Keep on to blur preview images; turn off to show them clearly
RT @iwiwi: AtCoder Conference 2025 来ました！基調講演を担当します！前半ではこれまでの競プロの進歩を振り返りつつ、後半ではAIの性能進歩の主要な起爆剤の1つとなるという想像しなかった形で更に人類に貢献している競プロとAIの最近の関係の話をします…
Building @SakanaAILabs 🧠
RT @ChiekoTazuke: Sakana AIの秋葉さん含めたR&Dチームの研究成果であるAB-MCTSは、AppliedチームのApplied Research Engineerの力によって、MUFG様向けソリューションで活用されています。世界的に評価された技術を社会…
Building @SakanaAILabs 🧠
Philosophically, I think there is a reasonable stance that the point of science is to educate and uplift people. AI religion sounds silly; if someone said AI could achieve nirvana the claim would be dumb at best. Similarly, an "AI scientist" may miss the point of science. AI algorithms to facilitate new discoveries are here already and will only grow more capable. But the human pursuit of science is only uplifted as far as these tools are capable of teaching deeper insights to scientists
Founder and CEO @deepforestsci. Creator of @deep_chem. Author @OReillyMedia. @stanford CS PhD. https://t.co/7LDcegrCsc
That said, efficient tokenizers still use compound Chinese words and even major substrings (I remember Qwen's one having, like, 50 characters for some typical CPC duckspeak line)
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
I sometimes wonder how differently LLMs feel in Chinese. Maybe the decay in "soulfulness" that happened to DeepSeek post-R1 just doesn't exist on the other side of the Great Language Wall.
That said, efficient tokenizers still use compound Chinese words and even major substrings (I remember Qwen's one having, like, 50 characters for some typical CPC duckspeak line)
RT @sariazout: Always Go To The Funeral “Always go to the funeral” means that I have to do the right thing when I really, really don’t fe…
ai agents @hud_evals | owned @AIHubCentral (1 million users,acq.) ex climate protester🦦 dont do the deferred life plan