El Pid ретвитнул
honey wake up! the new @redmonk language ratings just dropped!
redmonk.com/sogrady/2026/0…
English
El Pid
9.3K posts
El Pid
@pidster
Ingénieur. Chief Snarkitect. Saucerer.
The Moon Присоединился Kasım 2008
497 Подписки460 Подписчики
El Pid ретвитнул
Anyone else seeing Sonnet 4.6 insist on using 'go' instead of 'cargo' in their Rust projects? @AnthropicAI
English
El Pid ретвитнул
We’re recruiting for a Data Engineer to join our Research team at Liverpool FC. In my (perhaps biased) view, it’s a great opportunity to join an excellent team, work at the forefront of sports analytics, and have a tangible impact. linkedin.com/posts/conorqui…
English
@juliendubois Yes, very much understand. I'm just wary of instructions that some models can take too literally. Instead of the negative/restrictive : "Do NOT ... etc" I'd recommend a firm positive instruction: "The user must review these changes before they are committed to git", (or similar).
English
@pidster Huge discussion we've had many times with JHipster (where we have the opposite decision!).
As it's an AI Skill, the idea here is that the user is supposed to review everything first, before committing. Also, it could mess up stuff if you have multiple agents in parallel.
English
Just released Dr JSkill, an Agent Skill that helps AI tools create Java + Spring Boot applications.
It has tons of best practices that I gathered over the past 15+ years building JHipster!
Here's the GitHub repo + video:
github.com/jdubois/dr-jsk…
youtu.be/Z8tWJ0NTy08
YouTube
English
El Pid ретвитнул
Tip for the Google Gemini team: if you want to help Google truly get ahead in the AI era, use your hefty influence to get it so setting up API access to your own calendar doesn't involve THESE steps
Mario Zechner@badlogicgames
Need a Google Calendar CLI that works well with agents? Here you go: github.com/badlogic/gccli
English
El Pid ретвитнул
Devious new attack class defined by @wunderwuzzi23: "Cross-Agent Privilege Escalation"
If someone has both Claude Code and GitHub Copilot on their machine you can prompt inject one and have it edit the config for the other in order to further escalate your attack's privileges!
English
El Pid ретвитнул
El Pid ретвитнул
LLMs love XML — the most instructive of configuration languages. Leaked AI-startup system prompts all use it, as does The Universal Jailbreak from @hiddenlayersec hiddenlayer.com/innovation-hub…
El Pid@pidster
I updated github.com/pidster/person… to use XML tags in the persona instructions after a *quite interesting* conversation about reinforcing instructions to LLMs with @sublimino last week
English
@pidster An outrageously interesting conversation, with a frayed tapestry of threads to tie up next time! 🍻
English
I updated github.com/pidster/person… to use XML tags in the persona instructions after a *quite interesting* conversation about reinforcing instructions to LLMs with @sublimino last week
English
I have some ideas about how to improve this, but for now: npmjs.com/package/person….
MCP serving Personas that the agent can (and will) dynamically choose to adopt, depending on the task it is performing.
English
Tool creation to solve problem relating to agent productivity
Chip Huyen@chipro
I open sourced Sniffly, a tool that analyzes Claude Code logs to help me understand my usage patterns and errors. Key learnings. 1. The biggest type of errors Claude Code made is Content Not Found (20 - 30%). It tries to find files or functions that don't exist. So I restructured my code base for discoverability, and the average number of steps Claude Code needs for each instruction went from 8 to 7 steps.
English