Nate Showell

Karma: 493

Nate Showell 13 Jul 2025 21:10 UTC
2 points
−1
in reply to: Cole Wyeth’s comment on: Cole Wyeth’s Shortform
More useful. It would save us the step of having to check for hallucinations when doing research.

Nate Showell 12 Jul 2025 3:40 UTC
27 points
4
on: Generalized Hangriness: A Standard Rationalist Stance Toward Emotions
Another example of this pattern that’s entered mainstream awareness is tilt. When I’m playing chess and get tilted, I might think things like “all my opponents are cheating, “I’m terrible at this game and therefore stupid,” or “I know I’m going to win this time, how could I not win against such a low-rated opponent.” But if I take a step back, notice that I’m tilted, and ask myself what information I’m getting from the feeling of being tilted, I notice that it’s telling me to take a break until I can stop obsessing over the result of the previous game.
Tilt is common, but also easy to fix once you notice the pattern of what it’s telling you and start taking breaks when you experience it. The word “tilt” is another instance of a hangriness-type stance that’s caught on because of its strong practical benefits—having access to the word “tilt” makes it easier to notice.

Nate Showell 8 Jul 2025 3:29 UTC
3 points
0
in reply to: Ruby’s comment on: LessWrong Feed [new, now in beta]
It’s working now. I think the problem was on my end.

Nate Showell 5 Jul 2025 20:21 UTC
6 points
3
on: ‘AI for societal uplift’ as a path to victory
This strategy suggests that decreasing ML model sycophancy should be a priority for technical researchers. It’s probably the biggest current barrier to the usefulness of ML models as personal decision-making assistants. Hallucinations are probably the second-biggest barrier.

Nate Showell 4 Jul 2025 23:37 UTC
1 point
0
in reply to: Ruby’s comment on: LessWrong Feed [new, now in beta]
The new feed doesn’t load at all for me.

Nate Showell 22 Jun 2025 5:51 UTC
28 points
15
on: Consider chilling out in 2028
There’s another way in which pessimism can be used as a coping mechanism: it can be an excuse to avoid addressing personal-scale problems. A belief that one is doomed to fail, or that the world is inexorably getting worse, can be used as an excuse to give up, on the grounds that comparatively small-scale problems will be swamped by uncontrollable societal forces. Compared to confronting those personal-scale problems, giving up can seem very appealing, and a comparison to a large-scale but abstract problem can act as an excuse for surrender. You probably know someone who spends substantial amounts of their free time watching videos, reading articles, and listening to podcasts that blame all of the world’s problems on “capitalism,” “systemic racism,” “civilizational decline,” or something similar, all while their bills are overdue and dishes pile up in their sink.
This use of pessimism as a coping mechanism is especially pronounced in the case of apocalypticism. If the world is about to end, every other problem becomes much less relevant in comparison, including all those small-scale problems that are actionable but unpleasant to work on. Apocalypticism can become a blanket pretext for giving in to your ugh fields. And while you’re giving in to them, you end up thinking you’re doing a great job of utilizing the skill of staring into the abyss (you’re confronting the possibility of the end of the world, right?) when you’re actually doing this exact opposite. Rather than something related to preverbal trauma, this usability as a coping mechanism is the more likely source of the psychological appeal of AI apocalypticism for many people who encounter it.
What links here?
- Noosphere89's comment on Consider chilling out in 2028 by Valentine (27 Jun 2025 16:54 UTC; 6 points)

Nate Showell 14 Jun 2025 5:31 UTC
3 points
2
in reply to: Daniel Kokotajlo’s comment on: Distillation Robustifies Unlearning
Another experiment idea: testing whether the reduction in hallucinations that Yao et al. achieved with unlearning can be made robust.

Nate Showell 4 May 2025 22:44 UTC
3 points
0
on: What’s up with AI’s vision
Do LLMs perform better at games that are later in the Pokemon series? If difficulty interpreting pixel art is what’s holding them back, it would be less of a problem when playing later Pokemon games with higher-resolution sprites.

Nate Showell 26 Apr 2025 0:21 UTC
5 points
0
on: This prompt (sometimes) makes ChatGPT think about terrorist organisations
Have you tried seeing how ChatGPT responds to individual lines of code from that excerpt? There might be an anomalous token in it along the lines of ” petertodd”.

Nate Showell 9 Apr 2025 6:20 UTC
1 point
0
in reply to: Adam Zerner’s comment on: Against podcasts
Occasionally something will happen on the train that I want to hear, like the conductor announcing a delay. But not listening to podcasts on the train has more to do with not wanting to have earbuds in my ears or carry headphones around.

Nate Showell 5 Apr 2025 21:10 UTC
3 points
3
on: Against podcasts
I hardly ever listen to podcasts. Part of this is because I find earbuds very uncomfortable, but the bigger part is that they don’t fit into my daily routines very well. When I’m walking around or riding the train, I want to be able to hear what’s going on around me. When I do chores it’s usually in short segments where I don’t want to have to repeatedly pause and unpause a podcast when I stop and start. When I’m not doing any of those things, I can watch videos that have visual components instead of just audio, or can read interview transcripts in much less time than listening to a podcast would take. The podcast format doesn’t have any comparative advantage for me.

