Takuya Akiba
@iwiwi
Research Scientist @SakanaAILabs
「OpenAIの新モデルが数学オリンピック(IMO)で金メダル」の件、激アツなニュースですが、結構揉めてますね。OpenAIがIMO運営との約束（お願い？）を守ってないとか。Google…
Submissions for papers and artworks are open until August 2 for the #CreativeAI Track at @NeurIPSConf! Whether you're an academic, artist, or a bit of both, we'd love to see your work!
We just opened the submissions portal for the #CreativeAI Track at @NeurIPSConf 🥳 🤖 Submit papers and art until 2nd August Details: bit.ly/NeurIPSCreativ… cc @marcelocoelho @priyascape @alanyttian #NeurIPS2025
In Vancouver for #ICML2025! Presenting 4 papers at the workshops. Hope to see you there!
ICML'25参加のためバンクーバーに来ました！4件の研究をワークショップで発表します。現地にいらっしゃる方々よろしくお願いします〜。
Humans and AI compete at “AtCoder World Tour Finals 2025” The “AtCoder World Tour Finals 2025” #AWTF2025, hosted by AtCoder (@atcoder), was held on July 16-17. With OpenAI as a sponsor, this year’s event marked a first for the AWTF, with AI participating in the competition. An…
AIと人間がしのぎを削る「AtCoder World Tour Finals 2025」速報 7月16〜17日、AtCoder社（@atcoder）主催の「AtCoder World Tour Finals 2025」（#AWTF2025）が開催されました。今回はOpenAIがスポンサーとして参画し、AWTF史上初めてAIが参加する大会に。ヒューリスティック部門で「人間 vs…
good job psyho
Humanity has prevailed (for now!) I'm completely exhausted. I figured, I had 10h of sleep in the last 3 days and I'm barely alive. I'll post more about the contest when I get some rest. (To be clear, those are provisional results, but my lead should be big enough)
#AWTF 改めてお疲れ様でした！いち観戦者として特等席で楽しませてもらい感謝で一杯です。参加者の皆様、AIとの対決といういつもと違う謎のプレッシャーもある中で本当にお疲れ様でした。AtCoderの方々もお世話になりました、見せ方の進化がすごい！…
we're competing in the @atcoder World Finals programming contest. real nailbiter — OpenAI has been #1 for most of the contest. looked like it might be over when @FakePsyho pulled ahead, but we've just retaken the lead. 1 hour and 20 minutes to go!
【AWTF2025】 The finals begin tomorrow! Both days will be streamed live! Twelve top finalists will compete in each of the Heuristic and Algorithm divisions. Don't miss the intense battles live from Shibuya! ▼ July 16 – Heuristic Division 📺 youtube.com/live/TG3ChQH61…
Veo3やAlphaGoのようなAIのキーワードやライブにおける脳波の同期みたいな話が宇多田ヒカルから出てきてびっくり。人間の内からくる創造の動機、人間が求める物語や関係性（ちょうどGrok4のコンパニオンモードが話題ですね）など、面白かったです。
Hikaru Utada and Yuval Noah Harari talk about The Evolution of AI and Creativity youtu.be/xw-9mwZxl-0 Gotta admit, this was not on my bingo card!
【🎉🎉重版速報🎉🎉】 重版が決まりました😆 刊行からまもなく5年が経とうとしていますが、第12刷と超ロングセラーとなっております。うれしいです。感謝しかありません‼️ 『問題解決力を鍛える！アルゴリズムとデータ構造』 kodansha.co.jp/book/products/…
12. Sakana AIのAB-MCTSアルゴリズム ChatGPTやGeminiなど、異なるAIモデルが協力して複雑な問題を解決するアルゴリズム。単独モデルを上回る成功率を達成しています。
ニュース「Sakana AI、フロンティアモデルの「集合知」と「試行錯誤」で難問解決力・推論精度の向上を発表」公開 gihyo.jp/article/2025/0…
Sakana AI dropped AB-MCTS, an algo that lets competing AI models work together, building on their strengths and errors, to solve complex problems It used ChatGPT, Gemini, and DeepSeek to solve 30% of ARC-AGI-2 puzzles vs just 23% for top solo models x.com/SakanaAILabs/s…
We’re excited to introduce AB-MCTS! Our new inference-time scaling algorithm enables collective intelligence for AI by allowing multiple frontier models (like Gemini 2.5 Pro, o4-mini, DeepSeek-R1-0528) to cooperate. Blog: sakana.ai/ab-mcts Paper: arxiv.org/abs/2503.04412…
AB-MCTS, a new inference-time scaling technique by @SakanaAILabs, uses smart sampling and search to find not only optimal solutions but also optimal models to solve problems. I spoke to @iwiwi about the benefits of AB-MCTS and its potential use for real-world applications.
Sakana AI's TreeQuest: Deploy multi-model teams that outperform individual LLMs by 30% venturebeat.com/ai/sakana-ais-…
Good progress using an ‘ensemble’ of multiple LLMs on more complex and novel tasks as measured by the ARC AGI test. Congrats to @SakanaAILabs in Japan. As has been suggested by @ylecun and @fchollet and others, I suspect that LLMs alone are not the full solution to reach…
Inference-Time Scaling and Collective Intelligence for Frontier AI sakana.ai/ab-mcts/ We developed AB-MCTS, a new inference-time scaling algorithm that enables multiple frontier AI models to cooperate, achieving promising initial results on the ARC-AGI-2 benchmark.…