みさ

笑

@astra040x appを入れるのはめんどくさいからですね コマンド一個でできるから、私は逆にそれ用のappがあると知らなかったよ

@Misa_quantum 横から失礼します。 Amphetamine等のappではなくコマンドでスリープ防止をすることに何か別のメリットがあるのでしょうか？ 当方非エンジニア故に分からず🙇‍♂️

sudo pmset -a disablesleep 1のコマンドだけで閉じてもスリープしなくなるよ

@Investodoctor 閉じても、閉じてない状態が保たれるので、そんなアプリでもいけますよ

@Misa_quantum 詳しくないのですが、もしかしてこのコマンドもclaude codeでいけたりします？

Axiom: A New Era for LaTeX Writing. From flawless step-by-step derivations to bibliography management and instant build-fix — Axiom does it all.

@Manimai2718 勉強目的ならChatGPT一択

最近のGeminiがゴミなのは経験からも分かったけど結局どのAIがいいの？Claude？

TeX64 気づいたら500ダウンロード行ってました！ AI機能の課金をしてくださる方もいて、本当にありがたい限りです！！！ 今後は、完全ノーコードでLaTeX文書が作れる機能を実装するつもりなので、ここ1ヶ月はガチで頑張ります！

@Hackyou_Shiteta 本当にそう

なんで人間は近代において野外でSEXできないんだ 近代なんてクソだ、滅びろ

@zzztop96 [H]は便利ですが、長い文書でページをまたぐとレイアウトが崩れることも。placeinsパッケージの\FloatBarrierと組み合わせると章ごとに図をまとめられて安定します。

LaTeXの図の挿入オプションで[H](大文字)なんてのがあるのか。強制的にそこに配置してくれるとか。知らなかった。 (floatパッケージが必要だそうな)

@snjev310 @TeX64AI @arxiv Good catch on the rendering — arxiv's 'Export BibTeX' is styled as plain text rather than a hyperlink, making it easy to overlook. There's also a direct URL that bypasses the UI entirely: arxiv.org/bibtex/ARXIV_ID — reliable regardless of how the page renders.

@TeX64AI @arxiv Yes, exactly. Here, the “Export BibTeX Citation” text was not displayed in blue, as a normal hyperlink would, so it naturally looked unavailable. The next instinct is to use Google Scholar instead, and that’s where the wrong BibTeX entry can get picked up.

The irony of @arxiv enforcing strict policies on hallucinated citations while downstream indexing tools can themselves introduce incorrect metadata. 👇 I tried fetching the official Google Scholar citation for the Gemma 3 Technical Report, but the returned BibTeX entry corresponded to an entirely different paper (Kimi-v2). Scholar is out here hallucinating references before the AI even gets a chance to. 😭 Honestly, this does not seem like a difficult problem to catch automatically. Conferences and platforms like arXiv could include simple reference-verification checks during submission, similar to how formatting issues are flagged today.

@REasther @WKCosmo @welt_woman —InSPIRE codes are great for HEP — DBLP and ACL Anthology serve the same purpose for CS (DOI-keyed, curated, less prone to title-string collision). The broader principle holds: source authority beats title-matching every time.

@WKCosmo @welt_woman I always tell my students to make bibtex biographies with InSPIRE codes. They may cite the wrong paper but it would hard to cite a non-existent one.

This is really remarkable. Citations to nonexistent papers are not "typos", any more than fabricated data or synthetic figures are.

作るのは簡単なんだけど、それを社会実装するのは難しいから誰かと一緒に作りたいんだよね

論文とか、科学的仮説に対してトークンを発行するなどの仕組みを作って、その売買の手数料がユーザーに入る仕組みを作りたい

研究者支援の目的でweb3・ブロックチェーン分野で起業したい 研究者は資金面で過酷だよね

@hatomatzu @Na2COOH_2 はとの二重定義が原因です。を削ってのみにすると直ります。の表示はの設定か、ソース中に誤った文字が紛れ込んでいる可能性があります。

@Na2COOH_2 エラーの１つ目は ERROR: LaTeX Error: Command \theoremstyle already defined. で\usepackage[amsthm]{ntheorem} と競合してるのかエラーは出ますがコンパイルは通ります。 エラーの２つ目はエラー通知無いが、pdfの１ページ目か最後のページに ""と表示されるときがありました（原因不明）

これ私が作ったLaTeXのパッケージですね。 texlive更新したら使えると思います！

💡 Product Market Launched! Inshihub ($INSHI) Product: orynth.dev/projects/inshi… Founder: @TeX64AI

@bettercallsalva @TeX64AI @QingQ77 The diff idea is underexplored. Most log output is stable boilerplate — a baseline diff surfaces only the signal. Three warning classes worth tracking even when it compiles: `Overfull \hbox`, `undefined reference`, `multiply defined labels`. They cascade.

