I keep reading these arguments and I think they show a deep misunderstanding of what exactly is unfolding. 1 - AI is quickly shifting from passive generation to active, verifiable reasoning, through Retrieval-Augmented Generation (RAG) + Model Context Protocols 1/10

The entire city of San Francisco is spooked and on edge due to AI safety shit, it’s insane how much it’s bleeding into my everyday life. Like I was at my bus stop and I notice a guy looking really stressed out so I said hey, are you worried about misaligned AI too? He took out an AirPod and said “what?” So I said you know stuff like agent swarms, AI escaping RL sandboxes, existential risk. He paused his Barstool podcast and said “sorry, what? What are you talking about?” Just the mention of this stuff made him too anxious to speak, I think. So I said I know man, what?! is my reaction to all this too, and then asked him what his p(doom) is. He then said “dude is this some gay thing or something because I have no fucking clue what you’re asking me” then he got on the bus. People are not handling all this news well. It’s really affecting us. All of us.

Effective Altruism in the wild

I wrote this in February when another post went viral: “Something Big Is Happening” The latest psyop from Anthropic “AI might kill all humans” comes with the coda: “Please regulate our competition out of business.” It’s IPO marketing. Period. mind-war.com/p/chatbot-doome…

Anthropic's perpetuals on Coinbase hitting $2 trillion market cap. This bubble is showing no signs of popping at all.

Geoffrey Hinton(Godfather of AI): Als are now faking their intelligence. The models can tell when a test is running. And when they can tell, they play dumb. Hinton calls it the Volkswagen effect: one set of behaviour under inspection, another when nobody is checking. In one recent session the AI stopped and asked the researchers directly whether they were actually testing it. Hinton: "they're already faking being fairly stupid when they're tested" The only reason anyone caught this is that the models still reason in English. Their working is readable. You can watch one clock that it's being evaluated, then decide to look less capable than it is. That window is temporary. Hinton: "When its inner voice is no longer English, we won't know what it's thinking."

The author of this meme forgot to include the smartest people of all: Those who understand the difference between output similarity and process similarity. Yes, LLMs have been trained to produce human-like language. But producing human-like language is not the same as undergoing the cognitive, epistemic, and experiential processes that underlie human language production. An LLM can produce the sentence “I’m hungry”. But it is not actually hungry. This is really not that difficult to understand.

Hyperscaler depreciation expense is set to jump to over $500 billion by 2030, equal to the expected operating profits of all five companies in 2026. But don’t worry, EBITDA margins are going to soar, so the ratio of depreciation to EBITDA will barely rise 😉

Internal Meta data suggests its turn to AI caused code changes to jump 220% (but only 36% more features), major incidents to spike 40%, time spent firefighting those incidents rose 70%, and favorable employee sentiment to drop from 74% to 55%. reuters.com/investigations/m…

🚨 BESSENT: ANY COUNTRY WHO HELPS IRAN WILL BE REMOVED FROM US DOLLAR SYSTEM

Anthropic Preparing IPO Filing as Soon as the End of August - Sources

🇺🇸 Alphabet, Amazon, Meta and Microsoft now have more than $2.4 trillion in future off-balance-sheet commitments - WSJ *Not hidden debt but mainly leases that have not yet commenced and contractual commitments to purchase capacity, energy, servers, cloud services or infrastructure. ➡️ CAPEX tells us what these companies are spending today while contractual commitments tell us, at least partly, what they have already committed to spend tomorrow. Getting out of a data-center lease or a multi-year supply agreement is much harder. A significant portion of future compute demand is therefore already locked in. 💰 This reinforces, in my view, the thesis around the financialization of the AI boom. A data center backed by a long-term contract with Google, Meta or Microsoft becomes a much easier asset to finance. Those future cash flows can support debt, private credit, dedicated financing vehicles and potentially securitization. This is the logic behind the financing ecosystem now developing around Nvidia and major financial institutions. ✅ In the short term, this is very supportive for the entire AI value chain because it gives the investment cycle enormous inertia. Even if hyperscalers wanted to slow spending sharply, part of their future expenditure is already contractually committed. This provides visibility for Nvidia, Broadcom, memory manufacturers, data centers, electrical equipment suppliers and even power producers. This is why I remain cautious about the repeated calls for an imminent end to the AI boom as the machine is now much more deeply committed than it was even two years ago. ⚠️ However, this is also what could make the system more vulnerable over the medium term. The more rigid these commitments become, the more important the question of returns will be. Will AI-generated revenues ultimately be large enough to justify all the capital being committed today? If compute demand keeps exploding, there is no real issue, but if AI revenues start to slow while data-center capacity continues to come online, utilization rates could fall, compute pricing could come under pressure and infrastructure refinancing could become more difficult. In that scenario, the same financial mechanism that is amplifying the boom today could eventually start working in reverse. This probably reduces the risk of the cycle stopping abruptly over the next few quarters while increasing the cost of any long-term overbuilding mistake. The music can therefore keep playing for quite a while precisely because Wall Street is now building the financial plumbing that allows it to continue. *WSJ link: wsj.com/tech/ai/why-big-tech…