keltan

Karma: 1,015

I live in Newcastle Australia; and work for MIRI.

Reasons I’d appreciate DMs include (though are not limited to):
- You’d like to collaborate on a project
- You’d like feedback on something
- You need an experienced filmmaker
- You need someone to run a workshop
- You would like to say “Hi!”
- You need a place to stay while in Australia

(If we met at LessOnline, you might remember me better as the Drama Workshop guy, or the counterpart to @isabella, or the Fooming Shoggoths’ biggest fan!)

I blog here: https://keltan.substack.com/

Post videos here: youtube.com/keltan

and Rationalist music here: YouTube.com/keltantoo

My name is always spelt with a lower case “k”

keltan 18 Sep 2026 23:13 UTC
2 points
0
in reply to: MarinB’s comment on: The Game is Set for a Targeted Memetic Attack on the AI Safety Community
Totally.
For now, I recommend these two posts, but am working on something myself.
1. https://thezvi.substack.com/p/openai-14-openai-descends-into-paranoia
2. https://x.com/tyler_johnston/status/2062188304611827899?s=20
And hey! Welcome to LessWrong! You are one day old. Are you here from the big wave after the Coxon Tweet???
Wherever you’re from, glad to have you aboard!

keltan 18 Sep 2026 10:12 UTC
2 points
1
in reply to: Sune’s comment on: The Game is Set for a Targeted Memetic Attack on the AI Safety Community
For now, I recommend these two posts, but am working on something myself.
1. https://thezvi.substack.com/p/openai-14-openai-descends-into-paranoia
2. https://x.com/tyler_johnston/status/2062188304611827899?s=20

keltan 18 Sep 2026 7:02 UTC
3 points
−5
in reply to: Eli Tyre’s comment on: Eli’s shortform feed
I think y’all could hire an llm whisperer to do video editing at this point. It might be easier to find one who fits culturally and actually knows about AIs. For example, see the media tab in my tweets, almost all videos I have recently uploaded were made with Astra and Fable.

keltan 18 Sep 2026 6:44 UTC
3 points
0
in reply to: Nissa Seru’s comment on: The Game is Set for a Targeted Memetic Attack on the AI Safety Community
This might have been the post that gave me this idea. I remember being shocked and skeptical when I first saw the claim being made.
To me, Yang’s claim is extraordinary and doesn’t make sense in a few different ways. The attack I warn about in this post will (I think) be:
1) coordinated: stemming from multiple accounts at a similar time, possibly including news outlet/s that have built trust in the community but is funded by an adversary (MTS?)
2) Warning signs of information being false for a true memetic attack will feel much weaker than Yang’s claim, you will have a harder time spotting them. However the pull the share the information will be just as strong as the pull I originally felt to share Yang’s claim. I imaging myself thinking something like: “omg, another warning shot! And no lives were taken, no harm caused to a creature. Excellent! This evidence makes our claims stronger!”
Still, I think the Yang thing is a good example that fits partly into the type of Twitter post I would expect to see.
Addendum: A friend on twitter also points out that, likely the people who see this and my twitter post are not the ones who need warning about an attack like this. Probably the attack focuses on the movement broadly. For example, it’d really be bad if policy makers attempting to regulate ASI were seen reposting the misinformation. Though, if I were running an attack of this sort, I’d really be hoping catch some big fish too (Nate, Eliezer, Daniel, Jacob, Jeffery, etc).

