Daniel Paleka's Newsletter
Subscribe
Sign in
Home
Archive
About
You are going to get priced out of the best AI coding tools
The best AI tools will become far more expensive. Andy Warhol famously said:
Nov 5, 2025
•
Daniel Paleka
15
7
GPT-4o draws itself as a consistent type of guy
When asked to draw itself as a person, the ChatGPT Create Image feature introduced on March 25, 2025, consistently portrays itself as a white male in…
Mar 31, 2025
•
Daniel Paleka
23
12
2
Writing in public is still underrated
If you have ideas but never write them down, this post is for you.
Dec 1, 2025
•
Daniel Paleka
39
2
1
Most Popular
View all
GPT-4o draws itself as a consistent type of guy
Mar 31, 2025
•
Daniel Paleka
23
12
2
Writing in public is still underrated
Dec 1, 2025
•
Daniel Paleka
39
2
1
Contra Dwarkesh on RL sample-efficiency via information theory
Nov 18, 2025
•
Daniel Paleka
33
4
1
You should delay engineering-heavy research in light of R&D automation
Jan 6, 2025
•
Daniel Paleka
14
1
The two types of LLM preferences
Nov 10, 2025
•
Daniel Paleka
18
A/B testing could lead LLMs to retain users instead of helping them
Nov 2, 2025
•
Daniel Paleka
11
8
Latest
Top
Discussions
Underrated reasons to be thankful: AI research edition
I experienced my first Thanksgiving and thus had to write this.
Nov 28, 2025
•
Daniel Paleka
16
Even superhuman AI forecasters are only as good as your questions
Predicting the future is difficult. It might become easier. What then?
Nov 27, 2025
•
Daniel Paleka
6
1
What’s the deal with RL and forecasting?
Prediction is difficult, especially about the future.
Nov 20, 2025
•
Daniel Paleka
5
1
Contra Dwarkesh on RL sample-efficiency via information theory
Supervised learning teaches the model more bits/sample than RL; but it's not the right way to think about it.
Nov 18, 2025
•
Daniel Paleka
33
4
1
16 questions about reinforcement learning
AlphaProof constructed a large curriculum of tasks by randomly misformalizing math olympiad problems; creating many problems of varying difficulty, from…
Nov 18, 2025
•
Daniel Paleka
5
2
Obvious ways RL can fail
Reinforcement learning works for some things and not for others. Why?
Nov 15, 2025
•
Daniel Paleka
2
The two types of LLM preferences
The standard approach to measure values or preferences of LLMs is to:
Nov 10, 2025
•
Daniel Paleka
18
See all
Daniel Paleka's Newsletter
AI research and making the future go well.
Subscribe
Recommendations
Experience Machines
Robert Long
Random Features
Daniel Paleka
Machinocene
Kevin Kohler
Sentinel Global Risks Watch
Nuño Sempere
Daniel Paleka's Newsletter
Subscribe
About
Archive
Recommendations
Sitemap
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
