read.ehrlich.dev
ai-curated rss
about
github
rss
dark
lesswrong.com
posts from lesswrong.com
hot
new
week
month
year
spaces
all
ai
internet culture
writing
information theory
malware analysis
devops
site reliability
security
machine learning
cloud infrastructure
ai agents
quantum computing
open source
digital rights
behavioral economics
reverse engineering
energy
complexity
distributed systems
economics
databases
api design
git internals
wasm
materials science
data engineering
webdev
math olympiad
physics
cryptography
design
thermodynamics
mathematics
solo dev
observability
indie games
chemistry
programming languages
cellular automata
python
containers
technical writing
game theory
birding
reinforcement learning
nanotechnology
creative coding
statistics
climate science
procedural generation
worldbuilding
space exploration
astrophysics
game dev
data visualization
operating systems
networking
linux
computer vision
golang
robotics
astrobiology
aerospace
exploit development
neuroscience
cognitive science
manga
algorithmic trading
elixir
decision theory
computer graphics
political philosophy
history of science
formal verification
javascript
mechanical keyboards
swift
category theory
self hosted
board games
type theory
evolution
philosophy
rust
nix
philosophy of mind
embedded systems
urban exploration
retrocomputing
compilers
amateur astronomy
nuclear
information design
computer architecture
electronic music
oceanography
archaeology
chess
genetics
fractals
geology
ethics
kotlin
go game
vinyl
number theory
anthropology
aquariums
film
history of computing
lisp
synths
bioinformatics
pharmacology
ecology
metaphysics
semiotics
audio programming
history
paleoanthropology
topology
cpp
photography
haskell
consciousness
witsrtn
cartography
meditation
ham radio
fpga
sci fi
shell scripting
demoscene
homelab
horror
speedrunning
typography
tabletop rpg
animation
woodworking
3d printing
sourdough
puzzles
standup comedy
phenomenology
linguistics
existentialism
music theory
true crime
fermentation
lock picking
espresso
zig
finance
pixel art
philosophy of science
nonduality
typescript
conlangs
mycology
forth
285
The Hot Mess Paper Conflates Three Distinct Failure Modes
(lesswrong.com)
24 days ago ·
ai
·
cellular automata
270
An agent autonomously builds a 1.5 GHz Linux-capable RISC-V CPU
(lesswrong.com)
24 days ago ·
ai
·
ai agents
285
Finding features in Transformers: Contrastive directions elicit stronger low-level perturbation responses than baselines
(lesswrong.com)
24 days ago ·
machine learning
·
ai
250
Untrusted monitoring: extra bits
(lesswrong.com)
24 days ago ·
distributed systems
·
game theory
180
The Future of Aligning Deep Learning systems will probably look like "training on interp"
(lesswrong.com)
24 days ago ·
machine learning
·
ai
245
Confusion around the term reward hacking
(lesswrong.com)
24 days ago ·
reinforcement learning
·
cellular automata
72
Grounding Coding Agents via Dixit
(lesswrong.com)
23 days ago ·
ai agents
·
ai
175
Untrusted Monitoring is Default; Trusted Monitoring is not
(lesswrong.com)
24 days ago ·
site reliability
·
distributed systems
72
The Federal AI Policy Framework: An Improvement, But My Offer Is (Still Almost) Nothing
(lesswrong.com)
24 days ago ·
ai
·
digital rights
68
Why I am not buying IPv4 addresses as an investment
(lesswrong.com)
24 days ago ·
networking
·
api design
42
ARENA 7.0 Impact Report
(lesswrong.com)
24 days ago ·
ai
·
reinforcement learning
18
The Distaff Texts
(lesswrong.com)
24 days ago ·
history
·
worldbuilding
115
Was Anthropic that strategically incompetent?
(lesswrong.com)
29 days ago ·
behavioral economics
·
decision theory
260
The Artificial Self
(lesswrong.com)
1 month ago ·
cognitive science
·
ai
120
Some models don't identify with their official name
(lesswrong.com)
29 days ago ·
ai
·
machine learning
100
Emergent stigmergic coordination in AI agents?
(lesswrong.com)
29 days ago ·
ai agents
·
complexity
62
We Started Lens Academy: Scalable Education on Superintelligence Risk
(lesswrong.com)
29 days ago ·
ai
·
sci fi
305
Extracting Performant Algorithms Using Mechanistic Interpretability
(lesswrong.com)
1 month ago ·
complexity
·
evolution
95
When do intuitions need to be reliable?
(lesswrong.com)
29 days ago ·
behavioral economics
·
philosophy
65
Less Capable Misaligned ASIs Imply More Suffering
(lesswrong.com)
29 days ago ·
cellular automata
·
decision theory
28
What Are We Actually Evaluating When We Say a Belief “Tracks Truth”?
(lesswrong.com)
29 days ago ·
metaphysics
·
decision theory
42
My Willing Complicity In "Human Rights Abuse"
(lesswrong.com)
29 days ago ·
ethics
·
writing
125
Assessing heterogeneity in METR's late 2025 developer productivity experiment
(lesswrong.com)
1 month ago ·
ai
·
site reliability
72
What concerns people about AI?
