OpenAIのチーフサイエンティスト、実験にはAIを信頼するが複雑なシステム設計のレベルには達していないと指摘
本文の状態
日本語全文を表示中
詳細モードで約1分の本文を読めます。
同じ出来事の情報源
6媒体で確認
The Decoder · LY Corp Tech Blog · TechCrunch AI · Ars Technica AI · The Verge AI · TLDR AI
各社の報じ方を比較 ↓OpenAIのチーフサイエンティストJakub Pachockiは、かつて手作業でコードを書いていた実験をAIが1週間で処理できるようになったと認める一方、複雑なシステム設計はAIに任せられないと述べている。
Source Article
元記事を日本語で読む
本文に関係しない購読案内、埋め込み通知、サイト内プロモーションは除いています。

OpenAIの主任科学者であるヤクブ・パチョツキは、かつてはすべてのコードを手書きで書いていた。現在では、かつて1週間を要していた実験をAIが処理するが、彼はAIに主導権を渡すつもりはないという。
本記事「OpenAI主任科学者、実験はAIに任せるが複雑なシステム設計はまだ無理と語る」は、The Decoderに最初に掲載されました。
原文を表示
OpenAI Chief Scientist Jakub Pachocki used to write every line of code by hand. Now AI handles experiments that once took him a week, but he's not ready to let it run the show.
Just a year ago, Pachocki didn't even use autocomplete. He preferred working in the code text editor Vim, typing everything out himself. "I'm very pedantic about my code," he tells MIT Technology Review.
The latest models won him over. Experiments that used to eat up a full week of coding now get done by AI over a single weekend. "That's hard to argue with," he says. Still, Pachocki draws a line at complex design work. "I don't think it's at the level where I would just let it take the reins and design the whole thing."
He's also quick to point out that AI tools aren't a silver bullet for every developer; how useful they are depends on the person and the task. His goal now is to keep pushing the problem-solving capabilities of tools like Codex and bring them into new scientific fields.
OpenAI's path toward fully autonomous AI research
Pachocki's main focus right now is the autonomous AI research intern OpenAI announced last fall, "a system that you can delegate tasks to that would take a person a few days." It's set to ship in September. By March 2028, the company plans to have a full "AI Researcher": a multi-agent system that can independently tackle complex problems across math, physics, biology, chemistry, and even economics and politics.
"Of course, you still want people in charge and setting the goals. But I think we will get to a point where you kind of have a whole research lab in a data center," Pachocki says. He sees this going well beyond research. Eventually, these data centers could take over the work of entire corporations like OpenAI or Google. Work that once required massive human organizations would be done by a handful of people. That kind of shift creates "extremely concentrated power that's in some ways unprecedented," Pachocki says.
AI News Without the Hype – Curated by Humans
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.
Subscribe now
同じ出来事を6媒体で確認
同じ出来事を扱う別媒体の記事です。見出しと公開時刻を比較できます。
今日のまとめ
AIデイリーブリーフで今日の重要ニュースをまとめ読み