Hacker Newsnew | past | comments | ask | show | jobs | submit | chermi's commentslogin

What do you want? Some oversight board voting on what companies are allowed to do with their own money?

Wtf are the downvotes. It's a question.


Your perspective made sense until 2007.

We don't live in that world anymore. The bailouts changed the world forever. Since 2007 there is no such things as "with their own money" when it involves a big company risking bankruptcy. It's no longer private money.

Chancellor on the brink


I didn't want the bailouts.

And yet you got them.

Do you really want someone qualified for politics in that sense?


Last I checked labor was 15-30% for such projects, so it could help but only so much


Again, I asked ChatGPT for some sources for your post, but I couldn't find any reliable sources. Can you provide a reliable source for your claim?


I'm 99% they use nvidia at least for training


See nuclear topic, scroll just like 2 turns to see once again someone claiming nuclear is the one technology in all of history that doesn't improve with technology.


Damn what's happened since this? Presumably they scramble refusal intentionally somehow now? Like intentionally couple it to "directions" that effect performance if messed with? Or is it more like just don't rely on the model to refuse and instead capture bad responses between generation and delivery?


I've seen attempts to obfuscate the refusal direction, like here: https://arxiv.org/html/2608.18093v1

Also this one was interesting, training the model to give preambles with reasons for the reasons for refusal seems to make it less sensitive to modulating the single refusal direction: https://arxiv.org/html/2505.19056v1

My empirical observation is that when a new model is released on HuggingFace, an abliterated version with < 10/100 refusals (baseline usually 100/100) is uploaded the same day, so either these techniques don't work very well or the open-weight labs aren't applying them.

There's some defense-in-depth, like a lot of the "guardrails" people hit on cloud models are classifiers applied to prompt or output, not a refusal generated by the model. Also closed-weight models obviously try to avoid this by not letting you see or modify the weights.


I'm not active in this space but why do you think anything happened since this? As far as I know, it still works.


Why would I want AI to write diary entries for me? Maybe the other stuff is valuable, but this seems like an inversion. I would prefer it help me write an entry by giving me a useful view and context with extremely low friction entry. Maybe I'd want nudges and reminders or reframings based on my entries, pending privacy concerns, but I definitely don't want a diary not written by me. The writing itself is the value, the act of thinking, preserving your feelings and thoughts honestly.



I think you stopped at the wrong time with the wrong perspective. Why can't that info accumulation part also be made more self-contained?

I guess I'm having trouble unraveling your experience and personal usage vs. what you're concluding about the labs.


Almost but not quite I think. You can throw money at parts of problems. I think it's helpful to think it kind of like supercomputer MD/MC or electronic structure calculations. A tool that can get you valuable answers but not necessarily aid understanding. Simulations can be used to aid understanding also, and are integral to theory development. In the same way the approach to this result is.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: