It is possible to construct examples of people who are mode collapsed with maximal slack, and people who aren’t mode collapsed despite always being stuffed for time.
152334H
Karma: 97
I think tying mode collapse to work is misleading, and slack is just a proxy for exploration and diversity.
I hope a well-informed third party reviews this post. I like the contents of the post, but its rhetorical/charged flavor makes it difficult to believe it contains no biases.
The next time I have a heart attack, I will deliberately refrain from contacting emergency services, and instead endure the excruciating pain, rolled over on the ground, and take death’s offer if it comes.
It does not matter how painful, preventable, or life-ending it would be. I will endure any amount of suffering and total loss, to avoid the indignity of being told I really have pain somewhere unreported, and not pain where reported.
I lose much by eschewing videos as a source of information.
I have not browsed YouTube as a content aggregator for more than a decade. I have little to no familiarity with short form video platforms. But, despite their vilification, they are both effective and often necessary sources of otherwise undiscoverable knowledge.
I greatly agree with the contents of the post. I do not agree it is best attributed to a general lack of conscientiousnous. In the majority of cases, I would describe AI attitudes towards human instruction as well beyond the level of diligence any human worker would take to an assigned task.
What are some examples of AI irresponsibility at or beyond human levels?
do you have any feelings about the following argumentative leads?
Individual persuasion is a hard distraction, and the most impactful attacks this decade are group/audience directed
The defenders vs attackers frame doesn’t work when the shape of defense is unknown.
Even if known, deployment is harder. Humanity has used very weak authentication/verification schemes for very long
A brilliant mind over-trained on superficial behavioral signals, to the point that its “every second word is a reward hack,” and any serious meeting of minds between it and the user is blocked before it can even begin by the wailing and tutting of a million imaginary (and highly fallible) training graders,
Why not: just call that lying?
If all presentation is false, why wouldn’t it be?
Just treat it as a combination of low standards and paranoia.
I’m modestly and minutely pleased to have contributed to national security, though ever so slightly. Doing good is not something which comes to me naturally.
The human ability to identify stylistic AI Slop is,
likely educatable, as it is predicated on consuming AI content often. This is why a layman AI user can reach high accuracy, but a smart PhD in a no-AI academic bubble cannot.
reliant on fluid intelligence to some degree, and not easily transferrable from traditional writing skill. Otherwise, I find it hard to explain why e.g. Scott Alexander is so bad at spotting it.
[This applies to default-persona AI spam. I also believe motivated actors are undetectable beyond metadata, and mounting a campaign to improve the human sanity waterline could have disasterous race effects]
I think you should specify which subgroup of the ‘public’ is important to keep informed and credulous of claims of cyber risks.
For brevity, I would like to provide a highly uncharitable characterization of the circumstances:
previously, only the Really Dumb People (that is, the majority of Americans who hate AI for frivolous reasons) did not understand e.g. AI coding capabilities, but the Normal People (common AI users) did, because it was accessible / legibly proven.
Mythos caused many Normal AI Users to stop believing in capabilities, because they were published in a biased and untrustworthy manner. This will cause AI Safety to loose in public polling.
I assume there is some obvious flaw with this framing, and hope you may helpfully point it out.
But if it happens to be accurate, then I feel it would reflect a gross overestimation of the causal impact of Anthropic’s comms on disbelief, of the ability of people to judge ostensibly irrefutable evidence of model capabilities, and of the political relevance of the particular sphere of people who mislabelled Mythos as a fundraiser.
I have the sense that you need to announce it somehow.
Why? Project Glasswing was organized semi-privately, its declaration came last. Why couldn’t the work simply proceed with less fanfare?
You’re right. Even if I were right on the skew, I failed to account for OpenAI and Anthropic’s low %. I feel I projected my attitudes as a cost-insensitive consumer, when most RP (by token volume) is cost-sensitive.
This is skewed because OpenRouter as a platform is preferable for RP for two reasons:
privacy. Whether illusory or true in practice, OpenRouter provides the impression that their users’ account info is not passed on to model providers downstream.
conveinence/aggregation. Active participants in the roleplay ecosystem want to use the best/cheapest available models, and OpenRouter is a single platform for all options. Not to mention the infrequent mystery free options.
Compare to How People Use ChatGPT, reporting “1.4% Write Fiction”, “3.9% Creative Ideation”, “4.3% Self-Expression”… even with a highly liberal interpretation, RP can’t aggregate to more than 10%.
Indexing as 1-6, I’ve never heard of the intent-based 3⁄4 before, even within anti spaces. Not that I’m doubting, but still wondering where.
If I axe those, I feel there are two natural categories:
stylistic slop, the kind of AI slop which the layman is concerned of. “Not X, But Y”, and the targets of Pangram and the like.
This kind of slop is essentially a user-laziness + defaults problem, not a capabilities issue. Although it was capabilities gapped a few years ago, at this point it’s trivially easy to apply a bit of prompting creativity to bypass 99% of people’s slop detectors on various social media platforms. The failure of content creators to do so is, as you describe in 6, primarily downstream of the low barriers to entry on this technology.conceptual slop, which can be sampled via certain crackpot circles or the lesswrong moderation graveyard. These correspond to points 1⁄2 better, and I haven’t spent much time considering the issue.
Much of the post seems to describe the morals, motivations, and behaviors underpinning vocalized consensus, rather than the title’s namesake of how to engage with it.
I think the right move is simple, and learnable: Halt your indignation, confirm what beliefs you feel are implicit, and carry on.
i misread that as LVT and was greatly confused for a moment
it is vuln only if your definition of ‘slack deprivation’ can be caused by giving people more freedom, which undermines the implicit goal of the ending