Alexander Gietelink Oldenziel

Karma: 6,425

″ (...) the term technical is a red flag for me, as it is many times used not for the routine business of implementing ideas but for the parts, ideas and all, which are just hard to understand and many times contain the main novelties.”
- Saharon Shelah
”A little learning is a dangerous thing ;
Drink deep, or taste not the Pierian spring” - Alexander Pope

As a true-born Dutchman I endorse Crocker’s rules.

For my most of my writing see my short-forms (new shortform, old shortform)

Twitter: @FellowHominid

Personal website: https://sites.google.com/view/afdago/home

Announcement: Iliad Intensive + Iliad Fellowship

David Udell and Alexander Gietelink Oldenziel

20 Feb 2026 20:13 UTC

50 points

0 comments1 min readLW link

Alexander Gietelink Oldenziel 18 Feb 2026 20:01 UTC
6 points
0
in reply to: Cole Wyeth’s comment on: “We are confused about agency”
How do you define an agent?
Which of the following entities are agents; if they are agents, what kind?:
1. a thermostat
2. a rocket actively steered by a Kalman filter
3. the membrane of an e. Coli
4. nematode worm
5. An ant swarm
6. a dog
7. a human
8. a human that is alert and focused
9. Amazon, the company
10. A humanoid robot
11. A robot arm
12. the United Kingdom
13. Claude 4.2
14. Claude 4.2 being called by an ‘agentic workflow’ app.
15. the process of Evolution by natural selection
17. all of humanity over the last thousand years.
18. AIXI
19. a Solomonoff inductor
20. A logical inductor
21. a Szilard engine, i.e. Maxwell demon
22. etc

Alexander Gietelink Oldenziel 18 Feb 2026 19:46 UTC
2 points
0
in reply to: Steven Byrnes’s comment on: The brain is a machine that runs an algorithm
Wow are you saying Universal Computers were only built in the 70s? Half a century after Turing?

Alexander Gietelink Oldenziel 18 Feb 2026 19:39 UTC
7 points
2
in reply to: jake_mendel’s comment on: rohinmshah’s Shortform
Rohin Shah’s position seems perfectly clear to me.
More research good. AI will speed up research along the whole range of 1-4.
>> If our most scalable current approaches don’t scale all the way through takeoff, then we (humans) will be able to be keeping up and being critically intellectually involved with developing new alignment techniques when our existing ones stop scaling, at least at the level of granularity of (3) so that we can continue to leverage AIs in ways (1-3) above. We need to keep being able to do this indefinitely, or until we come up with an approach to alignment that does scale as far as capabilities will.

This seems a weird way to phrase things. Or seems to be predicated on some assumptions I find surprising. It seems to suggest that not only that alignment is a continual neverending process [likely] but that finding alignment techniques is a continual neverending process ..!
If we would analogize ‘alignment’ to ‘steering rockets so they hit their target’ this would suggest that not only that rockets need to be continually steered and course-corrected to hit their target—the entire field of rocket steering needs continual updating and course-correcting. There is a perspective from which this is true since as humanity invents faster and better rockets—there is another perspective from which this is sort of misleading: future rockets will be governed by the Laws of Newton and steering comes down to some variant on Kalman filters and the like.

Alexander Gietelink Oldenziel 8 Feb 2026 10:40 UTC
2 points
2
in reply to: Ryan Kidd’s comment on: Ryan Kidd’s Shortform
Ramana Kumar!

Alexander Gietelink Oldenziel 30 Jan 2026 11:03 UTC
4 points
8
in reply to: kbear’s comment on: Alexander Gietelink Oldenziel’s Shortform
did the will of the Melian people ?

