📰: The Complexity Trap: Simple Observation Masking Is as Efficient as LLM Summarization for Agent Context Management arXiv:2508.21433v1 https://moai.xrea.jp/dr-atom.html#68b57555189dc
【引用ツイート】
Miles Brundage
Yep we're on the same page, AI 2027 is indeed super unrealistic - the timeline are way off ᵇᵉᶜᵃᵘˢᵉ ᵗʰᶦⁿᵍˢ ʷᶦˡˡ ᵍᵒ ᶠᵃˢᵗᵉʳ ᵗʰᵃⁿ ᵗʰᵃᵗ
【引用ツイート】
VraserX e/acc
This is insane. A guy found out his therapist was literally pasting his words into ChatGPT and then parroting the answers back. By accident, he saw the therapist’s screen mid-session.
So people are paying $$$ an hour for someone to be a glorified ChatGPT wrapper.
Chubby @kimmonismus
Codex update: so we got two-codex-namings now?
OpenAI has greatly expanded Codex based on GPT-5. The model not only writes code, but also handles entire sessions with refactoring, testing, and bug fixes. New features include improved code reviews, a modernized CLI with diffs and to-do lists, deeper IDE integration (e.g., VS Code), and faster cloud environments.
2025年09月16日 02:09 https://pbs.twimg.com/media/G053PMBXgAA17a2.png