Tweet épinglé
StudyHour.uk
Time is ticking - Do not wait further.
GCSE Students and Teachers , take a look !
Perfect for Easter revision time.
English
StudyHour Labs
154 posts
StudyHour Labs
@studyhourlabs
Revision planning and AI study tools for UK students.
London Inscrit le Mart 2026
12 Abonnements6 Abonnés
Why so ? Computer use is missing in Codex on mac silicon
@OpenAI
Tibo@thsottiaux
Codex just got a lot more powerful. Computer use, in-app browser, image generation and editing, 90+ new plugins to connect to everything, multi-terminal, SSH into devboxes, thread automations, rich document editing. Learns from experience and proactively suggestions work. And a ton more.
English
Codex seems to be very precise when it comes to ideas. But long usage (2 days) of both Codex and Claude make them a mediocre developer. I use for ideas, but I tell what to do, never vibe.
Alex Ibragimov@alexwtlf
Has anyone actually switched from Claude Code to Codex?
English
Why is hitting ENTER taking a while now. The time b/w hit and submit is quite noticeable. It used to be super instant without any spinning.
Tibo@thsottiaux
Codex just got a lot more powerful. Computer use, in-app browser, image generation and editing, 90+ new plugins to connect to everything, multi-terminal, SSH into devboxes, thread automations, rich document editing. Learns from experience and proactively suggestions work. And a ton more.
English
New Codex update: Hitting ENTER is not imemdiately submitting, thinks and then submits !
Not pleased
English
English
The Claude Code Desktop app is an affront on software. As developers, we should be offended that they chose to ship something this awful.
Rushed out my video because I feel like I'm going insane.
English
Relying on a vision model to distinguish a small, sparse technical symbol from "noise" or "blank space" in a high-padding table cell is prone to false negatives.
English
@OpenSilverTeam I coded in wpf myself and have seen the whole chain from Win32. *Everything* is possible in Web. I am not sure why will anyone adopt this and how will you sell ? I am curious. Do you have real scenarios to prove and justify?
English
WPF is a unified, coherent UI model:
declarative XAML
+ a real UI designer,
strong MVVM/data binding that just works
without piles of extra libs or trial-and-error hacks.
On the JS side, there is a lot of churn,
overlapping choices,
and complexity.
It often feels like stitching together a UI stack from constantly moving pieces.
Even the HTML/CSS/JS model has limits for building rich GUIs,
this post explains it well: x.com/hasen_95dx/sta…
English
WPF got the developer experience right. The web got everything else right.
OpenSilver is the first UI framework that combines both. Write in WPF, ship real HTML DOM, not pixels on a canvas, and get accessibility, SEO, Ctrl+F, text selection, browser translation, mobile interactions, browser extensions, screen readers, and more.
Try it in seconds at XAML.io. No setup. No signup.
English
AI tooling, I have realised best works in macOS. Windows and Wsl is a bit of stretch. I have been a Win32 programmer but switched to macOS this year. Moreover npm ecosystem is blazing fast!
Max Weinbach@mweinbach
You can tell the Claude app devs are only focused on macOS and not Windows, because they just hard code the Mac command symbol for Claude so I see it on Windows That would be so confusing if I didn’t know what it was lmao
English
This is a very limited library - It's a fast pre-classifier for pipelines that need to decide before sending a PDF to an expensive OCR service.
Ramya Chinnadurai 🚀@code_rams
PDFs break almost every AI agent pipeline I've built. Not because of the AI. Because of the parser. Here's what actually happens when your agent reads a PDF: 1. Tables get flattened into garbage text 2. Formulas turn into random symbols 3. Multi-column layouts merge into one unreadable block 4. Speed tanks when the file is more than 10 pages The AI never sees what you think it sees. Firecrawl just shipped Fire-PDF, a Rust-based parser. What changed: 1. PDF to markdown, 5x faster than before 2. Tables preserved, not flattened 3. Formulas kept intact 4. Zero config, drop it in and it works I've had Firecrawl in my stack for months. The PDF part was always the weak link. Most "AI can't read this doc" problems are parser problems. Fix the input, the AI handles the rest.
English
Launch App via claude desktop on localhost
Toggle b/w desktop and mobile mode
Click on Task from top menu
After this step, the rendering goes bad
And then, do as in video
Causes bad rendering
Anthony Morris ツ@amorriscode
Today we're launching a rebuilt version of Claude Code on desktop. The app has been redesigned for the ground up to make it easier than ever to parallelize work with Claude. I haven't opened an IDE or terminal in weeks. Excited for you all to give it a shot!
English
For me, today, the Claude Plugin in @code is doing a decent job, CC Desktop went bonkers y'day, not following any instructions.
Kent C. Dodds 🏹@kentcdodds
Increasingly I'm discovering that Claude's Desktop app does a better job at doing what I need than ChatGPT's...
English
PyMuPDF + Docling does all the job for FREE, no need of $$$ LLMs
English
True; I have been dealing with this. Wrote my custom parser over docling core. EVERY pdf layout is different especially when you have to deal with Question papers.
It is a very hard problem.
Ramya Chinnadurai 🚀@code_rams
PDFs break almost every AI agent pipeline I've built. Not because of the AI. Because of the parser. Here's what actually happens when your agent reads a PDF: 1. Tables get flattened into garbage text 2. Formulas turn into random symbols 3. Multi-column layouts merge into one unreadable block 4. Speed tanks when the file is more than 10 pages The AI never sees what you think it sees. Firecrawl just shipped Fire-PDF, a Rust-based parser. What changed: 1. PDF to markdown, 5x faster than before 2. Tables preserved, not flattened 3. Formulas kept intact 4. Zero config, drop it in and it works I've had Firecrawl in my stack for months. The PDF part was always the weak link. Most "AI can't read this doc" problems are parser problems. Fix the input, the AI handles the rest.
English
Well, for me, I used AI more like a helper, reviewer. Parsing specific PDFs into clean render has been daunting. AI was a good enabler but not reliable!
David K 🎹@DavidKPiano
If you're reading this and thinking "skill issue", you're probably working on problems that LLMs have seen thousands of times. That's fine, but don't confuse "AI works for my work" with "AI works". Get out of the AI comfort zone.
English
export_to_markdown() walks the same body tree as my parser - orphaned TextItems (ring-around options Docling mis-classifies as PictureItems) are invisible to both. It also produces a flat markdown blob with no Q-boundaries, no marks extraction, and no keyword index.
#docling
Tom Dörr@tom_doerr
Converts documents into formats for agents via MCP github.com/docling-projec…
English