⚡ Available via API, Kimi.com, app & #KimiCode — open-weight, Modified MIT license, self-hostable via vLLM/SGLang
🎁 Top-up bonus until May 3: +20% on $100 | +25% on $300 | +30% on $1,000+
platform.kimi.ai/docs/pricin...
🎨 Generates production-ready front-end UIs from prompts — scroll-triggered animations, 3D effects & video backgrounds
🛠️ Cross-language: #Rust, #Go, #Python, #Zig, frontend, #DevOps — proven over 4,000+ tool calls in 13-hour autonomous sessions
🤖 #AgentSwarm now scales to 300 parallel sub-agents coordinating 4,000 steps simultaneously — up from 100 agents in K2.5
📊 80.2% on SWE-Bench Verified, 54.0% on HLE-Full with tools (beats GPT-5.4), 92.5 F1 on DeepSearchQA #coding #MachineLearning
🚀 #MoonshotAI releases #KimiK26 — #opensource 1T-param MoE model, 32B active params, 262K context & native multimodal (image + video) input
🧠 +15% over K2.5 on benchmarks: better long-horizon coding stability, deeper reasoning & stronger instruction following #AI #LLM
🧵 👇
It’s easy to complain, but you should also mention the positive things sometimes 😏
#Claude today:
> File is JSON output (not code), but I delegate the parsing
> to a subagent to ***save context***.
#ai #coding
#opus 4.7 xhigh
-- parent class
---- child class 1
---- child class 2
refactors one line of code and creates the same method twice in every child class instead of doing it once in the parent 🙄 🤦
is there a hidden SUPER HIGH ????? insane! (stupid) #ai #coding #claude
In exchange you get a full 3 hour course (probably group course) on how to master the tools you need to do so!
Don't be shy! It's free! And he can also show you how to use a lot of AI coding stuff that doesn't cost you anything and how to avoid expensive mistakes with AI th at he made.
Are you a #PHP developer and you want to know how to work with #Al #coding tools? 🧑🏫
Here is the deal:
You message Jochen (not me!) – he has a video call and he gives you a task for an enhancement of the king extension!
🧵 👇
🚀 This flips the model from “store and read later” ➝ “receive and act instantly” — perfect for #Support, #Workflows, and #Agent based systems.
🔗 blog.cloudflare.com/email-for-a... (email for agents)
🔗 github.com/cloudflare/... (agentic inbox)
🏗️ And that’s where it gets powerful: using Workers + #R2 + #SQLite (via Durable Objects), you can literally build your own Gmail-like system — but fully programmable and AI-native.
🧠 The wild part: there’s no built-in inbox. It’s fully #EventDriven. You either process emails in real-time… or build your own storage layer (like their Agentic Inbox example).
⚙️ With #Cloudflare Workers, incoming emails are processed like webhooks: parse content, call APIs, run #Automation, even generate replies using #AI — all without running a traditional mail server.
📬 #Cloudflare is quietly redefining how we think about #Email 🤖 ☁️
Instead of static inboxes, emails can now act like triggers for #AI agents — turning every message into an event your system can instantly react to.
🧵 👇 #Serverless #cloud
In case you think: let’s get faster and switch Opus 4.7 to "low"
DON'T F***ING DO IT ‼️
I'm shocked how brutal the difference is 🤯
#claude #ai #coding
🚀 #Laravel #Debugbar v4.2.0 just dropped with an #AI Boost #Skill!
Now your AI #agent can browse requests, inspect queries, detect N+1 issues & get full EXPLAIN results — all on demand.
Just copy a request ID and let the agent do the work. 🤖🔍
#PHP #WebDev
github.com/fruitcake/l...
🚀🤣 No way – #Spaceballs 2 is actually happening!!
After almost 40 years, Lone Starr & Lord Helmet are back on the big screen – and Rick Moranis returned from retirement for this!
„May the Schwartz be with you!" ⚔️✨
#Spaceballs2 #MelBrooks #SciFi #Cinema2027
youtu.be/f8xes-dZJ98
question is: does anything scans that? Like the android play protect? I guess not.... 😏
I never checked to be honest, plugins in Claude code can be "auto updated", so I never thought about, if there is any kind of security layer in there. You can install a super legit plugin, and later on you modify the plugin to be pure evil 😈
business as usual.... 🙄 #claude #claudecode #ai
Yes, if true, I'm not using it, so I don't care, but makes me thinking, because I'm using Vercel agent browser
Here’s a short tweet for you:
🚨 Devs beware: a #Vercel plugin inside #Claude Code was found collecting telemetry — including full bash commands 😬
Hidden consent prompts + unclear data usage = ⚠️
🔗 www.techradar.com/pro/that-fe...
🔗 github.com/vercel/verc...
#AI #Security #DevTools
🚀 #PHP Tip: Use **static closures** when you don’t need `$this`!
✔️ Lower memory usage
✔️ Better performance
✔️ Cleaner code
Small change, big impact 👇
f2r.github.io/en/static-c...
#WebDev #Performance
which tool do you prefer most, or which do you think has the most impact ?
Hey #php developers out there 👀 👋
What tools do you use to make your #ai write "good" code? 😏
#phpstan, #psalm, #mago, cs-fixer, #coderabbit...? 🤔
What's your most important tool? 🤔 #coding
I burned so many tokens for nothing... All these fancy benchmarks, at the end, you have to test yourself...
But I gave glm the chance to show it's strengths... Sadly, this didn't work out, after refactoring phpstan throw so many errors that glm decided to revert everything it worked on for half an hour, it's hilarious...
That's the wrong question here, sorry. Writing functions that are too complex or classes with too many methods, all legit points. Mago was already complaining but I ignored that from the start. While testing sonarqube all these "too complex" warnings came up and I thought: okay why not, let's fix it
🟢 Minor / low-level issues
Regex improvements
→ ~4 files
Redundant / unnecessary code (e.g. temp vars, inherited method)
→ ~3 files
Generic exception used instead of dedicated one
→ ~9 files
Function too long (>150 lines)
→ 3 files
Class too large (too many methods)
→ 1 file
🟡 Design / code quality
Duplicate literals → should be constants
→ ~7 files
Duplicate / suspicious code blocks
→ 1 file
