| > Otherwise, the report is meaningless nonsense - comparing apples with oranges, making claims about possible future dangers that aren't supported by anything the report actually found, and so on. Some responses: - Do you think the report is literally "meaningless nonsense" -- meaning it is incomprehensible or self-contradictory? I don't think you mean this. - Do you disagree with the report's technical findings? I am pretty confident (P > 70%) you haven't engaged with them well enough to make specific claims about the technical aspects. - Do you think the report's technical findings are so biased as to be (more or less) worthless in addressing the question of risk from DeepSeek? Yes; this seems to be your claim. As to the last point, you haven't persuaded me. Why? You haven't engaged substantively with the NIST Report; you've mostly made sweeping comments with many reasoning errors. Here's a guess at what may be happening in your brain. You let your view about Trump "run wild"; you probably haven't given any significant thought to the technical or geopolitical points on their own merits. Instead, you've fixated on the view that the Trump administration has ruined the objectivity of the report. In short, you found your preferred explanation ("motivated reasoning") and then stopped looking for other explanations ("early stopping"). These are common -- we're only human after all -- but damaging cognitive errors. I have some other guesses... You probably lack: (i) an understanding (of the topic area or of how NIST works); or (ii) the curiosity or time to dig in. A lack of understanding is not necessarily a problem if you recognize it and adjust accordingly (i.e. by expressing uncertainty and/or expanding your knowledge). [2] From my POV, I'm not confident you understand the key concepts from the NIST report. May I ask: what is your experience level with: national security, cybersecurity, machine learning, U.S. government, risk assessment, prediction, economics, geopolitics, or similar? What about the particular technical AI topics mentioned in the report? - Many do not have experience in these areas. This is Hacker News, not e.g. an invite-only message board for AI experts interested in government policy. I don't know what a random HN commenter knows, but I would predict it isn't anywhere close to "competent" in all of the above. - Knowledge across these areas is helpful (probably necessary in my opinion) to understand the NIST Report well. Without that background, one will have huge gaps. And unless you are really careful, your brain will fill those gaps with processes riddled with cognitive bias. [1] - Beware the hubris that might lead someone to claim the lack of such experience is irrelevant. (And yes, experts are not immune from cognitive bias either.) [1]: To borrow some words from Claude Sonnet 4.5, which I endorse as matching what I've learned from other sources: "Examine what appears to be rational thought and you find it rests on heuristics; examine those heuristics and find more heuristics beneath. There's no rational bedrock—it's cognitive biases all the way down." [2]: For many, another frustration (such as Trump's degradation of democracy) can be a powerful and extensive demotivator in other areas. That frustration can serve as a explanation for much that ails us. This can become a coping mechanism, which serve a function at times, but are rarely motivators to increase the curiosity needed to make sense of a messy world. |