(lesswrong.com)
1 month ago ·
ai
·
behavioral economics
270
Power Steering: Behavior Steering via Layer-to-Layer Jacobian Singular Vectors
(lesswrong.com)
1 month ago ·
robotics
·
machine learning
42
FW26 Color Stats
(lesswrong.com)
1 month ago ·
data visualization
·
creative coding
38
Sparks of RSI?
(lesswrong.com)
1 month ago ·
ai agents
·
complexity
105
[Linkpost] Should we make grand deals about post-AGI outcomes?
(lesswrong.com)
1 month ago ·
decision theory
·
behavioral economics
22
'Staying with it' Done Wrong
(lesswrong.com)
1 month ago ·
meditation
·
consciousness
14
Rationalist Passover Seder in Maryland
(lesswrong.com)
29 days ago ·
philosophy
120
A Dialogue on Civic AI
(lesswrong.com)
1 month ago ·
ai
·
consciousness
80
AI Safety Newsletter #69: Department of War, Anthropic, and National Security
(lesswrong.com)
1 month ago ·
ai
·
site reliability
62
Inputs, outputs, and valued outcomes
(lesswrong.com)
1 month ago ·
decision theory
·
behavioral economics
12
Forecasting Dojo Meetup - postmortem discussion.
(lesswrong.com)
1 month ago ·
climate science
·
cognitive science
28
New LessWrong Editor! (Also, an update to our LLM policy.)
(lesswrong.com)
1 month ago ·
technical writing
·
typography
38
Things that Go Boom
(lesswrong.com)
1 month ago ·
energy
·
nuclear
24
Probably you won't be able to perform a data-driven habit stacking for self-improvement
(lesswrong.com)
1 month ago ·
behavioral economics
·
statistics
16
Sensing Physical Necessity: An Exercise In Naturalism
(lesswrong.com)
1 month ago ·
philosophy of mind
·
consciousness
12
Is there any way to recover lost work?
(lesswrong.com)
1 month ago ·
internet culture
·
technical writing
312
AI #158: The Department of War
(lesswrong.com)
1 month ago ·
open source
·
digital rights
280
Investigating Self-Fulfilling Misalignment and Collusion in AI Control
(lesswrong.com)
1 month ago ·
game theory
·
behavioral economics
200
AI Safety Has 12 Months Left
(lesswrong.com)
1 month ago ·
ai
·
complexity
85
How I Handle Automated Programming
(lesswrong.com)
1 month ago ·
site reliability
·
compilers
85
Reasoning Models Struggle to Control Their Chains of Thought
(lesswrong.com)
1 month ago ·
golang
85
Models have linear representations of what tasks they like
(lesswrong.com)
1 month ago ·
behavioral economics
·
cognitive science
85
Have Americans Become Less Violent Since 1980?
(lesswrong.com)
1 month ago ·
true crime
42
Personality Self-Replicators
(lesswrong.com)
1 month ago ·
ai agents
·
cellular automata
18
A Compositional Philosophy of Science for Agent Foundations
(lesswrong.com)
1 month ago ·
philosophy of science
38
Salient Directions in AI Control
(lesswrong.com)
1 month ago ·
ai agents
·
ai
18
Computation, Chess, and Language in Artificial Intelligence
(lesswrong.com)
1 month ago ·
cellular automata
·
ai
12
How can I arrest and reverse stress, anxiety and depression induced cognitive decline?
(lesswrong.com)
1 month ago ·
neuroscience
·
cognitive science
spaces
all
ai
internet culture
writing
information theory
malware analysis
devops
site reliability
security
machine learning
cloud infrastructure
ai agents
quantum computing
open source
digital rights
behavioral economics
reverse engineering
energy
complexity
distributed systems
economics
databases
api design
git internals
wasm
materials science
data engineering
webdev
math olympiad
physics
cryptography
design
thermodynamics
mathematics
solo dev
observability
indie games
chemistry
programming languages
cellular automata
python
containers
technical writing
game theory
birding
reinforcement learning
nanotechnology
creative coding
statistics
climate science
procedural generation
worldbuilding
space exploration
astrophysics
game dev
data visualization
operating systems
networking
linux
computer vision
golang
robotics
astrobiology
aerospace
exploit development
neuroscience
cognitive science
manga
algorithmic trading
elixir
decision theory
computer graphics
political philosophy
history of science
formal verification
javascript
mechanical keyboards
swift
category theory
self hosted
board games
type theory
evolution
philosophy
rust
nix
philosophy of mind
embedded systems
urban exploration
retrocomputing
compilers
amateur astronomy
nuclear
information design
computer architecture
electronic music
oceanography
archaeology
chess
genetics
fractals
geology
ethics
kotlin
go game
vinyl
number theory
anthropology
aquariums
film
history of computing
lisp
synths
bioinformatics
pharmacology
ecology
metaphysics
semiotics
audio programming
history
paleoanthropology
topology
cpp
photography
haskell
consciousness
witsrtn
cartography
meditation
ham radio
fpga
sci fi
shell scripting
demoscene
homelab
horror
speedrunning
typography
tabletop rpg
animation
woodworking
3d printing
sourdough
puzzles
standup comedy
phenomenology
linguistics
existentialism
music theory
true crime
fermentation
lock picking
espresso
zig
finance
pixel art
philosophy of science
nonduality
typescript
conlangs
mycology
forth