Iba a testear GPT-6.2 Astra hoy, pero viendo las malas reviews voy a esperar unos días antes de dedicarle tiempo. ¿Lo has probado ya?
43
Replying to @alegoml
Yo lo tengo montado con Hermes + Telegram. Mi agente/asistente (Pepper) supervisa al resto de agentes. Veo práctico tener agentes especializados en ciertas tareas + un agente por encima que organiza y supervisa.
1
1
157
Vienen curvas
In the last 24 hours:
1. iPhone 11 through 16 are apparently vulnerable to remote attacks.
2. SSL certificates are broken until browsers remove eMudhra CA.
3. PHP >9 vulnerability CVE published (tomorrow).
The world is about to witness things that have never been seen before.
Readers added context they thought people might want to know
The post's iPhone claim is inaccurate; a leaked Cellebrite forensic tool targets a Bluetooth bug patched in iOS 26.5 and does not affect updated devices. The eMudhra CA and PHP issues are real incidents but framed sensationally.
justin.sh/blog/2026/10/c…
blog.mozilla.org/security/2026/…
php.net
109
Stephen De Winter 🤘 retweeted
Okay, and now the model card I've been waiting all week to build and share. This has been the question on my mind since Dev Day on Tuesday, so I'm guessing it's on other people's minds as well.
GPT-6.1 Sol vs. Astra vs. Opus 5.5, across every effort level on @VulcanBench Frontier v4.
And before I share my own insights, one note I think more benchmarks should be sharing. What accuracy score do you really need a model/effort to hit, for your routine daily coding tasks?
So many benchmarks show performance at Max effort, but I never use Max effort, never use Extra High, sometimes use High, but mostly use Medium and Low with the current frontier because that's more than powerful enough for most routine daily coding tasks.
And, we all know you don't need to use the best frontier model with the highest accuracy score. Sure you can do that, but you'll spend a fortune, and you'll find yourself in the 2025 agentic coding pattern where you give a model a task and just sit and wait.
On VulcanBench Frontier v4, I typically say that a model/effort combo between 80 - 85 is great for easy/medium difficulty tasks, 85 - 90 is a good choice for medium/hard tasks, and above 90, well, you may never need.
For me personally, I think I need model/effort combos that score above 90 maybe 5% of the time, the sweet spot for most of my routine daily coding work is model/effort combos that score in the low 80's on Frontier v4.
Do with this info what you will, we're all learning together too so take it with a grain of salt. And now, for the benchmark I've been waiting to share, see below.
Live long and benchmark 🖖
Mira que me gusta Hermes, pero ya van 3 actualizaciones que han dado problemas.
La última desactiva el plugin Kanban y toca reactivarlo.
Es solo una molestia, pero es una herramienta de trabajo y como tal debe ser productiva.
72
Los buenos tiempos!
This is also basically how everyone used the internet 20 years ago
You’d end up on some guy’s website about Civil War cannons, click his links page, discover a county historical society, then spend 40 minutes reading a forum for antique tractor owners
“What sort of website is this?” was half the experience
We called it surfing for a reason, now you open the same 4 apps every day
68
Stephen De Winter 🤘 retweeted
This is also basically how everyone used the internet 20 years ago
You’d end up on some guy’s website about Civil War cannons, click his links page, discover a county historical society, then spend 40 minutes reading a forum for antique tractor owners
“What sort of website is this?” was half the experience
We called it surfing for a reason, now you open the same 4 apps every day
Hoy deepseek-flash está respondiendo un porcentaje alto de los mensajes en chino (en Hermes), los días anteriores no pasaba. ¿Alguien más con el mismo problema?
No doy con la tecla para solucionarlo.
1
129
Testeando Deepseek v4.14-Flash.
Primeras impresiones: bastante más rápido que GLM-5.3-flash.
A nivel de escritura creativa, se queda un peldaño por debajo.
En tareas de programación, similares resultados (y es más rápido).
108
Stephen De Winter 🤘 retweeted
𝗜 𝗯𝗿𝗼𝗸𝗲 𝗱𝗼𝘄𝗻 𝘁𝗵𝗲 𝗲𝗻𝘁𝗶𝗿𝗲 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁 𝗺𝗲𝗺𝗼𝗿𝘆 𝗹𝗮𝘆𝗲𝗿: 𝗵𝗼𝘄 𝘁𝗼 𝘀𝘁𝗼𝗿𝗲 𝗶𝘁, 𝗳𝗶𝗻𝗱 𝗶𝘁, 𝗮𝗻𝗱 𝗸𝗲𝗲𝗽 𝗶𝘁 𝗰𝗹𝗲𝗮𝗻, 𝗶𝗻 𝗿𝗲𝗮𝗹 𝗰𝗼𝗱𝗲.
Free 30 min course on AI agent harness memories. Whiteboard first, then SQLite vs @mem0ai vs @zep_ai vs LangMem on the same facts.
🌟 Github Repo: github.com/ShenSeanChen/Waku…
💻 Join our Community: seanchen.io
Most people bolt a vector database onto a chatbot and call it memory. The actual layer is four operations:
Store → Retrieve → Consolidate → Retire
@NousResearch Hermes and my own Waku Agent run zero embeddings. Plain text and one SQLite file. Open it, it's yours.
Retire is the one nobody implements. Deleting a fact and invalidating a fact are not the same system.
Full video in the first reply 👇
0:00 - Why every LLM call starts with amnesia
2:31 - Store memory as text, tables, or graphs
5:04 - Find it 4 ways: do nothing, keyword, RAG, Graph RAG
7:29 - Maintain it: add, delete, retire, attribute, reflect
9:47 - Build the whole thing in plain text and SQLite
13:13 - Add a vector store only when you actually need one
14:38 - Row memory vs graph memory vs temporal graph
21:03 - Run the code for all 5
You Can Build Anything. You Can Learn Anything. 💪
Fantástico coste por tarea de deepseek-v4-flash, 46 millones de tokens por apenas 1€.
Estos días mi combinación es:
- gpt-5.6-sol para planificar
- deepseek-v4-flash para ejecutar
1
136














