Import AI jack-clark.net

Import AI 416: CyberGym; AI governance and AI evaluation; Harvard releases ~250bn tokens of text

jack-clark.net·47w

Import AI 415: Situational awareness for AI systems; 8TB of open text; and China’s heterogeneous compute cluster

jack-clark.net·48w

Import AI 414: Superpersuasion; OpenAI models avoid shutdown; weather prediction and AI

jack-clark.net·50w

Import AI 413: 40B distributed training run; avoiding the ‘One True Answer’ fallacy of AI safety; Google releases a content classification model

jack-clark.net·51w

Import AI 412: Amazon’s sorting robot; Huawei trains an MoE model on 6k Ascend chips; and how third-party compliance can help with AI safety

jack-clark.net·52w

Import AI 411: Scaling laws for AI oversight; Google’s cyber threshold; AI scientists

jack-clark.net·53w

Import AI 410: Eschatological AI Policy; Virology weapon test; $50m for distributed training

jack-clark.net·54w

Import AI 409: Huawei trains a model on 8,000+ Ascend chips; 32B decentralized training run; and the era of experience and superintelligence

jack-clark.net·55w

Import AI 408: Multi-code SWE Bench; backdoored Unitree robots; and what AI 2027 is telling us

jack-clark.net·56w

Import AI 407: DeepMind sees AGI by 2030; MouseGPT; and ByteDance’s inference cluster

jack-clark.net·57w

Import AI 406: AI-driven software explosion; robot hands are still bad; better LLMs via pdb

jack-clark.net·58w

Import AI 405: What if the timelines are correct?

jack-clark.net·59w

Import AI 404: Scaling laws for distributed training; misalignment predictions made real; and Alibaba’s good translation model

jack-clark.net·60w

Import AI 403: Factorio AI; Russia’s reasoning drones; biocomputing

jack-clark.net·61w

Import AI 402: Why NVIDIA beats AMD: vending machines vs superintelligence; harder BIG-Bench

jack-clark.net·62w

Import AI 401: Cheating reasoning models; better CUDA kernels via AI; life models

jack-clark.net·63w

Import AI 400: Distillation scaling laws; recursive GPU kernel improvement; and wafer-scale computation

jack-clark.net·64w

Import AI 399: 1,000 samples to make a reasoning model; DeepSeek proliferation; Apple’s self-driving car simulator

jack-clark.net·65w

Import AI 398: DeepMind makes distributed training better; AI versus the Intelligence Community; and another Chinese reasoning model

jack-clark.net·66w

Import AI 397: DeepSeek means AI proliferation is guaranteed; maritime wardrones; and more evidence of LLM capability overhangs

jack-clark.net·67w

Log in to enable infinite scrolling