I live in New York, work in finance, drink excessive amounts of coffee, and play chess.
- Merge and Conquer: Evolutionarily Optimizing AI for 2048
by xianshou on 10/27/25, 1:12 AM, with comments
- Stuck in the Matrix: Probing Spatial Reasoning in Large Language Models
by xianshou on 10/27/25, 1:11 AM, with comments
- Reflection AI Raises $2B to Build "American DeepSeek"
by xianshou on 10/9/25, 1:39 PM, with comments
- Nvidia-backed Reflection AI raising at $5.5B valuation
by xianshou on 10/8/25, 9:16 PM, with comments
- Unsupervised Elicitation of Language Models
by xianshou on 6/13/25, 9:29 PM, with comments
- DeepSeek V3 0324 is now the best nonthinking model (Reddit)
by xianshou on 3/26/25, 10:48 PM, with comments
- DeepSeek V3 0324 outpaces GPT 4.5 and Claude 3.7 in coding, other benchmarks
by xianshou on 3/26/25, 10:46 PM, with comments
- Practical RL (Yandex Data School)
by xianshou on 2/20/25, 4:56 PM, with comments
- InvestorBench: A Benchmark for Financial Decision-Making Tasks with Agents
by xianshou on 1/3/25, 12:56 AM, with comments
- An Evolved Universal Transformer Memory (Sakana.ai)
by xianshou on 12/17/24, 1:08 AM, with comments