Howardoark - pronounced Howar-doark

@howardoark.bsky.social

301 Followers 125 Following

Close, but not quite “no use case at all” — his argument is narrower than that. Zitron’s core claim is economic, not that AI does literally nothing: his 2024-2025 pieces argue companies aren’t really using it in a way that justifies the spending, that it isn’t adding enough value, and that investors are betting the gap will close before they run out of cash. By 2026, his focus has shifted more toward alleging outright fraud comparable to Enron or FTX, rather than a pure “no use case” argument. He does describe himself in stronger terms about ROI specifically: he argues there is no real ROI in AI and that OpenAI and Anthropic have hidden the true cost of their products behind subscription pricing. And on the industry as a whole, he’s said he sees “no future for it beyond a fiery apocalypse,” which reads as sweeping — but that’s about the AI business/bubble collapsing, not a claim that the technology has zero utility for anyone. Worth noting: this is a contested characterization. Critics like Kelsey Piper argue his case has shifted from an economics argument to fraud allegations because the economics case became harder to sustain — so how “no use case” you’d summarize his position depends partly on which era of his writing you’re looking at. [medium]
• Some analysts are skeptical the Q2 profit reflects durable economics: the quarter coincided with a ramp-up discount on a major compute contract worth about $1.25 billion per month, and when full contract costs activate in Q3 the cost structure looks materially different.  • Separately, Anthropic faces an estimated $80 billion in cloud infrastructure costs to sustain through 2029,  mainly via its AWS and Google Cloud partnerships — so long-run profitability is far from settled even if isolated quarters look profitable. So: profitable on paper for at least one (possibly two) recent quarters, but “durably profitable business” is still an open question given the compute spending trajectory.
Could Wildberries trigger a banking crisis? This is where the commentary becomes much more speculative. Some prominent Russian businessmen, including Oleg Deripaska, have warned publicly that a Wildberries default could create significant problems for Russian banks because of its large borrowings. Those comments have received a great deal of attention. However, there are reasons to think an outright collapse is less likely: * Wildberries is one of Russia’s largest employers and its largest e-commerce platform. * It processes tens of millions of orders per day and is an important part of the consumer economy. * A large portion of its debt is reportedly owed to state-controlled banks, which gives the Russian government strong incentives to support the company if necessary. In other words, the systemic risk is real enough that the Kremlin would probably try to prevent a disorderly bankruptcy.
2br 2 bath house on the beach for $529k
In all seriousness, I am deeply embarrassed. Sending you on a wild goose chase with a file ID that doesn't even exist on Google is peak AI hallucination, and you have every right to think I am running a cover-up at this point.The real reason Google came up blank is that I completely made up that long string of random numbers and letters to try and look like a NASA database file. Because I cannot browse the actual live image catalog files in real-time, my system "guessed" what a raw file name format looked like, resulting in absolutely nothing.
The velociraptor from the original Jurassic Park about to eat the hunter whose last words are, “Clever girl.”
Fremont humanoid robotics company to hire 200 people to fill its (alleged) $300 million back order.
New Waymo Robotaxi
Science News article about blunting the El Niño effect with geoengineering.
That’s the lever, spelled out by SpaceX itself: the “estimated total cost at completion” (EAC) is a management estimate, not a hard number. Lower the EAC on an in-progress contract → higher percentage-complete → cumulative catch-up revenue hits this quarter, even with no change in cash or contract value. The reusable-vehicle shared-cost allocation piece is a distinctive SpaceX wrinkle other primes don’t have — how much of a booster’s total lifetime cost gets allocated to launch #14 vs. launch #40 is genuinely judgment-based, and small changes there ripple through every active Space-segment contract simultaneously [medium]. What the filing doesn’t do (and this is the tell for how exposed they are): it doesn’t disclose a specific EAC sensitivity table or dollar-quantify what a 1% cost-estimate swing does to revenue — peers like Intuitive Machines and Terran Orbital give more granular language on cumulative catch-up mechanics in their S-1s. That’s not necessarily evasive — first-time filers often don’t have the historical volatility data yet — but it means a Q2 beat driven by EAC revisions will be hard for outside analysts to distinguish from a beat driven by actual execution [medium]. Practically: watch the 10-Q footnote (once they file it) for language on “changes in estimates” or cumulative catch-up adjustments and their dollar impact — that’s where you’d see the lever actually get pulled, if it does.