Nate Showell 5 Apr 2025 4:22 UTC
8 points
0
on: Nate Showell’s Shortform
Metroid Prime would work well as a difficult video-game-based test for AI generality.
- It has a mixture of puzzles, exploration, and action.
- It takes place in a 3D environment.
- It frequently involves backtracking across large portions of the map, so it requires planning ahead.
- There are various pieces of text you come across during the game. Some of them are descriptions of enemies’ weaknesses or clues on how to solve puzzles, but most of them are flavor text with no mechanical significance.
- The player occasionally unlocks new abilities they have to learn how to use.
- It requires the player to manage resources (health, missiles, power bombs)
- It’s on the difficult side for human players, but not to an extreme level.
There are no current AI systems that are anywhere close to being able to autonomously complete Metroid Prime. Such a system would probably have to be at or near the point where it could automate large portions of human labor.

Nate Showell 22 Mar 2025 20:48 UTC
3 points
−6
on: They Took MY Job?
I recently read This Is How You Lose the Time War, by Max Gladstone and Amal El-Mohtar, and had the strange experience of thinking “this sounds LLM-generated” even though it was written in 2019. Take this passage, for example:
You wrote of being in a village upthread together, living as friends and neighbors do, and I could have swallowed this valley whole and still not sated my hunger for the thought. Instead I wick the longing into thread, pass it through your needle eye, and sew it into hiding somewhere beneath my skin, embroider my next letter to you one stitch at a time.
I found that passage just by opening to a random page without having to cherry-pick. The whole book is like that. I’m not sure how I managed to stick it out and read the whole thing.
The short story on AI and grief feels very stylistically similar to This Is How You Lose the Time War. They both read like they’re cargo-culting some idea of what vivid prose is supposed to sound like. They overshoot the target of how many sensory details to include, while at the same time failing to cohere into anything more than a pile of mixed metaphors. The story on AI and grief is badly written, but its bad writing is of a type that human authors sometimes engage in too, even in novels like This Is How You Lose the Time War that sell well and become famous.
How soon do I think an LLM will write a novel I would go out of my way to read? As a back-of-the-envelope estimate, such an LLM is probably about as far away from current LLMs in novel-writing ability as current LLMs are from GPT-3. If I multiply the 5 years between GPT-3 and now by a factor of 1.5 to account for a slowdown in LLM capability improvements, I get an estimate of that LLM being 7.5 years away, so around late 2032.

Nate Showell 21 Mar 2025 4:08 UTC
6 points
4
on: Why White-Box Redteaming Makes Me Feel Weird
As you mentioned at the beginning of the post, popular culture contains examples of people being forced to say things they don’t want to say. Some of those examples end up in LLMs’ training data. Rather than involving consciousness or suffering on the part of the LLM, the behavior you’ve observed has a simpler explanation: the LLM is imitating characters in mind control stories that appear in its training corpus.

Nate Showell 14 Mar 2025 6:34 UTC
4 points
0
in reply to: cubefox’s comment on: Daniel Kokotajlo’s Shortform
There are sea slugs that photosynthesize, but that’s with chloroplasts they steal from the algae they eat.

Nate Showell 9 Mar 2025 20:33 UTC
2 points
0
in reply to: Carl Feynman’s comment on: What is the best / most proper definition of “Feeling the AGI” there is?
As I use the term, the presence or absence of an emotional reaction isn’t what determines whether someone is “feeling the AGI” or not. I use it to mean basing one’s AI timeline predictions on a feeling.

Nate Showell 8 Mar 2025 22:26 UTC
3 points
0
on: What is the best / most proper definition of “Feeling the AGI” there is?
Getting caught up in an information cascade that says AGI is arriving soon. A person who’s “feeling the AGI” has “vibes-based” reasons for their short timelines due to copying what the people around them believe. In contrast, a person who looks carefully at the available evidence and formulates a gears-level model of AI timelines is doing something different than “feeling the AGI,” even if their timelines are short. “Feeling” is the crucial word here.

Nate Showell 2 Mar 2025 20:13 UTC
14 points
1
on: Share AI Safety Ideas: Both Crazy and Not
The phenomenon of LLMs converging on mystical-sounding outputs deserves more exploration. There might be something alignment-relevant happening to LLMs’ self-models/world-models when they enter the mystical mode, potentially related to self-other overlap or to a similar ontology in which the concepts of “self” and “other” aren’t used. I would like to see an interpretability project analyzing the properties of LLMs that are in the mystical mode.

Nate Showell 15 Feb 2025 22:04 UTC
8 points
0
in reply to: tailcalled’s comment on: tailcalled’s Shortform
The question of population ethics can be dissolved by rejecting personal identity realism. And we already have good reasons to reject personal identity realism, or at least consider it suspect, due to the paradoxes that arise in split-brain thought experiments (e.g., the hemisphere swap thought experiment) if you assume there’s a single correct way to assign personal identity.

Nate Showell 14 Feb 2025 2:04 UTC
4 points
0
on: My model of what is going on with LLMs
LLMs are more accurately described as artificial culture instead of artificial intelligence. They’ve been able to achieve the things they’ve achieved by replicating the secret of our success, and by engaging in much more extensive cultural accumulation (at least in terms of text-based cultural artifacts) than any human ever could. But cultural knowledge isn’t the same thing as intelligence, hence LLMs’ continued difficulties with sequential reasoning and planning.