Alignment Is Capability
by drctnlly_crrct on 12/8/2025, 1:23:29 PM
https://www.off-policy.com/alignment-is-capability/
Comments
by: xnorswap
I've only been using it a couple of weeks, but in my opinion, Opus 4.5 is the biggest jump in tech we've seen since ChatGPT 3.5.<p>The difference between juggling Sonnet 4.5 / Haiku 4.5 and just using Opus 4.5 for everything is night & day.<p>Unlike Sonnet 4.5 which merely had promise at being able to go off and complete complex tasks, Opus 4.5 seems genuinely capable of doing so.<p>Sonnet needed hand-holding and correction at almost every step. Opus just needs correction and steering at an early stage, and sometimes will push back and correct my understanding of what's happening.<p>It's astonished me with it's capability to produce easy to read PDFs via Typst, and has produced large documents outlining how to approach very tricky tech migration tasks.<p>Sonnet would get there eventually, but not without a few rounds of dealing with compilation errors or hallucinated data. Opus seems to like to do "And let me just check my assumptions" searches which makes all the difference.
12/8/2025, 2:28:42 PM
by: delichon
> Miss those, and you're not maximally useful. And if it's not maximally useful, it's by definition not AGI.<p>I know hundreds of natural general intelligences who are not maximally useful, and dozens who are not at all useful. What justifies changing the definition of general intelligence for artificial ones?
12/8/2025, 2:26:13 PM
by: munchler
> A model that aces benchmarks but doesn't understand human intent is just less capable. Virtually every task we give an LLM is steeped in human values, culture, and assumptions. Miss those, and you're not maximally useful. And if it's not maximally useful, it's by definition not AGI.<p>This ignores the risk of an unaligned model. Such a model is perhaps less useful to humans, but could still be extremely capable. Imagine an alien super-intelligence that doesn’t care about human preferences.
12/8/2025, 2:31:59 PM
by: js8
I am not sure if this is what the article is saying, but the paperclip maximizer examples always struck me as extremely dumb (lacking intelligence), when even a child can understand that if I ask them to make paperclips they shouldn't go around and kill people.<p>I think superintelligence will turn out not to be a singularity, but as something with diminishing returns. They will be cool returns, just like a Brittanica set is nice to have at home, but strictly speaking, not required to your well-being.
12/8/2025, 2:21:16 PM
by: riskable
The service that AI chatbots provide is 100% about being as user-friendly and useful as possible. Turns out that MBA thinking doesn't "align" with that.<p>If your goal is to make a product as human as possible, don't put psychopaths in charge.<p><a href="https://www.forbes.com/sites/jackmccullough/2019/12/09/the-psychopathic-ceo/" rel="nofollow">https://www.forbes.com/sites/jackmccullough/2019/12/09/the-p...</a>
12/8/2025, 2:44:34 PM
by: podgorniy
Great deep analysis and writing. Thanks for sharing.
12/8/2025, 2:02:16 PM