I Will Never Respect A Website
We recorded 18 claims from this post. 18 can’t be checked. 7 more are held. Read the post at its source.
Claims from this post (18)
In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.
Assessment: Can’t be checkedSaid
Observation
“In any case, no previous real or imagined tech revolution has ever inspired such eager defensiveness, tribalism or outright aggression toward dissenters, nor such ridiculous attempts to obfuscate the truth about a product outside of cryptocurrency, an industry with obvious corruption and financial incentives.”
Our reading: Apart from crypto, no previous tech revolution inspired such defensiveness, tribalism and obfuscation as AI.
No horizon
Qualifier: except cryptocurrency
Assessment: Can’t be checkedSaid
Rhetoric
“LLMs are excellent digital busyboxes for those who want to come up with a way to work differently rather than actually doing work.”
Our reading: LLMs are excellent digital busyboxes for people who want to come up with a way to work differently rather than actually doing work.
No horizon
Scope: for those who want to come up with a way to work differently rather than actually doing work
Assessment: Can’t be checkedSaid
Observation
“They do not think, they do not have knowledge, they are conjuring up their own training data (or reflecting post-training instructions from those developing them or documents instructing them to act a certain way), and any time you try and make them do something more-complicated, they begin to fall apart, and/or become exponentially more-expensive.”
Our reading: LLMs do not think or have knowledge, and they fall apart or become exponentially more expensive on more complicated tasks.
No horizon
Qualifier: on more complicated tasks
Assessment: Can’t be checkedSaid
Observation
“LLMs are not “artificial intelligence” at all.”
Our reading: LLMs are not artificial intelligence.
No horizon
Assessment: Can’t be checkedSaid
Observation
“Large Language Models perform an impression of automating that process, but for the most part force you, the user, to do the shit that matters, even if that means “be responsible for the code that it puts out.””
Our reading: LLMs perform an impression of automating the process of building software, but for the most part force the user to do the things that matter, even if that only means being responsible for the code they put out.
No horizon
Qualifier: for the most part
Assessment: Can’t be checkedSaid
Observation
“Anyone you read talking about how LLMs “allowed them to look through a massive dataset” is likely using Python.”
Our reading: People who credit LLMs with letting them analyze massive datasets are likely really using Python.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“They should have to show you how much work it took to prepare the system, and why it’s superior to just doing it themselves.”
Our reading: People claiming others are being left behind by AI should have to show the setup effort required and why AI beats doing it themselves.
No horizon
Assessment: Can’t be checkedSaid
Causal claim
“I just think that they would be used dramatically less if there weren’t an industrial-scale publicity campaign run through the media and the majority of corporate America both incentivizing and forcing them to do so.”
Our reading: Developer use of LLMs would be dramatically lower without the media publicity campaign and corporate pressure.
No horizon
Qualifier: absent media and corporate pressure
Assessment: Can’t be checkedSaid
Causal claim
“Similarly, I’m not sure anybody would’ve been anywhere near as excited if OpenAI and Anthropic hadn’t intentionally sold them a product that was impossible to support long-term.”
Our reading: The author is not sure anybody would have been anywhere near as excited about LLMs if OpenAI and Anthropic had not intentionally sold them a product that was impossible to support long-term.
No horizon
Qualifier: I'm not sure
Assessment: Can’t be checkedSaid
Observation
“This also suggests that Anthropic’s (and OpenAI’s by extension) business model is fundamentally flawed.”
Our reading: Anthropic's inability to serve its paying customers at its current infrastructure scale, while still burning money, suggests that Anthropic's business model, and by extension OpenAI's, is fundamentally flawed.
No horizon
Qualifier: suggests
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“In fact, the ethical thing to do would’ve been to charge for and restrict the services in line with their actual costs so that users could have reliable and consistent access to the services in question.”
Our reading: Anthropic and OpenAI should have priced and restricted services in line with actual costs.
No horizon
Assessment: Can’t be checkedSaid
Observation
“Right now, I’d guess that OpenAI has access to around 2GW of capacity (as of the end of 2025), and Anthropic around 1GW based on discussions with sources.”
Our reading: The author guesses OpenAI had about 2GW of capacity at end of 2025 and Anthropic about 1GW.
No horizon
Qualifier: author's guess based on sources
Assessment: Can’t be checkedSaid
Forecast
“I imagine that at some point enterprise customers will be able to pay for an even higher priority tier, and that Anthropic’s “Teams” subscription (which allows you to use the same subsidized subscriptions as everyone else) will be killed off, forcing anyone in an organization paying for Claude Code (and eventually Codex) via the API, as has already happened for Anthropic’s enterprise users.”
Our reading: The author imagines that at some point enterprise customers will be able to pay for an even higher priority tier, and that Anthropic's Teams subscription will be killed off, forcing organizations paying for Claude Code (and eventually Codex) onto the API, as already happened for Anthropic's enterprise users.
No horizon
Qualifier: I imagine
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Both he and Mr. Amodei need to immediately stop overstating the capabilities of Large Language Models.”
Our reading: Altman and Amodei must immediately stop overstating LLM capabilities.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Mr. Amodei should immediately cease any and all discussions of AI potentially or otherwise eliminating 50% of white collar jobs, as Mr. Altman should cease predicting when Superintelligence might arrive, as Mr. Amodei should actively reject and denounce any suggestions of AI “creating a white collar bloodbath.””
Our reading: Amodei should stop discussing AI eliminating 50% of white-collar jobs and Altman should stop predicting superintelligence timelines.
No horizon
Assessment: Can’t be checkedSaid
Observation
“Both Altman and Amodei engage in the language of AI doomerism as a means of generating attention, revenue and investment capital, actively selling their software and future investment potential based on their ownership of a technology that they say (disingenuously) is potentially going to take everybody’s jobs.”
Our reading: Altman and Amodei both use the language of AI doomerism to generate attention, revenue and investment capital, selling their software and future investment potential on a technology they disingenuously say may take everybody's jobs.
No horizon
Assessment: Can’t be checkedSaid
Causal claim
“These are acts of violence that could have been stopped had ChatGPT and similar applications not been anthropomorphized by design, and trained to be “friendly.””
Our reading: The acts of violence by people twisted by LLMs into taking their own and others' lives could have been stopped had ChatGPT and similar applications not been anthropomorphized by design and trained to be 'friendly'.
No horizon
Qualifier: had ChatGPT and similar applications not been anthropomorphized by design and trained to be 'friendly'
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Stop talking about AGI. Start talking about this like regular old software, because that’s all that ChatGPT is.”
Our reading: AI executives should stop talking about AGI and describe ChatGPT as the regular software it is.
No horizon
Held from this post (7)
These records did not pass every check, so nothing about them is published. Held is not a finding.