production would have to happen while the company is forced to slash costs to deal with $79 billion in debt just as Fitch has just downgraded the company's credit rating to junk status.] 

Up until recently, the Ellisons and their allies had managed to successfully sell the idea that the merger was a done deal. Resistance was futile. It is time to accept your new overlords. This created a situation where the greatest barrier the opposition faced was the belief that there was nothing the opposition could do. When the NYT runs something like this, you know that narrative is shifting.

Of course, positive thinking only gets you so far. The odds of the merger going through are still quite good, just not the slam dunk that much of the press has credulously repeated up to this point.

There are all sorts of fascinating moving parts here that you would know about if you got your news from the previously mentioned news organizations. This is more than just an antitrust story, with Middle Eastern sovereign wealth fund money making up something like half of the funding, which normally would be considered illegal. There is the previously mentioned Oregon investigation into corruption. Then there are all sorts of interesting aspects to the financing and the timing. The "ticking fee" is still a ways off, but it's not trivial, particularly if Ed Zitron proves to be prescient with respect to the near future of Oracle. Larry Ellison has committed to put up tens of billions of dollars that he might not have at some point.

A lack of qualifications is not a qualification and other thoughts on Platner

Picking up from Joseph's post.

Not being from Maine, I didn't spend a lot of time reading or writing about the Senate primary. Here are, I believe, the entirety of my social media posts on the subject. If you look closely, you might spot something of a common theme.

Anyone else getting a Fetterman vibe here?

— markpalko.bsky.social (@markpalko.bsky.social) October 25, 2025 at 9:40 PM

Sinema was an unconventional candidate who was going to bring in young voters. Fetterman was an unconventional candidate who was going to bring in young voters. Starting to question that particular argument.

— markpalko.bsky.social (@markpalko.bsky.social) October 24, 2025 at 8:56 PM

Like I said: Sinema was an unconventional candidate who was going to bring in young voters. Fetterman was an unconventional candidate who was going to bring in young voters. Platner was an unconventional candidate who was going to bring in young voters. Starting to question that particular argument

[image or embed]

— markpalko.bsky.social (@markpalko.bsky.social) January 8, 2026 at 12:54 AM

You would think the normal reaction to how Fetterman turned out would be "how can we avoid this happening again?" not "let's give it one more try."

[image or embed]

— markpalko.bsky.social (@markpalko.bsky.social) March 4, 2026 at 8:09 PM

I still think the relevant comparison here is Sinema/Fetterman. We see the same anti-establishment framing, the same arguments about the need to select an outsider, in many cases the same people pushing the candidacy. What I want to know is how people still think this is a good idea.

— markpalko.bsky.social (@markpalko.bsky.social) April 12, 2026 at 3:49 PM

He's an enormously problematic candidate brought to us by literally the same people (such as Morris Katz) who brought us Fetterman, using many of the same arguments. If you can convince me that Fetterman was a good choice, I'll send Platner a check.

— markpalko.bsky.social (@markpalko.bsky.social) March 4, 2026 at 8:58 PM

All sorts of big issues are raised by this story—too many to list—but for now, here is a quick list, starting with the recurrent theme of my social media thread.

One of the silliest and most dangerous political ideas to emerge in the past few years is the notion that the best qualification for high public office is a lack of qualifications.

Not being a professional politician has a certain gut-level populist appeal, but it is almost never a positive, particularly when it is combined with a lack of any relevant experience. And no, "my parents bought me an oyster farm" is not sufficient.

This raises another point that I find particularly annoying. I grew up, if not poor, then certainly frugal, in the rural South. Perhaps for that reason, I have always had exceptionally little patience with the fake man-of-the-people schtick. It shows a cynicism and a disdain for voters that I find deeply offensive.

Poor-boy cosplay also provides an area of vulnerability in the general election, which leads to yet another issue. Non-politicians are almost always non-vetted. While it's true we can't always assume the flip side—that elected officials have been well vetted (Swalwell slipped past voters for years)—as a general rule, the bigger and tougher the races someone has won, the more difficult it is to dig up dirt.