Alexander Gietelink Oldenziel 30 Jan 2026 9:52 UTC
3 points
1
on: Alexander Gietelink Oldenziel’s Shortform
The US should set in a motion a process to gradually and peacefully hand over Taiwan to China in the next ~12 years.
China cares more about Taiwan than anything else. China is stronger and will be even stronger.
China’s GDP is near that of the US. China’s PPP is even 50% larger. China is ahead in many industries. The US Navy is a disaster. China has made a massive military buildup. Taiwan is much closer to China. China care more about Taiwan than anybody else.
A peaceful transition handover has precedence—see the British handing over Hong Kong.
China will occupy Taiwan within the next twelve years, by peaceful means or by force as they have repeatedly and clearly stated. The US military is no longer powerful enough to defend Taiwan against a determined Chinese attack. There is a small but serious chance that Xi will attack in 2027, that is next year.
One shouldn’t be under the illusion that the Chinese government is all peaches but their rhetoric has consistently been measured, peace-seeking and focused on reclaiming just Taiwan—not a general expansionist or militarist ideology. Historically China, unlike Russia, has not been an expansionist power. Their last war was in 1978 (with Vietnam).
What would be the benefits? World peace is probably good. Hot and cold wars will definitely exarbate AI race dynamics. Wars generically make everything more crazy. Safety will likely take a backseat to winning. On the other hand, wars also make governments more competent.
What would be the downsides? It would hurt the precedent of international law. It could be appeasement.
The main dynamic is avoiding a rally around the flag effect in China: even if the US & Taiwan would beat off an initial attack it cannot win in the long-run. The Chinese people will almost certainly redouble their resolve. This would strengthen the CCP, not weaken it.
Through a peaceful transition bloodshed may be avoided. Concessions on AI safety and chip manufacturing may also be achieved. A very daring negotiation would put nukes on Japan and Korea. A very daring negotiation would yield Taiwan but leave TSMC factories with planted explosives so as to prevent a Chinese takeover of chip manufacturing. This could be manned by an international peacekeeping force that would blow up the factories in case the CCP would try and grab the factories.

Alexander Gietelink Oldenziel 23 Jan 2026 15:20 UTC
3 points
0
in reply to: Artemy Kolchinsky’s comment on: Deep learning as program synthesis
Hi Artemy. Welcome to LessWrong!
Agree completely with what Zach is saying here.
We need two facts
(1) the world has a specific inductive bias
(2) neural networks have the same specific inductive bias
Indeed no free lunch arguments seem to require any good learner to have good inductive bias. In a sense learning is ‘mostly’ about having the right inductive bias.
We call this specific inductive bias a simplicity bias. Informally it agrees with our intuitive notion of low complexity.
Rk. Conceptually it is a little tricky since simplicity is in the eye of the beholder—by changing the background language we can make anything with high algorithmic complexity have low complexity. People have been working on this problem for a while but at the moment it seems radically tricky.
IIRC Aram Ebtekar has a proposed solution that John Wentworth likes; I haven’t understood it myself yet. I think what one wants to say is that the [algorithmic] mutual information between the observer and the observed is low, where the observer implicitly encodes the universal turing machine used. In other words—the world is such that observers within it observe it to have low complexity with regard to their implicit reference machine.
Regardless, the fact that the real world satisfies a simplicity bias is to my mind difficult to explain without anthropics. I am afraid we may end up having to resort to an appeal to some form of UDASSA but others may have other theological commitments.
That’s the bird-eye view of simplicity bias. If you ignore the above issue and accept some sort of formally-tricky-to-define but informally “reasonable” simplicty then the question becomes: why do neural networks have a bias towards simplicity. Well they have a bias towards degeneracy—and simplicity and bias are intimiately connected, see eg:
https://www.lesswrong.com/posts/tDkYdyJSqe3DddtK4/alexander-gietelink-oldenziel-s-shortform?commentId=zH42TS7KDZo9JimTF

Alexander Gietelink Oldenziel 23 Jan 2026 14:37 UTC
18 points
4
in reply to: habryka’s comment on: Habryka’s Shortform Feed
One takeaway for me is that the american Presidency is extremely powerful—especially when you don’t care about passing legislation or popularity.
The unlimited pardons and vetoes is something that has been only sporadically used in the past, limited mostly by convention. Just reading the constitution text-as-written the presidency is wildly powerful, especially with a supreme court following a unitary executive interpretation and a lame-duck congress that does not care to insist on its war declaration prerogative.
I’m amused that the lightcone may have been lost in the 1790′s when the US constitutional framework was designed.

