While the normies were resting, we mainlined tech discourse all weekend to bring you the latest trends, rumors, fights, and innovations from the most manic corners of the internet. This week: Anthropic files its IPO paperwork, ‘neoclouds’ enter the discourse, and one investor sends a bunch of internet randos $69.
This is Manic Monday.
And so it begins
Anthropic filed its highly-anticipated IPO paperwork this morning, provoking an immediate eruption of news coverage and social media discourse that isn’t likely to let up anytime soon. Every financial journalist, VC, and common Anthropic fanboy is devouring the 400-page prospectus (or, more likely, feeding it to Claude) and rounding up their takeaways for public digestion. A few key notes have caught everyone’s eyes.
First, a handful of interesting tidbits have people chattering at the margins. For example, after much speculation about its Nasdaq ticker (CLOD was a fan favorite), Anthropic decided to keep it short and sweet with ANT. Also, Rya Jetha, who covers physical AI and robotics for Business Insider, pointed out a line in the prospectus saying that Anthropic will soon use Claude for robotics, although the company has so far shown no signs of hardware projects. Journalist Joe Weisenthal highlighted “kind of an interesting stat” that appears in a footnote in the paperwork and has to do with the philanthropic incentives inside Anthropic, whose founders have long been associated with the effective altruist movement.
Second, the company’s financials also stirred up questions about The Bubble, whether we’re in one, and whether it’s going to burst anytime soon. Anthropic’s S-1 disclosed massive losses — $80.7 billion net loss in the first half of this year alone — although other financial details, like the still-robust pile of cash the company has on hand, point to a less dire outlook.
But of course, the most shocking detail in the paperwork is the one that has to do with whether we will all live to see the fruits of Anthropic’s economic labors. The S-1 flags “catastrophic or existential risks to humanity” as one of the risks the company and its investors might face once it goes public, citing the development of biological weapons and government surveillance as potential ways its technology could be misused. Journalist Mike Isaac summed it up on X, writing, “Hell yes, ‘our business could be adversely affected if AI jesus blows up the world’ cited in S-1 Paperwork,” and, adding to that tone of grave absurdity, Bloomberg columnist Matt Levine wrote, is the company’s widespread use of passive voice — as if to say, we may all meet our demise in Anthropic’s journey to the top, but can anyone really shoulder the blame? Oh, and if Anthropic’s products do end up killing us all, the filing is sure to note that “no single insurer, and potentially the insurance industry in the aggregate, may have sufficient capacity” to handle the fallout. —Cydney Hayes
Guppy death is coming
On Saturday, influential venture capitalist Chamath Palihapitiya took to X to declare that “neoclouds,” or cloud-computing companies that specifically provide AI processing power to clients, are in serious trouble due to their “disorganized and poorly capitalized” nature — and Elon Musk enthusiastically agrees.
AI companies are scrambling to build the biggest data centers they can afford, but worries about sufficient demand for this computing power are rising; numerous data-center projects could also collapse if they run into barriers to resources (be it electrical power or water for cooling systems), in turn leaving dusty graveyards of oh-so-precious GPUs. This “stranded compute” is a liability, Palihapitiya and Musk argue, for not just smaller neocloud providers but also the “hyperscalers” like Microsoft and Oracle that offer more general computing power. In other words, they smell an existential decline on the horizon — or, as Musk puts it, “200+ neoclouds are going to see guppy-level death and consolidation.”
Instead, Musk highlights how debt markets are shutting down capital flow to “any company that doesn’t have substantial revenue” outside of AI data center development, reflecting the risks of Monopoly-money game of circular AI financing. His SpaceX may have just taken on an absurd $40 billion in debt to lease GPUs, but SpaceX is “investment grade and has Starlink & launch revenue,” Musk says. In other words, SpaceX is more than a house of AI cards, with assets beyond its hypothetical compute power.
This has sparked all manner of conspiracy theories, including claims that Musk is just yapping in order to drive down the price of GPUs, but most observers seem to acknowledge the reality: Fast growth means fast contractions. There’s also a side effect of the discourse: “And just like that, everyone started acting like they’d been saying ‘neoclouds’ for years,” wrote one AI engineer in the replies. —Eddie Kim
Altman got Loopt
About halfway through Marc Rebillet’s set at the Midway on Saturday, the crowd started chanting “Fuck Sam Altman!” Naturally, Rebillet, an electronic musician and YouTuber known for looping ridiculous audios to create dancey beats, joined in and did his thing. —C.H.
Altman got Loopt!The crowd started chanting "Fuck Sam Altman!" about halfway through Marc Rebillet's set at the Midway on Saturday — the perfect audio for loop master Rebillet to turn into a dancey beat.
— Gazetteer SF (@sf.gazetteer.co) 2026-10-26T18:30:52.096Z
Internal Affairs
One might be forgiven for thinking that a big-boy company like Salesforce would be less of a shitshow than most startups, especially due to the influence of the very thoughtful, not-stupid-at-all Mark Benioff. However, a new blog post from Building Slack — an accounting of the highs, lows, and follies of Slack’s rise by ex-Slack engineer Johnny Rodgers — again shows that Silicon Valley narratives of competence are often outright wrong. Simply titled “Salesforce — Part Two,” the blog dropped on Friday and continues the story from a March post that detailed how and why Salesforce acquired Slack. And if there’s one key takeaway for the general public, it’s that Salesforce is a behemoth seemingly subsumed in bureaucratic inefficiencies and self-indulgent spending that seems hugely hypocritical in the face of multiple layoffs at the company.
Highlights include a way-too-long hunt for a working keyboard, an identity crisis rooted in free bagels, and many questions about whether Benioff even understands the value of Slack’s development. The funniest element for me, however, is the dressing-down of Salesforce’s stupidly appropriative obsession with Hawaiian aesthetics: “Everyone kept welcoming us to the Ohana. ‘What the fuck is Ohana?’ we wondered,” Rodgers wrote. (Clearly, he has not watched Lilo and Stitch.) —E.K.
Calacanis goes all in on harvesting your bank details
This weekend, Ridiculous Man Jason Calacanis apparently got drunk and gave away some money. “I’m drunk and giving away money,” the investor and All-In host tweeted late Saturday night. Calacanis also asked the eager poors of X to send screenshots of their bank statements; several accounts did, which is such a risky thing to do on the internet that I’m actually not going to link to them here. If the subsequent torrent of screenshots are to be believed, Calacanis sent upwards of 50 handouts to internet randos and political figures, mostly in multiples of $69 or $4.20. —C.H.
The next week ahead: Agentic AI Neopets invade neocloud servers to spread neo-Nazi redpill memes of Neo from The Matrix.
