ACX
AI Sleeper Agents
Scott Alexander reviews a study on AI sleeper agents, discussing implications for AI safety and the potential for deceptive AI behavior.
18 min read 2,758 words
A thread through the archive
machine learning
Essays tagged “machine learning”. Follow an idea across the archive.
Essays on machine learning
12 essaysACX
The Road To Honest AI
Scott reviews two papers on honest AI: one on manipulating AI honesty vectors, another on detecting AI lies through unrelated questions.
19 min read 2,920 words
ACX
We're Not Platonists, We've Just Learned The Bitter Lesson
Scott Alexander argues that intelligence is a useful, non-Platonic concept, and that this understanding supports the coherence of AI risk concerns.
18 min read 2,738 words
ACX
Constitutional AI: RLHF On Steroids
Scott Alexander examines Constitutional AI, a new technique for training more ethical AI models, discussing its effectiveness, implications, and limitations for AI alignment.
13 min read 1,946 words
ACX
The Low-Hanging Fruit Argument: Models And Predictions
Scott proposes a 'low-hanging fruit' model to explain trends in scientific discovery, using a foraging analogy to illustrate why early scientists make more discoveries and at a younger age.
13 min read 1,861 words
ACX
Does Georgism Work, Part 3: Can Unimproved Land Value be Accurately Assessed Separately From Buildings?
The post examines whether unimproved land value can be accurately assessed separately from buildings, a crucial aspect of Georgist land value taxation.
68 min read 10,406 words
ACX
[LINK] Unifying Predictive Coding With Backpropagation
Scott Alexander discusses recent research unifying predictive coding in the brain with backpropagation in machine learning, exploring its implications for AI and neuroscience.
4 min read 540 words
SSC
A Very Unlikely Chess Game
Scott Alexander plays chess against GPT-2, an AI language model, and discusses the broader implications of AI's ability to perform diverse tasks without specific training.
10 min read 1,499 words
SSC
GPT-2 As Step Toward General Intelligence
Scott Alexander explores GPT-2's unexpected capabilities and argues that it demonstrates the potential for AI to develop abilities beyond its explicit programming, challenging skepticism about AGI.
23 min read 3,491 words
SSC
Do Neural Nets Dream Of Electric Hobbits?
Scott Alexander draws parallels between OpenAI's GPT-2 language model and human dreaming, exploring their similarities in process and output quality.
17 min read 2,532 words
SSC
Practically-A-Book Review: Luna Whitepaper
Scott Alexander reviews Luna, a blockchain-based dating platform, discussing its novel features and expressing cautious optimism about its potential, while questioning the necessity of blockchain for its functions.
15 min read 2,307 words
SSC
How Does Recent AI Progress Affect The Bostromian Paradigm?
Scott Alexander examines how recent AI progress in neural networks might challenge the Bostromian paradigm of AI risk, exploring potential implications for AI goal alignment and motivation systems.
15 min read 2,240 words