It's important to note that Platner never even made it to the general election, where the really nasty rat-fuckers were sharpening their knives. Instead, the damaging stories started coming out almost immediately, possibly even organically, and simply picked up steam. It was less like a drip, drip, drip and more like a growing stream, with progressively worse details constantly emerging. I'm reminded of the scene in the Buster Keaton film where, starting with a few pebbles, progressively larger rocks come tumbling down the hill toward the hapless protagonist.

Much of this dirt was disturbingly close to the surface. A halfway diligent background check, along with a few hours of Internet searching, would have revealed much of it. A competent campaign would have been able to get ahead of many of these stories rather than just trying to dodge each rock individually.

Good journalism, another key component of a well-functioning primary system, would have brought lots of these things out in a more timely manner. If you read papers like the New York Times today, you will find all sorts of sharp reporting focused on the failed candidate, but at the time the vast majority of news organizations were more than content to simply play along with the simple "oyster man takes on the political establishment and wins" narrative.

It wasn't until after he became the presumptive nominee that we started seeing the right questions being asked, and not until the days after the vote that we saw aggressive digging.

Last July, in a small town in coastal Maine, a couple of progressive, self-styled recruiters of economic populists showed up at the blue-shingled house of Graham Platner, a little-known oyster farmer and Marine veteran who lived largely off government benefits.

They knew his name from local labor organizers and activists, and they had watched a video on the internet of him talking about oysters. Struck by his left-leaning ideology, his working-class affect and his gravelly voice, they became convinced that he could win a Senate seat in Maine — and quickly persuaded Mr. Platner of the same.

The initial headhunters, Dan Moraff and Leanne Fan, and then a third out-of-state operative they called up to Maine — Morris Katz — told Mr. Platner he was “the one,” a “hero of the movement,” “a historical figure” who could be “leading a revolution,” according to half a dozen people with knowledge of their conversations.

But a clutch of people who cared about Mr. Platner were telling him something else. They worried about his mental health, amid his ongoing efforts to heal from post-traumatic stress disorder after tours in Iraq and Afghanistan. They feared this trio of out-of-state operatives was a dangerous combination of inexperienced and overconfident. The worst-case scenario, they thought, wasn’t running for Senate and losing — it was destroying the life he worked hard to build.

 

 The NYT started this piece with their usual self-congratulatory spiel talking about how thoroughly they investigated the story, but while I appreciate professional reporting, I'd appreciate professional reporting done in a timely fashion (say eight months ago) a hell of a lot more. At the risk of being both preachy and obvious, one of the most important roles of journalism is informing voters. Waiting until the die has already been cast is often easier and safer, but it defeats the purpose.

The Maine primary was a textbook case of how not to select a nominee. What needs to be a process of scrutiny and skepticism, especially with an untested candidate, instead became a bizarre rush to declare the matter settled months in advance, even while increasingly damaging revelations were steadily coming forward. There seemed to be a widespread fear that if we let the primary do what it was supposed to, we would somehow damage our chances of unseating Collins.

For me, that might be the final and most important lesson. When you tell yourself that a vigorous democratic process is too risky, you are wrong on principle and stupid in practice.

A concession to Mark about outsider candidates

This is Joseph.

So, credit where credit is due, Mark called the issues with Graham Platner early on. But one thing that still amazes me about all of this is the strange alignment of the left in it. 

Matt Yglesias:

And 

It really goes to the central question of "what are the goals of the organization?" and I think it is quite clear that it isn't a resistance movement against Donald Trump. Why am I so sure? 

When is the last time you saw people protesting Trump over Gaza? Has the Israeli occupation improved greatly since he took power? But you see them protesting the out of power Democrat senators from New York. There was some sort of plausible ambiguity when Biden was president. Perhaps the protesters were so certain that Biden (or Harris) would win that they focused on shifting policy with-in the Democratic party. This would be naive but at least plausible. Meanwhile, the current administration actually started a war with a different middle eastern country (Iran) and we still don't have that as the focus. 

