Scott Alexander shows how he won his 2022 bet about AI image generation capabilities, tracking the progress from early failures to complete success in 2025, using this to argue against AI skeptics. Longer summary

Scott Alexander describes the resolution of a bet he made in June 2022 about AI image generation capabilities. The bet claimed that by June 2025, AI would master image compositionality and be able to accurately generate specific complex scenes. The post shows the progression of AI image generation from 2022 to 2025, starting with early failures by DALL-E2, through various partial successes with Google Imagen and DALL-E3, and ending with ChatGPT 4o's complete success in May-June 2025. Scott uses this to argue against critics who claimed AI was just a 'stochastic parrot' that couldn't achieve true understanding, though he acknowledges some remaining limitations with very complex prompts. Shorter summary

Scott tests OpenAI's o3 model's ability to identify locations from photos, finding it has remarkable success even with minimal visual information, raising questions about AI capabilities. Longer summary

Scott tests OpenAI's o3 model on increasingly difficult GeoGuessr-style location guessing challenges using his own photos. Starting with a Google Street View image of a featureless plain, progressing through personal photos of Nepal mountains, a dorm room, and extremely zoomed-in shots of grass and river water, Scott finds that o3 shows remarkable ability to identify locations from minimal visual cues. While it fails on some challenges like identifying a specific house address, its success rate and reasoning process on most images is impressive enough to make Scott question whether this represents a qualitatively different level of AI capability. Shorter summary

Scott investigates the correlation between intelligence and neuron count, exploring various theories before suggesting that more neurons allow for less polysemantic (overlapping) representations of concepts. Longer summary

Scott explores why intelligence correlates with neuron count across species, humans, and AI models, despite this correlation not being immediately intuitive. He examines various hypotheses about how having more neurons could help with complex pattern-matching tasks like IQ tests. After discussing several possibilities including pattern storage and matching, he settles on polysemanticity as a potential explanation: fewer neurons means each neuron must encode multiple concepts, reducing precision. The post concludes with insights from an expert suggesting that larger neural networks (biological or artificial) can better approximate complex functions and maintain multiple hypotheses simultaneously. Shorter summary

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. Longer summary

Scott Alexander describes a chess game he played against GPT-2, an AI language model not designed for chess. Despite neither player performing well, GPT-2 managed to play a decent game without any understanding of chess or spatial concepts. The post then discusses the work of Gwern Branwen and Shawn Presser in training GPT-2 to play chess, showing its ability to learn opening theory and play reasonably well for several moves. Scott reflects on the implications of an AI designed for text prediction being able to perform tasks like writing poetry, composing music, and playing chess without being specifically designed for them. Shorter summary

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. Longer summary

This post discusses GPT-2, a language model AI, and its implications for artificial general intelligence (AGI). Scott Alexander argues that while GPT-2 is not AGI, it demonstrates unexpected capabilities that arise from its training in language prediction. He compares GPT-2's learning process to human creativity and understanding, suggesting that both rely on pattern recognition and recombination of existing information. The post explores examples of GPT-2's abilities, such as rudimentary counting, acronym creation, and translation, which were not explicitly programmed. Alexander concludes that while GPT-2 is far from true AGI, it shows that AI can develop unexpected capabilities, challenging the notion that AGI is impossible or unrelated to current AI work. Shorter summary

Scott Alexander hypothesizes that mystical experiences, hallucinations, and paranoia might be linked to an overactive pattern-matching faculty in the brain. Longer summary

Scott Alexander explores the relationship between mysticism, pattern-matching, and mental health. He suggests that hallucinations, paranoia, and mystical experiences might all be related to an overactive pattern-matching faculty in the brain. The post begins by discussing how the brain's failure modes differ from computers, then explains top-down processing and pattern matching using visual examples. It then connects these concepts to hallucinations, paranoia, and mystical experiences. Scott proposes that certain practices like meditation, drug use, and religious rituals may strengthen the pattern-matching faculty, leading to experiences of universal connectedness or enlightenment. He acknowledges that this hypothesis doesn't explain all aspects of mystical experiences and their benefits. Shorter summary

Scott Alexander examines the process of 'crystallizing patterns' in thinking, discussing its benefits and potential pitfalls across various domains. Longer summary

Scott Alexander explores the concept of 'crystallizing patterns' in thinking, using examples from the Less Wrong sequences and C.S. Lewis's writings. He discusses how naming and defining patterns can make them easier to recognize and think about, potentially changing how people view certain issues. The post examines whether this process can ever be wrong or counterproductive, concluding that while it can sometimes be misleading, it's generally beneficial if done carefully. Scott uses various examples to illustrate his points, including political correctness, mainstream media, and religious concepts. Shorter summary