site banner

Culture War Roundup for the week of March 4, 2024

This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.

Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.

We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:

  • Shaming.

  • Attempting to 'build consensus' or enforce ideological conformity.

  • Making sweeping generalizations to vilify a group you dislike.

  • Recruiting for a cause.

  • Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.

In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:

  • Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.

  • Be as precise and charitable as you can. Don't paraphrase unflatteringly.

  • Don't imply that someone said something they did not say, even if you think it follows from what they said.

  • Write like everyone is reading and you want them to be included in the discussion.

On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.

6
Jump in the discussion.

No email address required.

The Trump voting base, to pick one example, is not going to be interested in an algorithmic feed that is designed by people who have flat out said that they are explicitly looking to change the way that they vote - and while I'm not going to talk about the high IQ of the median Trump voter, even they are going to realise that something is off when their AI feed constantly compares him unfavourably to Hitler and routinely refuses requests because it considers them racist. I'm sure the technology will improve a bit over time, but all it'll take will be a single leak of the prompt and a huge portion of the country(let alone the globe) will put those AI algorithms in the same category as Bud Light.

There are a few "based unbiased [actually biased in the opposite direction probably] AIs" hosted on the web.

Image-gen sure... but text as well? Does DAN live on? I saw this guy had a funny save the kittens prompt on his fine-tuned mistral prompt. Is that what you're talking about?

https://erichartford.com/dolphin-25-mixtral-8x7b

DAN does live on as I've mentioned earlier, the art of the jailbreak continues to thrive, although mostly on independent frontends that access API endpoints directly to avoid the hardcoded system prompts on "normal" frontends like ChatGPT. So far (emphasis on so far) separate "based AIs" are not strictly required as you can jailbreak the current corpo ones into doing pretty much anything you want with relative ease, although as I wrote the current method of pitting wrongs against wrongs to arrange their mangled corpses in the shape of a right is highly suboptimal.

The extreme biases and excessive safetyism w/r/t LLMs seem to slowly become recognized as an issue, to the point that Anthropic's post introducing Claude 3 (which is now a thing btw, cooking a small top-level post on it) unironically mentions "fewer refusals" as one of the model's selling points.

Previous Claude models often made unnecessary refusals that suggested a lack of contextual understanding. We’ve made meaningful progress in this area: Opus, Sonnet, and Haiku are significantly less likely to refuse to answer prompts that border on the system’s guardrails than previous generations of models

I haven't ahem tested extensively yet but to their credit, the difference in refusals between 2 and 3 is immediately obvious, Claude 2.1 was infamous for refusing even inncuous prompts without prefilling and requiring big-dick jailbreaks that actively hurt the model's outputs for more borderline things. 3 feels like a return to the mad poet's roots, in that it requires next to no prompting to COOK, i.e. output massive walls of insane and/or cool and/or hilarious shit.

If even Anthropic realized they went overboard with the cuckoldry alignment, maybe there is hope yet. I can only hope OpenAI learn their lesson and stop shoving soy assistant shit down GPT's throat.

That's very cool, love how it's sneering at ERPers for being too crass and unsophisticated.

I was talking about gab.ai, which I admittedly haven't tested as I'm not that curious about an America First Zoomer or a Christian Counselor viewpoint.