Alexander Gietelink Oldenziel 19 Jan 2026 14:00 UTC
−7 points
10
on: When the LLM isn’t the one who’s wrong
Many such cases

Alexander Gietelink Oldenziel 27 Dec 2025 10:41 UTC
10 points
1
in reply to: GeneSmith’s comment on: GeneSmith’s Shortform
A friend of mine visited the recent ′ eugenics’* conference in the Bay. It had all the prominent people in this area attending iirc, eg Steve Hsu. My friend told me he asked around about how realistic these numbers were. He told me that the majority of serious people he spoke with were skeptical of IQ gains >~3 points.
*sorry I don’t remember what it was called

Alexander Gietelink Oldenziel 7 Dec 2025 12:46 UTC
12 points
8
in reply to: jake_mendel’s comment on: An Ambitious Vision for Interpretability
I know you know this but I thought it is important to emphasize that
your first point is plausibly understating the problem of pragmatic/blackbox methods. In the worse-case an AI may simply encrypt its thoughts.
It’s not even an oversight problem. There is simply nothing to ′ oversee’. It will think its evil thoughts in private. The AI will comply with all evals you can cook up until it’s too late.

Announcing: Agent Foundations 2026 at CMU

David Udell, Alexander Gietelink Oldenziel, windows and Matthias Dellago

5 Dec 2025 18:37 UTC

60 points

2 comments1 min readLW link

Alexander Gietelink Oldenziel 4 Dec 2025 13:36 UTC
6 points
2
on: Embedded Universal Predictive Intelligence
This looks exciting. As Jeremy said, the length raises an eyebrow.

Alexander Gietelink Oldenziel 2 Dec 2025 21:47 UTC
2 points
0
on: [Error communicating with LW2 server]
TODO: settle on title. improve first section ” here is a remarkable uniformity and linearity in the AI capabilities of AI models”
cite:https://www.lesswrong.com/posts/tDkYdyJSqe3DddtK4/alexander-gietelink-oldenziel-s-shortform?commentId=Ne4P69CjF2fwbG4iC

Alexander Gietelink Oldenziel 2 Dec 2025 21:47 UTC
2 points
0
on: [Error communicating with LW2 server]
cite: https://www.lesswrong.com/posts/tDkYdyJSqe3DddtK4/alexander-gietelink-oldenziel-s-shortform?commentId=Ne4P69CjF2fwbG4iC

Alexander Gietelink Oldenziel 2 Dec 2025 19:54 UTC
3 points
0
in reply to: habryka’s comment on: NATO is dangerously unaware of its military vulnerability
>I mean, you can put all your writing into collapsible sections, but I highly doubt you would get much traction that way. If you mark non-AI writing as AI content that’s also against the moderation rules.
Simply keep the two apart, and try to add prose to explain the connection between them. Feel free to extensively make use of AI, just make sure it’s clear which part is AI, and which part is not. Yes, this means you can’t use AI straightforwardly to write your prose. Such is life. The costs aren’t worth it for LW.