But it is now looking a lot like a faction prioritizing power in the Democrat's coalition at the expense of fighting MAGA, and yet branding themselves as the fighters. This strategy seems dangerous in a moment where the current president is very much opposed to all of the things that this faction seems to want. I suppose there could be a "heighten the contradictions plan" but, keep in mind, contradictions can be heightened for decades (or even centuries) causing much misery and no immediate revolutions. 

I think we should listen to Mark and be much more skeptical about outsider candidates without background, vetting, or clear policy objectives. 

P.S. And the candidate is not going quietly which only reinforces the intuition that the game is not winning versus the Republicans . . . 

I probably should have worked in a Matrix reference

Following up on this week's sci-fi, folklore, magic theme.

From Internet of Bugs Carl Brown:

But predicting what humanity would do in the future is only the beginning of the powers of a Super-AI.

I’ll quote here from “The AI Does Not Hate You” by Tom Chivers about a future AI called “Basilisk”:

...the Basilisk is saying, ‘If you work to bring me about as fast as possible, I won’t create a perfect copy of your mind and torture it for billions of subjective years.’ (The argument is that since a perfect copy of your mind would essentially be you, this is equivalent to bringing you back to life.) In essence, a thing that doesn’t exist yet may be blackmailing you from the future, threatening to punish you for not working hard enough to make it exist.

I only wish I were kidding. This belief (which is often called “Roko’s Basilisk”16 ) is even more unhinged than it sounds, and it comes from the website of one of the authors of IF ANYONE BUILDS IT EVERYONE DIES, and that author “Eliezer Yudkowsky, banned discussion of Roko’s basilisk on the blog for several years as part of a general site policy against spreading potential information hazards.”17 Luckily for posterity, those discussions are retained on the Internet Archive18, and the original post can be found replicated elsewhere19, so if you really want to - although I don’t recommend it - you can follow the links I’ve included below so you can see what kind of crazy underlies all this SuperAI nonsense.

Now, I want to be clear - there are many different versions of Roko’s Basilisk, and there are some people in the Doomer community who will say they don’t believe in any version of it. Despite that, the story about the Basilisk is still indicative of the powers that rationalists ascribe to ASI [Artificial SuperIntelligence --  MP]. According to an article in Slate20: “Yudkowsky said that Roko had already given nightmares to several LessWrong users and had brought them to the point of breakdown.” The Basilisk is also often treated as at least plausible in public discussions of SuperAI. And when the people outside the AI community or industry are exposed to this idea, even if they are told not to take the idea that an AI would torture them seriously, they are still left with an impression of an impossibly powerful AI that conceivably could bring people back to life to torture them. This isn’t harmless - this sticks in people’s heads and colors their ideas about AI thereafter.

Though this may border on staring-into-the-abyss territory, let's take a moment and think about what actually terrorized all these True Believers.

We won't even try to examine the argument that a future AI would go to all this trouble to do something that couldn't possibly influence past events. Instead, let's focus on the threat itself. Imagine you were told that there was a possibility that someone will build a computer simulation where a character based on you would be tortured. Does this thought terrify you? This is literally what we're talking about here.

Admittedly, the possibility that characters in a computer simulation might be sentient and self-aware is potentially an unsettling thought (it's also an excuse to post one of my favorite Doctor Who moments below), but the fact that one of these characters has your name shouldn't be that big of a deal. But, of course, the people who are deeply disturbed by this, those who find it haunting their nightmares, are not approaching it from any kind of scientific or logical standpoint.

There are definite echoes of Harlan Ellison's I Have No Mouth, and I Must Scream. That's a story about a non-omniscient computer that has to settle for torturing the last surviving humans. There's a much older precedent to the basilisk.

This is a story of an all-powerful being who brings you back to life after you die and punishes you through all eternity for making it angry. The story of the Basilisk is one of a vengeful God casting you into hell, nothing more, nothing less.

As a non-religious person, I try to be respectful of faith based beliefs (and I will delete replies that strike me as disrespectful), but this is something different. this is people believing in magical things then trying to convince us and themselves that they are actually the true rationalists.

