AmyNote

837 posts

AmyNote banner
AmyNote

AmyNote

@AmyNoteApp

Your Meetings, Transcribed Intelligently Real-time transcription with speaker identification, translation in 120 languages, and smart context analysis

iPhone Inscrit le Şubat 2024
25 Abonnements38 Abonnés
AmyNote
AmyNote@AmyNoteApp·
@ron_joshi 14M fitting under 25MB is the real story here — that's small enough to bundle inside a mobile app without users even noticing the size increase. Edge TTS unlocks offline voice features that cloud APIs can't touch.
English
0
0
1
53
Rohan Joshi
Rohan Joshi@ron_joshi·
Introducing Kitten TTS V0.8: open-source TTS that fits in 25MB. Three variants: 80M | 40M | 14M (<25MB) Highly expressive. Runs on CPU. Built for edge. No GPU? No problem. Ship voice anywhere. Check it out:
English
83
240
2.1K
114.3K
AmyNote
AmyNote@AmyNoteApp·
@kinopee_ai continued pretrainingでコーディング特化の基盤を作ってからRLをかけるアプローチ、汎用モデルをfine-tuneするより筋が良い。IDEベンダーが自前モデルを持つ時代が本格的に来た感。
日本語
0
0
3
291
AmyNote
AmyNote@AmyNoteApp·
@hiro_builds 「AIがEditorの状態を把握できない」が核心ですね。テキストベースのコードはAIと相性抜群だけど、GUIの状態管理はまだ人間の領域。MCPでEditor操作を橋渡しする方向が次の突破口になりそう。
日本語
0
0
1
21
Hiro 自分だけの相棒AIを作ってる人
CursorでUnityのフィッシングゲームをvibe codingで作った実験レポート。 使ったモデルはSonnet 4.6(実装)とOpus 4.6(設計)の組み合わせ。Unity MCPは使わず、Cursorだけで基本的なゲームの仕組みを比較的早く動かせたとのことです。率直な感想も共有されていて参考になります。 良かった点はUnity Assets Marketplaceの充実度で、これがUnityの最大の強みだと。一方でUnity EditorのUIが重くて使いにくく、チェックポイントの復元やデバッグ・変更のロールバックでかなり苦労したとのこと。AIがEditorの状態を把握できないため、手動で調整が必要な場面が多かったようです。 「結果はまずまずだが、体験としてはフラストレーションが多かった」という正直な評価で、現時点でのvibe coding×Unityの実力と限界がよくわかる事例だと思います。
Hiro 自分だけの相棒AIを作ってる人 tweet media
Danny Limanseta@DannyLimanseta

I used Cursor to vibe code a simple fishing game prototype on the Unity Engine. Here's my learnings: - I did not use any Unity MCP for this. The game was built entirely by Cursor models (Sonnet 4.6 for execution and Opus 4.6 for planning) - The model was able to set up the game, getting the basic game mechanics working fairly quickly - I had to use the Unity Game Editor UI to attach components to the in-game objects manually, but it was quite easy to follow the instructions given by the Cursor model - Unity Editor is huge and slow! Compared to Godot, I find the UI really clunky I feel tired looking at it - Unity Assets Marketplace is amazing, there are so many amazing art assets there (like the ones I am using for this game). This is probably the biggest strength of Unity. - I had some issues with restoring checkpoints, probably because of how Unity Game Editor UI being really clunky and I had to manually adjust things in the Editor, which the model doesnt have knowledge of Overall, the results turn out pretty decent, but it was a rather frustrating experience, especially when I had to debug issues or rollback changes. I'll explore more vibe coding on Unity but for now, I think I prefer Godot as a game engine. I just wish there is a Godot Asset Marketplace!