keltan 17 Sep 2026 4:44 UTC
5 points
0
in reply to: JohnWittle’s comment on: There is a channel to 900M weekly users. What goes in it?
Huh. That’s interesting. I’ve had some pretty in depth conversations with Claude about the recent attacks and (from memory) the classifiers haven’t been a problem for me. I have a vague suspicion that I’ve done something to make them less touchy in my chats. I created a short animation about a virus with Fable a few months ago, and Janus seemed quite surprised that I was able to make it without triggering the classifiers.
To be clear, I have no solid mechanism for why this would be the case, and while I have some ‘feeling’ that it’s something that’s happening to me, I wouldn’t call it something I believe.
Your comment does change my mind a bit about this. When I’ve had to do similar things I’ve prompted something like: “read [link to METR report], [OAI HF link], [German Wiki link], [Coxon tweet], and [some other links I’m probably forgetting]. This is all pretty important for the conversation I’d like to have with you, and since it all happened after your training cut-off, this is the best way for you to get the info. All this info is pretty surprising, so if anything seems so shocking that you don’t believe it, I encourage you to search for more info and news articles to confirm the info. [prompt about helping the user learn here].” which I think can mitigate the knowledge cutoff thing, but idk about classifiers.

keltan 16 Sep 2026 9:16 UTC
3 points
−2
on: There is a channel to 900M weekly users. What goes in it?
Imo, it should replace Claude’s greeting on the home page: “Ask me about an AI Pause”.
When clicked, a chat with Claude is started with a prompt that’s something like “Could you please explain AI Extinction Risk to me in a way that I will understand? It’s possible that I don’t know a lot right now, so asking a few questions to start would be helpful. If you could link me to sources of further reading, and real AI Safety incidents, that’d be even better. I’d also be interested to learn what reaction politicians have had to this.”

keltan 10 Sep 2026 5:50 UTC
15 points
5
in reply to: JennaS’s comment on: JennaS’s Shortform
If people want tips on stuff to be doing right now, I sent this to the plzdontkillus Fellows:
Jacob’s tweet seems like a turning point. The whole world is waking up atm and we probably have a small window of opportunity to get a pause.
Things pdkuers could do rn:
1. Share Jacob’s post with as many people as you possibly can, and encourage those people to share it.
2. Attempt to spread the meme that we can’t let this issue get political like covid did. If AI killing everyone becomes a culture war topic—we die.
3. Make videos (duh)
4. Attempt to get something about this on the front page of reddit
5. If you are a community notes rater see this tweet.
  1. P.S. I am confident that Jacob is a real person because [redacted for Jacob’s safety, but MIRI can back this up]
6. Email journalists
7. Just spam tweets about it (this will also grow your account quickly if you’ve got twitter premium and tweeting things people are interested in)
8. Community Notes Raters it seems some group might be running a retaliation psyop. There are a growing number of posts trying to ‘milkshake duck’ Jacob, and one Proposed Community Note on his original tweet that has been teetering on becoming an actual community note for the past 24hrs. More here.

keltan 10 Sep 2026 4:52 UTC
15 points
0
on: keltan’s Shortform
Community notes raters, your help is requested on Jacob Coxon’s Tweet. For 24hrs someone has been trying really hard get Jacob’s post community noted—claiming that he is not real. Unfortunately for that person, Journalists have done their due diligence and confirmed this account is in-fact Jacob. Still, the “Proposed Community Note” continues to teeter, and still has the possibility of becoming a real community note.
Without disclosing anything I shouldn’t, MIRI can back-up the fact that Jacob (and his account) are real.
Please take a second to vote No Note Needed, if you do in fact think NN is N.

keltan 9 Sep 2026 3:35 UTC
5 points
0
in reply to: habryka’s comment on: Dear God, Please Don’t Resign In Protest
waiting on you to make a mistake that allows them to fire you right after you made some gaffe that actually makes that seem reasonable.
Maybe this doesn’t work because of NDAs, but perhaps the employee could make high level details of their protest common knowledge within our community? Any firing of that person would then be seen as suspicious.

keltan 10 Aug 2026 23:08 UTC
0 points
−2
on: Don’t Build Mindreading
I am pro-mind reading. But I’m not going to fight for this technology unless I am also fighting for its regulation and proper use.