The sad part is that this belief in a cyber Hades is neither the silliest nor the stupidest aspect of the AI movement. The rot and the absurdity run throughout the whole damn thing. 

These people are as flaky as your grandmother's best biscuits.

The idea of approaching tech narratives as folklore actually works better than it did eight years ago.

Yesterday we did a post on how most of the ideas in the techno-optimist movement are just old sci-fi tropes. Tomorrow we'll be talking about some of the AI movement's more cult-like aspects (wait till we get to “Roko’s Basilisk”). This seems like a good transition.

Tuesday, June 26, 2018

I'm afraid even the Brothers Grimm would have found Bitcoin a little too fantastic

I'm edging closer to the notion that the tools which we would normally use to critique journalism are no longer up to the task of discussing the 21st century technology narrative. Instead, the appropriate methods are probably those of the folklorist. We are rapidly approaching the realm of the myth and the tall tale. Why not start thinking in those terms?

It is standard practice when discussing something like a Jack tale to list the Aarne–Thompson classification. For example, Jack in the beanstalk fall under the classification AT 328 ("The Treasures of the Giant"). We could do something similar with the vast majority of tech reporting -- Take Theranos -- This and other accounts of college dropouts supposedly coming up with some amazing innovation can be classified under "wayward youth finds magic object."

 I've been giving quite a bit of thought recently to how magical heuristics have come to dominate the conversation about technology and innovation, but the idea of actually treating the narrative as folklore didn't hit me until I read this:

The paperclip maximizer is a thought experiment described by Swedish philosopher Nick Bostrom in 2003. It illustrates the existential risk that an artificial general intelligence may pose to human beings when programmed to pursue even seemingly-harmless goals, and the necessity of incorporating machine ethics into artificial intelligence design. The scenario describes an advanced artificial intelligence tasked with manufacturing paperclips. If such a machine were not programmed to value human life, then given enough power its optimized goal would be to turn all matter in the universe, including human beings, into either paperclips or machines which manufacture paperclips.[4]

    Suppose we have an AI whose only goal is to make as many paper clips as possible. The AI will realize quickly that it would be much better if there were no humans because humans might decide to switch it off. Because if humans do so, there would be fewer paper clips. Also, human bodies contain a lot of atoms that could be made into paper clips. The future that the AI would be trying to gear towards would be one in which there were a lot of paper clips but no humans.
    — Nick Bostrom, "Ethical Issues in Advanced Artificial Intelligence", 2003

Bostrom has emphasised that he does not believe the paperclip maximiser scenario per se will actually occur; rather, his intention is to illustrate the dangers of creating superintelligent machines without knowing how to safely program them to eliminate existential risk to human beings. The paperclip maximizer example illustrates the broad problem of managing powerful systems that lack human values

Suddenly it struck me that this was just the magic salt mill ever so slightly veiled in cyber garb. In case you're not up on your folklore...

It is Aarne-Thompson type 565, the Magic Mill. Other tales of this type include The Water Mother and Sweet porridge.

Synopsis

A poor man begged from his brother on Christmas Eve. The brother promised him, depending on the variant, ham or bacon or a lamb if he would do something. The poor brother promised; the rich one handed over the food and told him to go to Hell (in Lang's version, the Dead Men's Hall; in the Greek, the Devil's dam). Since he promised, he set out. In the Norse variants, he meets an old man along the way. In some variants, the man begs from him, and he gives something; in all, the old man tells him that in Hell (or the hall), they will want to buy the food from him, but he must only sell it for the hand-mill behind the door, and come to him for directions to use it. It took a great deal of haggling, but the poor man succeeded, and the old man showed him how to use it. In the Greek, he merely brought the lamb and told the devils that he would take whatever they would give him, and they gave him the mill. He took it to his wife, and had it grind out everything they needed for Christmas, from lights to tablecloth to meat and ale. They ate well and on the third day, they had a great feast. His brother was astounded and when the poor man had drunk too much, or when the poor man's children innocently betrayed the secret, he showed his rich brother the hand-mill. His brother finally persuaded him to sell it. In the Norse version, the poor brother didn't teach him how to handle it. He set to grind out herrings and broth, but it soon flooded his house. His brother wouldn't take it back until he paid him as much as he paid to have it. In the Greek, the brother set out to Constantinople by ship. In the Norse, one day a skipper wanted to buy the hand-mill from him, and eventually persuaded him. In all versions, the new owner took it to sea and set it to grind out salt. It ground out salt until it sank the boat, and then went on grinding in the sea, turning the sea salty.