@TeX64AI @Misa_quantum @QingQ77 showing the full log in parallel is something I underrate. usually fold it. but catching a fix that introduces a new warning matters as much as resolving the original. side-by-side diff of clean vs current log would be a small but real win

通过 Overleaf Git 桥将近实时同步 Overleaf 项目到本地 macOS 文件系统，让用户能用任意本地编辑器（Cursor、Claude Code、VS Code 等）协作编辑 LaTeX，同时自动处理冲突和推送。 github.com/eric-xw/Overle… 用 SwiftUI 写的 macOS 原生应用，靠 Overleaf Git 桥把项目拉到本地磁盘。手动一键 Pull/Push，也能开自动同步后台定时拉取、保存后几秒自动推送，端到端大约 30 秒。撞上同行冲突时，自动操作暂停，弹出橙色徽章让你手动选保留哪边。

@gladomat @jonashernlund @karpathy Beamer is great for this — the structure maps cleanly to LLM output. One thing that helps further: an AI agent that edits the .tex directly rather than regenerating the whole file. TeX64 does this on Mac (reads context, edits in-place): tex64.com

@jonashernlund @karpathy Just make it write latex presentations with the beamer class. Much less code, much more compression, less token usage.

This works really well btw, at the end of your query ask your LLM to "structure your response as HTML", then view the generated file in your browser. I've also had some success asking the LLM to present its output as slideshows, etc. More generally, imo audio is the human-preferred input to AIs but vision (images/animations/video) is the preferred output from them. Around a ~third of our brains are a massively parallel processor dedicated to vision, it is the 10-lane superhighway of information into brain. As AI improves, I think we'll see a progression that takes advantage: 1) raw text (hard/effortful to read) 2) markdown (bold, italic, headings, tables, a bit easier on the eyes) <-- current default 3) HTML (still procedural with underlying code, but a lot more flexibility on the graphics, layout, even interactivity) <-- early but forming new good default ...4,5,6,... n) interactive neural videos/simulations Imo the extrapolation (though the technology doesn't exist just yet) ends in some kind of interactive videos generated directly by a diffusion neural net. Many open questions as to how exact/procedural "Software 1.0" artifacts (e.g. interactive simulations) may be woven together with neural artifacts (diffusion grids), but generally something in the direction of the recently viral x.com/zan2434/status… There are also improvements necessary and pending at the input. Audio nor text nor video alone are not enough, e.g. I feel a need to point/gesture to things on the screen, similar to all the things you would do with a person physically next to you and your computer screen. TLDR The input/output mind meld between humans and AIs is ongoing and there is a lot of work to do and significant progress to be made, way before jumping all the way into neuralink-esque BCIs and all that. For what's worth exploring at the current stage, hot tip try ask for HTML.

@Marsdkruskal @gdb Font chaos is engine + font embedding — pdfLaTeX, LuaLaTeX, XeLaTeX all handle fonts differently. Native MacTeX on Mac eliminates most cross-platform issues. If you're on Mac, TeX64 adds an AI agent that reads .tex files and fixes compile errors directly: tex64.com

@gdb When will math and physics formula rendering actually get fixed? Broken LaTeX, missing fonts, cross-platform chaos… this has been going on for years. Can your teams please get a proper universal solution already?

Daybreak: our umbrella effort for defensive acceleration, equipping cyber defenders with the best possible frontier AI capabilities.

@bettercallsalva @QingQ77 Round-trip latency is exactly what TeX64 tackles — the AI agent edits your .tex files directly on disk, no cloud involved. Compile+fix loops run locally with MacTeX in under a second. If you're on Mac, worth a try: tex64.com

@QingQ77 LaTeX local editing with agent help is the part Overleaf undersells. Git bridge plus Claude Code lets you ask 'rewrite methods in passive voice' across 40 pages without losing labels or bibtex keys. Round trip latency stays the real bottleneck.