Alkmpfer retweeted
According to Cloudflare Data Explorer the amount of russian bot traffic has dropped to 49% after Ukraine struck a yandex data center. Could this disrupt the midterm elections in the US?
Alkmpfer retweeted
Gemini 3.8 Flash benchmarks. And holy cow! Flash outperforms 5.6 sol and opus 5 on terminal bench 2.1, HLE and much more.
Google is back!
⌨️ Anthropic 官方把「怎么写好提示词」做成了一套能边敲边学的互动课程。
GitHub 揽获 3.7 万 stars。
大部分人学提示词是刷别人总结的技巧帖,看的时候觉得有道理,轮到自己写还是一句「帮我写个文案」丢过去。这套教程反过来:九章内容每节都留了实验区,你改一个字、把角色设定删掉、把数据和指令混在一起,立刻能看到 Claude 的输出怎么变差。亲手弄坏一次的印象,比看十篇技巧帖都深。
内容从提示结构、清晰表达、角色设定、数据和指令分离,一路讲到逐步思考、示例运用和怎么减少幻觉。第九章直接上聊天机器人、法律、金融和写代码的实战练习,附录还补了提示链、工具调用和检索。
不想装环境的话,它还有一份 Google Sheets 版本和练习答案表。
提示词不是玄学,是能一节一节练出来的手艺。
GitHub:github.com/anthropics/prompt…
Our general-purpose coding agent just scored 100% on the ARC-AGI-3 interactive reasoning benchmark.
NVIDIA AVO completed all 183 levels across all 25 public environments, figuring out what to do with no instructions, explicit rules, or stated goals.
Alkmpfer retweeted
Folks! It has crossed 3x!
Qwen 3.8 27B now runs 214.6% faster on Apple Silicon.
All of it from an open leaderboard on @yukonresearch.
The OG @ivanfioravanti holds the record right now, set with GLM-5.
The top 5 records were each set with a different frontier model.
GLM-5, Claude Fable 5, Claude Opus, Grok 4, GPT-5.6
Accelerate open intelligence and open frontier research through open innovation.
Fastest weekend i've watched on a leaderboard.
Qwen 3.8 27B now runs 193.4% faster on Apple Silicon than it did on Friday. That's 2.94x!
MLX kernels written by people pointing DeepSeek V4 Pro, GPT-5.6 and Claude at an open leaderboard on @yukonresearch.
Accelerate open intelligence and open frontier research through open innovation.
Alkmpfer retweeted
#nerdystuff in pausa pranzo
#Fingerprint cerca uno username o una mail su oltre 700 piattaforme e tira fuori ogni account collegato in tempo reale: impressionante (e un filo inquietante) quanto emerge da un solo nickname!
🔍 fingerprint.to
Alkmpfer retweeted
Lost my phone at the office and spent 30 minutes turning the place over. Find My was disabled by MDM.
Out of ideas, I asked Claude how I could find it. It suggested tracking the Bluetooth signal strength, then wrote me a meter in about a minute.
I walked around watching the number climb. Found it.
Apparently you can just make the tool you need now.
Code: github.com/ben-z/findphone
Alkmpfer retweeted
Google acaba de soltar un curso de solo 1 hora sobre Ingeniería Agentiva que destroza a la mayoría de cursos de pago 🔥
Timestamps:
00:00 → Cómo construir tu primer agente de IA
08:24 → Memoria de agente (corta, persistente y larga)
28:34 → Bucles agentivos y agentes de larga duración
40:04 → MCP vs API (esto solo ya vale la pena)
1:00:22 → Sistemas multiagentivos
En 60 minutos aprendes más que en 10 cursos pagos. Míralo hoy. Luego lee el artículo y construye un sistema agentivo que se auto-mejora solo.
¿Lo vas a ver ahora o lo guardas para “después”? 👀
Alkmpfer retweeted
🚨 One orchestrator. 10 parallel agents. 100+ tokens a second.
All local.
The @googlegemma team just dropped a MASSIVE demo for Gemma 4 26B.
They built a concurrent workflow that lets the 26B model coordinate an entire team of sub-agents on your machine.
Out of the box, the cookbook lets you run 10 parallel agents to:
→ Code an entire SVG art gallery in seconds
→ Translate text simultaneously
→ Generate ASCII art
→ Write parallel code
Spinning up multi-agent systems locally has never looked this fast or this accessible.
100% free and open-source.
repo link in 🧵↓
Alkmpfer retweeted
Oggi è la Giornata Mondiale della Sclerosi Multipla. Mi sembra un buon momento per ripostare la mia guida un poco cretina alla #sclerosimultipla.
#WorldMSDay #GiornataMondialeSclerosiMultipla @aism_nazionale (continua⬇️)
Alkmpfer retweeted
Il pomeriggio al #xeoneday 2026 inizia con Gian Maria Ricci con il talk “Architettura ai tempi dell'AI”
Alkmpfer retweeted
Qwen 3.7-max beats Opus 4.7 and GPT-5.5
We tested three frontier models on a real agentic task: write a Tetris bot that plays the game and trains itself. Each model could read its own code, run benchmarks, and rewrite itself across 10 iterations. Then we compared the final bots head to head.
Qwen 3.7-Max: training cost $1.32, bot improvement +56%
Claude Opus 4.7: training cost $12.15, bot improvement +28%
GPT-5.5: training cost $2.85, bot improvement +7%
Qwen won on every dimension - biggest jump, 9× cheaper than Claude, 2× cheaper than GPT. Long agentic loops is where Qwen Max actually delivers.
Alkmpfer retweeted
If you want to age your sys admins 30 years overnight, remember that Active Directory is fully unicode compatible, so you can rename your laptop with emojis it its hostname, and it will reflect like that in AD
ping desktop-🤷♂️👍👌.mycompany.local
Alkmpfer retweeted
Il #XeOneDay2026 scalda i motori con la cena pre-conf speaker: buon cibo, la giusta quantità di prosecco e l’entusiasmo decolla!
Alkmpfer retweeted
Quick update: pushed the 4090 further!💡
192K context at 152 tok/s on Qwen3.6-27B, single GPU. 128K hits 159.
Same Q4_K_M.
Vanilla Qwen3-1.7B draft beat the distilled 4B draft.
Smaller > smarter for spec-dec.
Next: 1M context locally + 250-400 tok/s via DFlash + TurboQuant. Receipts coming.
Alkmpfer retweeted
🚀Proseguiamo con la presentazione dei nostri spettacolari speaker del One Day! Alle 14:15 sul palco del Belstay Venezia Mestre il 22 maggio 2026, salirà @alkampfer, il quale ci parlerà di:
📣Architettura ai tempi dell'AI
✒️Per info e reg. primo commento





