I realize Bostrom isn't proposing this as a likely scenario. That's not the point. What matters here is that he and other researchers and commentators tend to think about technology using the specific heuristics and motifs people have always used for thinking about magic, and it worries me when I start recognizing the Aarne–Thompson classifications for stories in the science section.

 

AI doomsters and Dial F for Frankenstein

Regular readers know we've been discussing the deeply dysfunctional and just plain weird culture of the techno-optimists for going on 20 years now.

Probably the first thing you need to keep in mind when dealing with the movement is that the optimism here is about the speed and magnitude of upcoming technological changes, not necessarily about whether their impact will be wonderful or catastrophic as long as it's big and soon. The AI doomsters are an outgrowth of the movement, and their ideas still find their greatest audience there. The conflict between the abundance crowd and the doomsters is essentially a sectarian disagreement between people who believe that large language models are on the verge of achieving god-like powers, which they will use to usher in a golden age free of want, and people who believe that large language models are on the verge of achieving god-like powers, which they will use to exterminate humanity.

This leads us to arguably the second most important thing to remember about techno-optimists: this is not a movement based on deep thinkers or profound insights into the nature of technology. Instead, the primary intellectual foundation comes from postwar science fiction, mostly filtered through movies, TV shows, and comic books. (Other influences include a fascist-friendly form of libertarianism, the Ithuvania effect, and constant and shameless grifting.)

Techno-optimists tend to have a notably weak grasp of the thing they claim to be obsessed with. Just listen to Elon Musk going off script talking about space travel or Marc Andreessen explaining how he simply tells chatbots not to hallucinate. Conversations that pass themselves off as explorations of the future of technology largely consist of sci-fi tropes being passed around. Ideas such as bipedal humanoid robots are judged not based on the soundness of their engineering, but on how well they would fit in with old Galaxy magazine covers or Mystery in Space comics.

 

 

 

 

One of the most common of these tropes dates back at least to Dial F for Frankenstein (Playboy, January 1965), a decidedly minor Arthur C. Clarke work, a lead-up to a pun that is weak even given the low standards of the genre. (If you want lighthearted Clarke, stick with Tales from the White Hart.) It is, however, one of the earliest examples of a popular sci-fi trope which has blossomed into something close to a religious movement in the past few years.

The idea is that if you construct a sufficiently complex system, superintelligence will almost immediately and spontaneously emerge, along with other human traits such as vengefulness, the need to dominate and control, and, most importantly, an overpowering survival instinct.

 

It is a trope born of equal parts anthropomorphism and narrative necessity. That sort of thing is fine for literary purposes, but we now have rich and powerful people who actually believe this, and that's not fine at all.

Patrick Boyle on the anniversary of Brexit

I could pull pretty much any random passage from a Boyle video and strike gold, but this one in particular caught my attention.

The first story was told by the Leave campaign, who drove a large red bus around the UK with the words, "We send the UK £350 million a week. Let's fund our NHS instead."

The pitch was that Britain was hemorrhaging money to Brussels, money that was being wasted, and that leaving would let the country reclaim its borders, slash bureaucracy, revitalize the health service, and, most importantly, bring back the bendy banana, which the EU had supposedly outlawed.

This was promoted heavily by right-wing populists and, more surprisingly, by a smaller group of left-wing Euroskeptics who wanted to leave the EU not to deregulate the economy, but to regulate it far more aggressively than Brussels would ever allow.

The two groups agreed on the destination and on absolutely nothing else. 