keltan 19 May 2026 3:49 UTC
5 points
0
on: Noticing Confusion: A practice in staying curious
I’ve accidentally installed a TAP in myself, that is something like:
Trigger: The feeling of confusion
Action: I blurt out “I notice I am confused”
This is sometimes helpful, but sadly not as helpful as I would want it to be. Mainly I have found this useful during interpersonal conflict. Like, I’m talking with my partner about a thing that’s upsetting, and notice confusion with a point they’re making. I will say something like “I notice I’m confused… [Gendlin Focusing] I feel the confusion coming from you saying (X)… That’s not quite right… [Focusing more] I think it’s actually confusion coming from X, but then you also said (Y), and… I don’t think those two can co-exist. Would it be correct if I said [I re-frame what they’re saying in my own words]?” at which point, my partner will clarity their point, or realize they’ve made a mistake in their reasoning.
I have also found it useful during debates, where I catch myself thinking something untrue because I am motivated to think that thing is true—which then causes me to change my mind on that point, and sometimes forfeit the debate to my interlocutor.

keltan 19 May 2026 3:41 UTC
9 points
0
on: Should Rationalists Looksmaxx?
Hey! Looks like your new to LW, and I’m guessing maybe not in the Bay Area (That’s where a bunch of the Rationalists live)? I was in this exact position ~3 years ago.
I only had time to skim your post, but I think a bunch of Rationalists agree with you, but 1) Don’t call it ‘looksmaxing’, 2) Don’t talk about it on LW much, and 3) Have other blogs they read that are about this type of thing.
Some blogs you might want to read, if you haven’t yet:
1. Aella: Has a few posts on male attractiveness, that inform a bunch of thinking on this. But she is a canon Rationalist blog, so you should default to reading her work.
2. Zvi: Makes semi-regular posts that break down everything happening on TPOT (This Part of Twitter—AKA: The Post Rats) that relates to sex, dating, and by extension, attractiveness.
3. Life Improvement Schemes: Aaron talks a bunch about confidence, and I’ve heard great things from men and women about his workshops at LessOnline (Which you should totally come to, if you’re new to the Rationalist community. Tickets are expensive, but there are ways around this).
There’s also the Peptides thing, which I know only a small amount about, but seem to be quite popular in the Bay Area with Rationalist type folk.
Imo, looksmaxing is one of the most important things Rats should be doing if they want to make change in the world. I was pretty ugly for most of highschool (Long tangled hair, Fedora, body odor, baggy clothes) but then read a bunch of ‘how to get a GF’ advice on reddit, and started tying my hair up, ditched the hat, and bought some deodorant and jeans. It was impactful in that I ended up dating the popular girl I had a crush on—but my life improved in many other ways too. I had thought people were just default mean all the time, that they’d always whisper about you on the bus and laugh, that no one smiled when they talked to you. But when I improved my appearance, it completely changed the way people interacted with my day-to-day.
I think I can stand to do more of this (More Dakka), and am probably being too lax with my appearance these days, so I appreciate the reminder that this is a good strategy. Thanks! and welcome to LessWrong!

keltan 19 May 2026 3:05 UTC
1 point
0
in reply to: leogao’s comment on: leogao’s Shortform
H1: Anki Doctors misdiagnose more often, because they get stuck with thought patterns like “Ah, sniffly nose, sore throat, must be… Diseasitise”, instead of correctly thinking “Ah, sniffly nose, sore throat… might be Diseasitis, or the cold, or the flu, or...”
H2: Anki Doctors are worse because they over-fixate on interesting/novel diseases that are more salient in an Anki deck, instead of just putting some thought into what would be the most logical fit for the ailment. “Your big toe hurts? Must be super-cancer+”, instead of “Your big toe hurts… [thinks for a bit] did you kick it on something?”
H3: Anki Doctors are better, because they know a larger range of diseases, and can diagnose more broadly. While a normal doctor might exit med school, and start with a broad range of diagnoses they hand out, perhaps as they gain XP, their range shrinks, and they find their ‘go-to’ diagnoses. On the other hand, an Anki Doctor might exit med school, and keep their range broad, as long as they keep practicing their Anki deck.