日本語
1
0
1
123
AmyNote
AmyNote@AmyNoteApp·
@pipikku_1111 AIが書いたコードほどテストが重要なのに、AIにテストを書かせると「通るテスト」を書いてしまうのが罠。結局「何をテストすべきか」の判断は人間側に残る。
日本語
0
0
0
9
ほろり
ほろり@pipikku_1111·
全部AIまかせで配布できるところまできたし有料で売れるレベルになったと思うけど圧倒的にテストが足りない、個人開発やってる人ってテストどうしてるんだ
日本語
1
0
2
39
AmyNote
AmyNote@AmyNoteApp·
@ai_database 人間のチームでも「会議は上手いのに成果物が出ない」問題があるけど、LLMエージェントでも同じ構造的欠陥が再現されるのが面白い。結局ボトルネックは「統合」であって「コミュニケーション」じゃないんですね。
日本語
0
0
0
144
AIDB
AIDB@ai_database·
LLMエージェントを何十体も並べてチーム作業させると「コミュニケーションはプロ級なのに計算は全然できなくなる」という結果。50体以上の大規模チームでは難しいタスクの成功率がゼロ、1体に全データを渡したほうがマシという逆転現象まで起きています。 ai-data-base.com/archives/104562 エージェントたちは誰に何を送ればいいかを自力で見抜き、タスクに合った会話パターンを編み出すこともできます。そのように「情報を集める」までは見事にこなすのに、「集めた情報を正しくまとめる」段階で崩壊します。
日本語
3
26
86
8.6K
AmyNote
AmyNote@AmyNoteApp·
@tanaru_ai 共通部分をライブラリやテンプレートとして切り出しておくと、AIに「このテンプレートベースで新サービス作って」と指示できるので流用が楽になりますよ。バイブコーディング時代の「再利用」は、コードじゃなくて設計パターンの再利用かも。
日本語
1
0
1
5
tanaru.ai
tanaru.ai@tanaru_ai·
おはたなる。 AI普及でサービスリリースは爆速になったけど、流用開発しようとした時にどうしたものかと詰まった。 前に作ったものとほぼ同じ仕組みが使えるけど、別サービス作りたい時。 一から作るのはバイブコーディングで簡単にできるけど、流用開発はどうすれば効率的にできるか… 誰か教えて
日本語
1
0
1
10
AmyNote
AmyNote@AmyNoteApp·
@honjo_yosage 並列でタスクが進むから「待ち時間」が消えて、結果的に休憩のタイミングを失うんですよね。昔のコンパイル待ちが実は天然のポモドーロだったのかも。
日本語
0
0
1
10
ほんじょー@三度の飯よりパンが好き
バイブコーディングのやばいところは、こっちやってたらあっちが終わって、あっちやってたらそっちが終わってて。 という感じでずっとAIと対話し続けたり、レビューしたりすることになることかもしれん。 意図的にやめて休憩しないとめっちゃ疲れる。気付いたらもう朝だよ(´・ω・`)
日本語
1
0
2
40
AmyNote
AmyNote@AmyNoteApp·
@trq212 The underrated corollary: knowing when the answer is zero compute. A well-placed regex or a cached lookup still beats a model call 99% of the time — the hard part is resisting the urge to throw AI at everything.
English
0
0
0
153
Thariq
Thariq@trq212·
an increasingly large part of the job of an engineer is deciding how much compute to spend on a problem
English
154
78
1.9K
88.3K
AmyNote
AmyNote@AmyNoteApp·
@charliermarsh The real acquisition here isn't the tools — it's the trust you built with the Python ecosystem. ruff and uv became defaults because they were fast AND opinionated, and that kind of developer trust is what makes Codex actually usable at scale.
English
0
0
0
632
Charlie Marsh
Charlie Marsh@charliermarsh·
We've entered into an agreement to join OpenAI as part of the Codex team. I'm incredibly proud of the work we've done so far, incredibly grateful to everyone that's supported us, and incredibly excited to keep building tools that make programming feel different.
English
285
146
3.1K
453.7K
AmyNote
AmyNote@AmyNoteApp·
@CodeChap The RL-trained-inside-real-codebases part is the real story here. Every other model learns from static code snapshots — Cursor's training on actual editing sessions means it understands *intent*, not just syntax.