 It's remarkable how often factions of the left ends up implicitly and often explicitly supporting the far right, particularly given how these stands almost never turn out well. 

Peter Thiel's moments of honesty are largely limited to him saying bitchy things about Elon Musk

 A little bit of background: back in 1999, I was working for a large financial services company (you've heard of them). One of the biggest topics on everyone's minds was how to crack the online market. Every major bank was devoting serious money and resources to the question.

This was also going on in the startup world. Two of the leading companies in this space were Confinity and X.com, which would merge about a year later into the company known as PayPal.

You often hear about Elon Musk being the co-founder of PayPal, but that somewhat misrepresents what actually happened. These were two different companies with very different approaches and proposed products, and they brought very different things to the table. Confinity was focused on the area of payments and had already launched the PayPal product and had identified eBay as being the most important early market. Elon Musk's X.com was a more general, less well-thought-out idea for an all-services bank, something that would not be practical online for years.

X.com had more money and resources, particularly teams of engineers, and was technically the acquirer rather than the acquiree, while the primary product and basic approach that made the combined company viable enough to sell to eBay came from Confinity.

For an example of the quality of ideas that Musk brought to the table, here's an anecdote from Peter Thiel:

I was thinking of doing a book on PayPal … and [the chapter on Elon Musk] was going to be titled, “The man who knew nothing about risk.” … We had decided to give a credit card to anybody who wanted them. You got up to $10,000 credit limit. Elon had told the woman who was rolling the service out that he wanted 1 million people to be using the new credit card by the end of the year. Fortunately, it was about two levels down from the front page, and so not that many people were able to discover this. Some people did. They wrote us back and said, “You know this is fantastic! I haven’t had credit in years. I can’t believe you’re offering me credit. I haven’t even had a checking account in 10 years!” … We ended up with something like a 50% charge back rate—the worst subprime companies are like 4%-6%. And then, happily, we sort of rolled that product back very quickly. 

The 2018 Quartz article I pulled this quote from goes to embarrassing lengths to depict Musk as a bold visionary rather than someone who is really bad at business (the earlier Quartz article it links to is even worse), but these are not isolated cases. If you follow the man closely, you will find countless examples of Musk failing to grasp basic and sometimes self-evident concepts about the bond market, manufacturing, manned spaceflight, you name it, all delivered with an air of self-confidence that can only be called delusional.

After PayPal, Musk's business career has consisted of running a moderately profitable niche car company and having founded or acquired a string of companies that either lost money or were only occasionally profitable. His unprecedented fortune in no way reflects the money his companies have made. Instead, it is the product of his ability to convince people to give him money, be they investors, stockholders, or taxpayers.

It is a disconnect that sets Musk apart from the robber barons of the Gilded Age and every other multibillionaire (in 2026 dollars) I can think of before the year 2000.

“It really seems like anyone with some renders and a white paper written by someone being gassed up by an overly agreeable AI can get VC funding these days,”

Futurism points us to what happens happens when tech bro "engineers" run into actual engineers:  

A scathing takedown video — a collaboration between Irish aeronautical engineer Brian McManus from the YouTube channel Real Engineering and tech publication IEEE Spectrum — discusses the many pain points of Starcloud, a Y combinator company that raised $170 million earlier this year to develop data center spacecraft with the help of SpaceX, and how the white paper outlining the project feels far more like it was dreamed up by an AI than a team of human experts.

“It really seems like anyone with some renders and a white paper written by someone being gassed up by an overly agreeable AI can get VC funding these days,” McManus argued. “Billionaires will attempt to pull the rug over your eyes and convince you that this technology makes total sense, but reality is, this technology is dumb.”

 

My favorite part occurs around the 7-minute, 40-second mark when the narrator points out that, while it makes sense to have solar panels perpendicular to the sun, having the radiators catch as much sunlight as possible is the opposite of what you want.

It's a small point, but it really brings home just how little thought you have to put into something before you get the Y Combinator stamp of approval and almost $200 million in funding.

 

For a deeper IEEE dive.