Samuel Ratnam
32 followers · 44 following · 1376 views
on the atlas — 41
- Did Claude 3 Opus align itself via gradient hacking? — LessWrong1 savers
- [2410.01463] Selective Aggregation for Low-Rank Adaptation in Federated Learning1 savers
- Postscript on the Societies of Control | The Anarchist Library3 savers
- ⿻ Symbiogenesis vs. Convergent Consequentialism — LessWrong2 savers
- Encyclical Letter of His Holiness Leo XIV Magnifica Humanitas (15 May 2026)46 savers
- Attribution-Based Control for AI ⬩ OpenMined1 savers
- Synthetic Persona Pretraining: Alignment from Token Zero — LessWrong1 savers
- Current AIs seem pretty misaligned to me — LessWrong9 savers
- Teaching Claude why \ Anthropic6 savers
- The Bitter Lesson78 savers
- Do not conquer what you cannot defend — LessWrong6 savers
- DEBT INC.: GUILT, CREDIT, AND THE ALGORITHMIC FUTURE1 savers
- Eating Meat and Eating People2 savers
- Natural Intelligence: the Gaia architecture [v2 draft - April 2023]1 savers
- Towards a scale-free theory of intelligent agency4 savers
- Lachlan Marvit4 savers
- More people should write « the jsomers.net blog28 savers
- Did Claude 3 Opus align itself via gradient hacking? — LessWrong20 savers
- Russellcause.pdf1 savers
- Mechanical-Intelligence-Volume-1.pdf2 savers
- the trouble with making predictions as an agent1 savers
- 100 Papers to Inspire Wonder - Google Docs2 savers
- [2410.08020] Efficiently Learning at Test-Time: Active Fine-Tuning of LLMs1 savers
- Concrete Projects in AGI Preparedness1 savers
- Run a team of coding agents - Conductor1 savers
- kitty1 savers
- The Bitter Lesson for Software - by Uzay1 savers
- The Second Story Of Echo And Narcissus5 savers
- I would really like it if you had a personal website and I think it would make the world better11 savers
- let code die1 savers
- Designing Freedom3 savers
- Practically-A-Book Review: Byrnes on Trance1 savers
- Wu wei2 savers
- Curius / Onboarding2621 savers
- 101 things I would tell my self from 10 years ago28 savers
- A Website To End All Websites | Henry From Online16 savers
- The Idealists Collective11 savers
- Shipping at Inference-Speed | Peter Steinberger9 savers
- Seeing more whole - Joe Carlsmith7 savers
- Helena Tran3 savers
- Why should ethical anti-realists do ethics? - Joe Carlsmith3 savers
highlights — 21
One way I would market it is for it to pay dividends along the way.
⿻ Symbiogenesis vs. Convergent Consequentialism — LessWrongIf, at that point, we figure out that it is technically impossible to climb another ladder, then a strong, horizontally aligned group intelligence would stop there instead of committing suicide. Which is one of the good endings.
⿻ Symbiogenesis vs. Convergent Consequentialism — LessWrongthe sharp-left-turn generating dynamics need to be not just improbable, but information-theoretically impossible
⿻ Symbiogenesis vs. Convergent Consequentialism — LessWrongA lot of things seem to be converging
⿻ Symbiogenesis vs. Convergent Consequentialism — LessWrongBut if the core issue is domination by rent-seeking lords over enclosed commons, then the task becomes different: not to repair markets, but to contest enclosure; not to humanize extraction, but to reassert collective control over infrastructures and resources; not to defend an already exhausted ideal of capitalism, but to build renewed forms of collective, anti-feudal politics.
DEBT INC.: GUILT, CREDIT, AND THE ALGORITHMIC FUTUREThe central figure is no longer the capitalist entrepreneur competing in a market, but the lord who owns the gate, the channel, the platform, the territory, and who can therefore extract tribute from all who pass through.
DEBT INC.: GUILT, CREDIT, AND THE ALGORITHMIC FUTUREwe now have algorithmic capital as a new or additional form of credit system—one that involves a digital form of debt concerning our desires, emotions, and habits.
DEBT INC.: GUILT, CREDIT, AND THE ALGORITHMIC FUTUREForgiveness has a perverse way of entangling us even further in indebtedness
DEBT INC.: GUILT, CREDIT, AND THE ALGORITHMIC FUTUREA sin that is forgiven, in this logic, amplifies rather than abolishes our debt.
DEBT INC.: GUILT, CREDIT, AND THE ALGORITHMIC FUTURENote that this definition is a matter of degree: if we apply a high bar for internal consistency, then each subagent will be small (e.g. beliefs and desires about a single object) whereas a lower bar will lead to larger subagents (e.g. a whole ideology).
Towards a scale-free theory of intelligent agency'I had come here to shoot at "Fascists", but a man who is holding up his trousers is not a "Fascist", he is visibly a fellow-creature, similar to yourself, and you do not feel like shooting at him.
Eating Meat and Eating PeopleForecasts should be viewed as commitments / endorsements.
the trouble with making predictions as an agentWhatever you build, start with the model and a CLI first.
Shipping at Inference-Speed | Peter SteinbergerHelp people. Some of them may not remember, but you will feel good about doing it anyway. Some of them will, and it will create a strong bond between you.
101 things I would tell my self from 10 years agoNot desire, not need, love — to see them wholly, with gentleness and acceptance.
21 observations from people watching - by shaniPolite has a mechanical quality to it, like carrying out all the right movements to replace batteries in a remote. Happy has a boundless quality: unpredictable, even when it is at a low level. There is an openness, allowing another person to surprise and delight them. The easiest way to say this: there is no script for happy. It tumbles out of the body. Polite comes from the mind –it is restrained and calculated – measured lines and pauses.
21 observations from people watching - by shaniPhilosopher Alan Watts believed that wu wei can be described as "not-forcing."[81] Watts also understood wu wei as “the art of getting out of one’s own way” and offered the following illustration: “The river is not pushed from behind, nor is it pulled from ahead. It falls with gravity.”
Wu weiOr to put things another way: money-pump arguments try to shoe-horn “rationality mandates self-modifying to become an agent with a consistent utility function” into an implication of that more familiar slogan, “rationality is about winning.” But messing with your values (without the aim of maximizing for some other set of values) is actually very different from “winning.”
Why should ethical anti-realists do ethics? - Joe CarlsmithJohn and Helen both show alienation: there would seem to be an estrangement between their affections and their rational, deliberative selves; an abstract and universalizing point of view mediates their re- ponses to others and to their own sentiments
Alienation, Consequentialism, and the Demands of Morality - https://cla.csulb.edu/departments/philosophy/wp-content/uploads/2022/05/rdgrp_ethicsRailton3.pdfMy formula for greatness in a human being is amor fati: that one wants nothing to be different, not forward, not backward, not in all eternity. Not merely to bear what is necessary, still less conceal it ... but love it.
Eternal returnWhat if some day or night a demon were to steal after you into your loneliest loneliness, and say to you, "This life as you now live it and have lived it, you will have to live once more and innumerable times more; and there will be nothing new in it, but every pain and every joy and every thought and sigh and everything unutterably small or great in your life will have to return to you, all in the same succession and sequence" ... Would you not throw yourself down and gnash your teeth and curse the demon who spoke thus? Or have you once experienced a tremendous moment when you would have answ…
Eternal return