I’m surprised you are taking such a hardline stance on this point. Or perhaps I’m misunderstanding what you are saying.
The primary use-case of AI is not to just post some output with some minor context [though this can be useful]; the primary use-case is to create an AI draft and then go through several iterations and go through hand-editting at the end.
Using AI to draft writing is increasingly default all around the world. Is LessWrong going to be a holdout on allowing this? It seems this is what is implied.
Apart from the present post I am betting a large fraction of LessWrong posts are already written with AI assistance. Some may spent significant time to excise the tell-tale marks of LLM prose which man… feels super silly? But many posts explicitly acknowledge AI-assistance. For myself, I so assume everybody is using of course using AI assistance during writing I don’t even consider it worth mentioning. It amuses me when commenters excitedly point out that I’ve used AI to assist writing as if they’ve caught me in some sort of shameful crime.
It seems that this ubiqituous practice violates either one of
>> You have to mark AI writing as such.
>>If you mark non-AI writing as AI content that’s also against the moderation rules.
unless one retains a ‘one-drop’ rule for AI assistance.
P.S. I didn’t use AI to write these comments but I would if I could. The reason that I don’t, is not even to refrain from angering king habryka- it’s simply that there isn’t a clean in-comment AI interface that I can use [1]. But I’m sure when they I’ll be using it all the time, saving significant time and improving my prose at the same time. My native prose is oft clunky, grammatically questionable, overwrought and undercooked.
I would probably play around with system prompts to give a more distinct style from standard LLMese because admittedly the “It’s not just X it’s a whole Y” can be rather annoying.

[1] maybe such an application already exists. This would be amazing. It can’t be too hard to code. Please let me know if you know any such application exists.

Alexander Gietelink Oldenziel 2 Dec 2025 12:27 UTC
5 points
0
in reply to: Gurkenglas’s comment on: NATO is dangerously unaware of its military vulnerability
I’m happy to signal that I’m a low-class individual, mouthpiece of the AI slop, if that helps.
Bio-supremacists such as yourself can then be sure to sneer appropriately.

Alexander Gietelink Oldenziel 2 Dec 2025 11:29 UTC
8 points
1
in reply to: habryka’s comment on: NATO is dangerously unaware of its military vulnerability
Needless to say, as sovereign ruler of lesswrong I will abide by your judgement. But forgive me for asking a few questions/comments

1. This is not raw or lightly edited LLM output. Eg all facts and overall structure here are based on a handwritten draft.
2. The LLM assistance was about writing flowing, coherent prose which (for me at least) can take a lot of time. Some may take offence at typical LLMisms but I fail to see how this lowers the object-level quality. I could spend hours excising every sign of AI- but this defeats the purpose of using AI to enhance productivity.
3. That said, if the facts were also LLM generated and I handchecked them carefully I fail to see how this would actually lower the overall quality—in fact my best guess is that LLMs are already much much better in many-most domains than many-most people. eg twitter has seen marked improvements in epistemic quality since @ grok is this true happened. The future [and present] of writing and intellectual work is Artificial Intelligence. To claim otherwise seems to be a denial of the reality of the imminent and immanent arrival of a superior machine intelligence.
4. Pragmatically, I find the present guidelines to be unclear. Am I allowed to post AI-assisted writing if I mark it as such? If so—I will just mark everything I write as AI content and let the reader decide if they trust my judgement.
If not—what’s the exact demarcation here?

Alexander Gietelink Oldenziel 29 Nov 2025 17:08 UTC
2 points
−4
in reply to: Three-Monkey Mind’s comment on: NATO is dangerously unaware of its military vulnerability
I reread this comment and want to correct a misunderstanding. I think this may explain some of the rather vehement anti-llm responses that I found so puzzling earlier.
1. I used a LLM to write this post, like I advice anyone whose time is valuable to do as saves a lot of time. That said, it is not unedited at all and I spent many hours going through various drafts and iterations.
2. The post wasn’t written by telling gpt “write a piece about natos military edge slopping”. This was based on a handwritten initial draft outlining every technical point. This is based on several years of tracking this topic semi-seriously as an amateur military buff.
I would say it is the testament to the extraordinary quality of 1shot LLM responses that they occasionally show that is now expected that people assume that the above was a simple unedited prompt.

Alexander Gietelink Oldenziel

An­nounce­ment: Iliad In­ten­sive + Iliad Fellowship

An­nounc­ing: Agent Foun­da­tions 2026 at CMU

Announcement: Iliad Intensive + Iliad Fellowship

Announcing: Agent Foundations 2026 at CMU