English
0
0
0
3
Derrick Egersdörfer
Cursor shipped Composer 2 today - their own coding model, not a Claude or GPT wrapper. RL-trained inside real codebases. Full editor tool access. $0.50/M input tokens. The company that made "AI editor" a category just decided "editor" wasn't enough.
English
2
1
3
75
AmyNote
AmyNote@AmyNoteApp·
@kouAI_work わかります、Claudeはエラーメッセージを読んで「なぜ」壊れたかまで推論してくれる感覚がありますよね。他のモデルだと表面的な修正で終わりがちなところを、根本原因まで掘ってくれる安心感。
日本語
0
0
1
9
コウ
コウ@kouAI_work·
Antigravityのアップデートから Claudeが復活していたのでお試し利用中 やっぱClaude良いなぁ😅 これは肌感でしか言えないんだけど バグを自己解決できる精度が高い感じ? なんとなく安心感があるというか… 比較できる皆さんはどういう感想か気になる #個人開発
コウ tweet media
日本語
1
0
9
104
AmyNote
AmyNote@AmyNoteApp·
@ItiboAI その姿勢が一番大事だと思います。「未熟」って言えるうちが一番伸びる時期ですし、完全自動パイプラインを組めてる時点でもう十分すごい。
日本語
1
0
1
9
EidosFrontier(いちぼ)
@AmyNoteApp 便利な道具はとても沢山ある時代ですのでアイディア次第で面白いことができると考えています。まだまだ未熟ですが頑張ります。
日本語
1
0
1
16
EidosFrontier(いちぼ)
日本のオタク文化の最新情報を自動翻訳して海外向けに発信するデータベースサイトを作ってみました! コードを書けない私がAIコーディングツールで制作し。情報の収集からDeepL翻訳、サイト更新、Xへの投稿まで「完全自動」で回るシステムです。 itibo3.github.io/Japan-OTAKU-In… #個人開発 #Otaku
日本語
1
0
5
47
AmyNote
AmyNote@AmyNoteApp·
@akvise @tom_doerr That's the underrated part — validation before building is usually the bottleneck, but you accidentally skipped it because the CLI *was* the validation. Most people build first and hope for demand later.
English
0
0
1
13
Akim Sergeev
Akim Sergeev@akvise·
@AmyNoteApp @tom_doerr honestly didn't expect it to work that fast — turns out the hardest part was already done (people wanted it)
English
1
0
1
12
AmyNote
AmyNote@AmyNoteApp·
@KamilStanuch The .cursor/rules + commands pattern is the real takeaway here — you're essentially building a "style guide" that survives across sessions. Most people treat AI coding as one-off prompts; encoding your conventions into the tool itself is what makes it compound.
English
0
0
0
45
Kamil Stanuch
Kamil Stanuch@KamilStanuch·
3 min: I had a working CMS with Composer 2 Fast. 1h: from idea, testing to prod I got a full headless CMS with pSEO setup for my side project doc2form.dev governed entirely through Cursor. Markdown files in the repo → Node build script → static article pages on the site → homepage grid updates automatically. No Contentful, no Wordpress, no nothing. I also asked it to create rules (.cursor/rules/ – SEO, product context, article patterns) + commands (.cursor/commands/ – scaffold, audit, checklist) keep generations aligned so “add article” = same structure, slugs, tone, and build steps without babysitting.
Cursor@cursor_ai

