文章
文章
これらのノートが書ける部分。研究は同じ流れのソース。ツールは残り —— 自分で歩かないといけない。
スタジオ出力
ノートブック級のデスク
スタジオ出力は jubit.ai/terminal の右サイドバー。PPT、PDF、MP4、YouTube —— それに NotebookLM の Video Overview、Slide Deck、Infographic、Audio。そこで生成し、ダウンロードする。この庭は偽ファイルを置かない。
出力
PPT
jubit ターミナルのスタジオを開く。そこで生成またはダウンロード。
スタジオを開く ↗出力
PDF
jubit ターミナルのスタジオを開く。そこで生成またはダウンロード。
スタジオを開く ↗出力
MP4
jubit ターミナルのスタジオを開く。そこで生成またはダウンロード。
スタジオを開く ↗出力
YouTube
jubit ターミナルのスタジオを開く。そこで生成またはダウンロード。
スタジオを開く ↗NotebookLM
Video Overview
jubit ターミナルのスタジオを開く。そこで生成またはダウンロード。
スタジオを開く ↗NotebookLM
Slide Deck
jubit ターミナルのスタジオを開く。そこで生成またはダウンロード。
スタジオを開く ↗NotebookLM
Infographic
jubit ターミナルのスタジオを開く。そこで生成またはダウンロード。
スタジオを開く ↗NotebookLM
Audio
jubit ターミナルのスタジオを開く。そこで生成またはダウンロード。
スタジオを開く ↗
ローカルドロップ
この端末に .ppt / .pptx / .pdf / .mp4 を並べるか、YouTube URL を貼る。アップロードしない。キットをパックして公開。
NotebookLM 級のソース
ターミナルのリサーチタブ / featured フィードからピン。dseek.ai/terminal?tab=research で開く。
Semiconductor hegemony 2029: ASML, TSMC and the Terafab pivot
Three dated protections — memory price floors, TSMC's Low-NA runway, and ASML's uncontested light source — all approach expiry around 2029, the same year the Terafab/xLight challenger stack targets maturity. Until then incumbents keep pricing power; the trade is watching whether the challengers' milestones hold, because each one is dated and checkable while the narrative is neither.
AI software chain — modality economics
The pricing unit decides fate (seat loses, usage gains, outcome wins), modality decides unit economics (text deflation does not rescue per-second video), and in generative modalities legal usability now outprices quality.
AI hardware chain: the constraint migrates, and the money follows
Excess profit in the AI hardware chain always piles up on whichever pipe is narrowest right now — and that pipe moves every 6 to 12 months, which in 2026 puts it at grid power rather than anything made of silicon.
HK luxury MPV — the moat moved
The Alphard’s moat was Hong Kong’s tax-and-scarcity structure, not the vehicle — and the binding constraint has migrated from acquisition tax to charging capacity and residual-value discovery.
Example: dseek.ai/terminal?tab=research&symbol=00700
ノートを読み込み中…