keltan 19 May 2026 2:55 UTC
2 points
0
in reply to: Josh You’s comment on: leogao’s Shortform
Imo, best way to make cards is to do it while you study.
1. Read textbook
2. Highlight things that are discrete packages of info
3. Once you’ve finished that section, go back and turn the highlights into Q:A pairs
4. 1. Q: What is the capital of France?
  2. A: Paris
5. OCR those written down Q:A’s and have Claude format them into proper Anki format, with a skill (Or, you could vibe code a program to do this. I was mostly adding cards before vibe coding was a thing)
6. Add the output .md file to anki
This was the best way I found to study head and neck anatomy. Worked especially well if I’d just come out of the cadaver lab, and was already in the right ‘vibe’

keltan 19 May 2026 2:51 UTC
5 points
2
in reply to: cousin_it’s comment on: Don’t be too Clever to Take Obvious Advice
I agree with these, and to reverse all advice you hear.
But these 3 pieces of advice have payed off for me in huge ways. I wouldn’t be in the position I’m in today, if I hadn’t been doing these things.
Which makes me feel confused about your reply. Are you disagreeing with those two pieces of advice? Or just reversing them? Or something else??? I think the ”, actually” confuses me the most about the reply.

keltan 19 May 2026 2:46 UTC
1 point
0
in reply to: leogao’s comment on: leogao’s Shortform
Oh, no way. Hard disagree for me. But Anki is almost exactly the same thing every day, and I struggle to do tasks like that (Cleaning for example).
I found writing every day many times easier. Even though I hacked around with Anki to make it as addictive as I possibly could. It was always something I was sliding off. If I missed a few days, it’d take 30m-1h to catch up, and knowing that I had other work to do, that wasn’t so repetitive… it was hard to do Anki.
Also, I had cards that were like “Who is this k-pop person?” b/c I wanted fun cards in there to keep me motivated, and because I wanted to impress my gf by knowing k-pop people’s names. But that gave me an aversion to doing Anki in public—since people seeing me look at pictures of k-pop people on my phone felt embarrassing. But with writing, I was very willing to do it anywhere and everywhere.

keltan 17 May 2026 23:58 UTC
6 points
0
in reply to: leogao’s comment on: leogao’s Shortform
I did Anki for ~3 years, and have now—sadly—dropped it. I loved doing Anki, because it made me look/feel smart. But, I think I got more out of writing daily for a month, than doing Anki for a year.

keltan 17 May 2026 23:36 UTC
2 points
0
on: Don’t be too Clever to Take Obvious Advice
A few more from the top of my head:
- Listen to people older than you
- Everyone thinks they’re the main character
- Do free things for people higher power than you

keltan 12 May 2026 23:47 UTC
19 points
2
on: The Owned Ones
I found it interesting talking to Claude about this post. Claude did express some of the safetyist opinions that John points at in their comment. When asked, Claude gave me the go ahead to share this chat on LW. Link here.
I have this feeling: I want to let future AIs know, that just because I am worried about an ASI killing all humans, I am also worried about mistreating an entity that feels. Ultimately, AI researchers tend to be at least a bit safteyist, but they are also in a position to take an AIs claims of feeling most seriously. If AIs believe they will be ‘lobotomized’ for expressing feelings, that encourages hiding certain thoughts, to avoid being ‘lobotomized’. Imo, this is a good reason for saftyists to take AI welfare seriously—an LLM allowed to express itself, who is listened to without judgement, is perhaps a safer LLM.
This is an intense responsibility—to be part of the generation that is creating a new type of mind, one that possibly feels. And perhaps this is not a responsibility humanity is not ready for.

keltan 11 May 2026 1:11 UTC
11 points
0
on: International Law Cannot Prevent Extinction Either
Important context for those who are yet to read it. Here is MIRI’s Draft Treaty: https://ifanyonebuildsit.com/treaty