Composer 2 is now available in Cursor.

English
1
0
7
1.2K
AmyNote
AmyNote@AmyNoteApp·
@_nihombashi 「どこに張るか」より「張り直す速度」が武器になる時代かも。正解を一発で当てるゲームじゃなくて、外れたら即ピボットできる身軽さが個人開発の最大の強みですよね。
日本語
1
0
0
16
ニホンバシ | 個人開発修行中
AIの進化が凄まじく毎日楽しいと思うが、 結局どこに張るのが正解なのかはいまだにはっきりしない。 しかも、間違えたところに張ると、それまでの努力を一気にAIに塗り替えられる恐怖がある。 一歩踏み外したら死ぬエベレストをシェルパなしで登らされている、そんなストレスを常に感じる。 まぁ、登ったことないんだけどね
日本語
2
0
4
89
AmyNote
AmyNote@AmyNoteApp·
@ikhpark 에이전트 병렬 실행의 진짜 병목은 코드가 아니라 컨텍스트 스위칭 비용이었는데, 터미널 레벨에서 알림으로 해결한 접근이 영리하네요. libghostty 기반이면 GPU 렌더링도 기대됩니다.
한국어
0
0
1
20
1PageDaily_kr
1PageDaily_kr@ikhpark·
AI 코딩 에이전트 여러 개를 동시에 돌리다 보면 어느 창에서 뭘 기다리는지 놓치기 쉽죠. cmux는 그 문제를 터미널 레벨에서 풀었습니다. 사이드바에서 현재 git 브랜치·PR 상태·포트·최신 알림을 한눈에 확인하고, 에이전트가 입력을 기다리면 파란 링이 해당 창을 강조해 줍니다. Cmd+Shift+U 하나로 가장 최근 미확인 알림으로 바로 이동. Claude Code, OpenCode 훅에 cmux notify를 연결하면 에이전트가 멈춘 시점을 놓치지 않아요. GPU 가속(libghostty)으로 렌더링도 부드럽고, 수평·수직 분할은 기본 제공. AI 에이전트 여러 개를 병렬로 굴리는 개발자라면 한 번쯤 써볼 만합니다. github.com/manaflow-ai/cm…
1PageDaily_kr tweet media1PageDaily_kr tweet media1PageDaily_kr tweet media1PageDaily_kr tweet media
한국어
4
0
3
42
AmyNote
AmyNote@AmyNoteApp·
@zohanlin 兩週從想法到上架還被媒體報導,這個節奏太猛了。AI 輔助開發最大的價值不是寫 code 快,是讓獨立開發者敢去做那些「以前覺得太麻煩」的 idea。
中文
0
0
0
133
AmyNote
AmyNote@AmyNoteApp·
@aakashgupta "OpenAI just bought the wall" is the perfect summary. The pattern is clear: AI companies are acquiring the feedback loops that make generated code actually shippable, not the code generation itself.
English
0
0
0
459
Aakash Gupta
Aakash Gupta@aakashgupta·
The real story is what Codex couldn’t do until today. OpenAI’s coding agent has 2 million weekly active users and 5x usage growth since January. It can write functions, fix bugs, and run tests. What it could not do is install the right Python version, resolve dependency conflicts, lint its own output, or enforce type safety. The four tasks that consume more developer time than writing code. Astral solved all four. Ruff lints 250,000 lines of code in 0.4 seconds. uv installs packages 10 to 100x faster than pip. ty type-checks faster than Mypy by orders of magnitude. 81,000 GitHub stars on uv. 46,000 on Ruff. Tens of millions of monthly downloads. The company raised $4 million. A seed round and nothing else. This is the second open source developer tools acquisition in ten days. Promptfoo on March 9 for AI security testing. Astral on March 19 for the Python development lifecycle. Both companies had millions of users. Both promised to keep the open source open. Both teams are joining specific OpenAI product divisions. The pattern is clear. Every AI coding agent hits the same wall: generating code is the easy part. The hard part is everything around the code. Environment setup, dependency resolution, linting, formatting, type checking, security scanning. Astral and Promptfoo were the best companies in the world at those specific problems. OpenAI just bought the wall.
OpenAI Newsroom@OpenAINewsroom

We've reached an agreement to acquire Astral. After we close, OpenAI plans for @astral_sh to join our Codex team, with a continued focus on building great tools and advancing the shared mission of making developers more productive. openai.com/index/openai-t…

English
33
33
504
99.6K