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.
Jump in the discussion.
No email address required.
Notes -
As someone who is not nearly as impressed with AI as you, thank you for the Turing test link. I've personally been convinced that LLMs were very far away from passing it, but I realize I misunderstood the nature of the test. It depends way too heavily on the motivation level of the participants. That level of "undergrad small-talk chat" requires only slightly more than Markov-chain level aptitude. In terms of being a satisfying final showdown of human vs AI intelligence, DeepBlue or AlphaGo that was not.
I still hold that we're very far away from AI being able to pass a motivated Turing test. For example, if you offered me and another participant a million dollars to win one, I'm confident the AI would lose every time. But then, I would not be pulling any punches in terms of trying to hit guardrails, adversarial inputs, long-context weaknesses etc. I'm not sure how much that matters, since I'm not sure whether Turing originally wanted the test to be that hard. I can easily imagine a future where AI has Culture-level intelligence yet could still not pass that test, simply because it's too smart to fully pass for a human.
As for the rest of your post, I'm still not convinced. The problem is that the model is "demonstrating intelligence" in areas where you're not qualified to evaluate it, and thus very subject to bullshitting, which models are very competent at. I suspect the Turing test wins might even slowly reverse over time as people become more exposed to LLMs. In the same way that 90s CGI now sticks out like a sore thumb, I'll bet that current day LLM output is going to be glaring in the future. Which makes it quite risky to publish LLM text as your own now, even if you think it totally passes to your eyes. I personally make sure to avoid it, even when I use LLMs privately.
Well remember even passing the basic casual Turing test used to be extremely difficult. It took at least 65 years between the creation of the test and systems beginning to pass it consistently. And I still remember science articles and science fiction stories from the 90s and 2000s talking about it like it was the holy grail. It’s only in the past few years that it’s started to seem like an inadequate measurement of an AI’s capabilities.
Interestingly your motivated Turing test starts to sound a lot like the Voight-Kampff test from Bladerunner.
Is there any reason the test was treated as a holy grail other than the "Turing" name brand? I can't see any theoretical justification for it.
The theoretical justification for it is something analogous to the idea of a Universal Turing Machine, though obviously not rigorous.
If we come up with any other test to determine "human-level intelligence", a test that can't be beaten by a "spiky" non-general intelligence that outperforms in unexpected areas (I'm old enough to remember when chess performance was a generally-accepted sign of intelligence!), then someone judging a Turing test can just use that other test. If it turns out that for some reason an AI really can't understand how to respond to a weird hypothetical about upside-down tortoises, then the judge can ask them about upside-down tortoises. If computers had sucked at chess, a judge could have asked the AI to play chess. Computers only start to beat a Turing test reliably when there's nothing a judge can come up with that they can't beat.
More options
Context Copy link
This is the way I always understood it. Lacking the ability to detect any internal experience other than our own, the way we distinguish between 2 different things is by applying input to them and seeing if there's differences in output, e.g. we shine light on it and detect what qualia the light that reflects off of it and into our eyeballs generate in our minds. Detecting intelligence isn't as simple as detecting the color or shape of something and wouldn't involve inputting light rays but rather words to see what words get returned in response. If there's no way to distinguish between 2 different entities in this way, then it makes no sense to say that 1 has human-level intelligence while the other lacks it. For that to be the case, there must be some way to induce different outputs from those 2 things with the same input. In something relating to intelligence, anyway; input-output of words probably don't cover the entirety of all possible detection mechanisms, but they do seem to me to cover a lot.
More options
Context Copy link
Because it was an impossibly high bar. Nothing was able to do that, for years. The idea that you’d be able to talk to a computer program and not recognize it seemed like science fiction.
More options
Context Copy link
Very intuitive, sensible, and wasn’t surpassed for 80 years.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
Interesting idea! Although there is definitely CG from the '90s that still looks downright good. Jurassic Park comes to mind as a masterpiece, which largely worked because the artists understood what worked well with the technology of the time: night shots (few light sources, little global illumination) of shiny-but-not-reflective surfaces (wet dinosaurs), used sparingly and mated with lots of practical effects.
CG only became a negative buzzword when it got over hyped and stretched to applications that it just wasn't very good for at the time. In some ways it's improved since (we can render photoreal humans!), but it still does get stretched in shots that are IMO just bad movie making ideas ("photorealistic, yet physics-defying").
I could see AI slop going the same way: certain "tasteful" uses still look good, but the current flood of AI art (somehow all the girls have the same face, and I've definitely spotted plenty of online ads that felt cheap from obvious AI use) will be "tacky" and age poorly.
More options
Context Copy link
More options
Context Copy link