If AI large-model talent in China is looking for domestic career opportunities, WeChat is actually worth considering. Not many people know that the WeChat team also has its own large model, WeLM, which is closed-source and has already iterated to version V4.
At this stage, joining established, scaled large-model teams is the more conventional choice. But there are also people who choose to join WeChat’s relatively small-scale large-model effort, which is somewhat like trying to bottom-fish in the stock market—of course, it still depends on individual judgment and ability to evaluate the situation.
Revisit this post in three months.
显示更多
This is really big news. Google introduced the Open Knowledge Format (OKF) - a standardized way to store information in a directory of markdown files. Makes it really easy to make a digital brain that agents can use.
These files can serve as a living wiki. You can give agents the ability to query them or edit them. They can interlink.
Seems to me this could replace Notion or Obsidian. I can think of so many uses for this.
Google's blog post:
An easier to understand explanation is the SPEC.md file:
I gave those two links to Antigravity and asked how we could use it for any of the projects we're working on. It came up with so many ideas. I would imagine Claude Fable 5 would whip up some pretty amazing things based on this system.
Currently creating an OKF library of our pepper garden. It's going to be a fun weekend.
显示更多
We heard you wanted to use Codex rate limit resets on your own time.
Starting today, we’re rolling out the ability to save rate limit resets to use later.
We’re starting Go, Plus, Pro, and Business users with one free reset:
显示更多
A Kobold Story : Trenchcoat Adventurer the classic designed dungeon crawler that changes what you think about the genre now has
The Full 1.0 Release now available on steam for PC. Mac version coming shortly.
Supporting Indies has never been so much fun. Pick up a copy today.
Handcrafted Levels, Human Art, Randomized Loot and difficulty settings for replay-ability.
Lets you play as a cozy game or a hardcore min / max crawler experience.
Bringing back the concept of Shareware the entire 1st chapter is playable as the Demo. Try before you Buy.
Indies change the game. Support your favorite indie today.
显示更多
Dario Amodei, anthropic's CEO, just answered the question everyone is asking. should you still learn to code:
1. coding is going away first. the AI models are doing it already. the broader task of software engineering takes longer but that's going too. if you're learning to code purely for job security, you're learning the wrong thing.
2. even at 5% of the task you're still valuable. if AI does 95% and you do 5%, you become 20 times more productive. comparative advantage is surprisingly powerful even when the gap is massive.
3. the professions with the most runway are human-centered ones. things that mix people, the physical world, and analytical skills together. he uses the radiologist example. the doctor who understands patients and context, not just reads scans.
4. critical thinking might be the most important skill of the next decade. when AI can generate anything, the ability to tell what's real from what's fake becomes rare and valuable. you don't want false beliefs. you don't want to get scammed. that's his actual advice to a 25 year old.
5. AI can make you stupider if you use it carelessly. anthropic ran studies on this. depending on how you use the model, de-skilling in coding is measurable and real. the tool doesn't cause it. carelessness does.
6. the semiconductor space is his pick for a capitalistic win in the next decade. physical world, traditional engineering, direct AI tailwind. not software but chips.
显示更多
Many iPhone users love their cameras. But Apple AI’s ability to improve images has lagged ChatGPT and Google Gemini.
Spurs HC Mitch Johnson on Wemby's ability to dissect defenses:
"Over time, I believe there just becomes a comfortability and that leads familiarity and then confidence."
SAS (1-2) NYK Game 4: Wed, 8:30pm/et, ABC
显示更多
Last year DOJ published a report.
For every 10,000 blacks, 153 violent crimes are committed by a black person each year against a white person.
The same figure for violent crimes committed by whites against a black person: 3.4.
A black person is 50-times more likely to commit a violent crime against a white person than for a white person to commit a violent crime against a black person.
This is an extraordinary difference in the rates of violence committed between these two groups.
We are told that whites are systemically racist.
We are told that whites are successful because they oppress others.
But the facts tell the opposite story.
Blacks are not the ones being victimized.
Whites are.
Western civilization has been successful not because it was oppressive.
It has been successful because whites created, starting thousands of years ago, systems that are impartial and fair.
Today these systems are being dismantled.
Criminals are being released because of their skin color.
Law-abiding citizens are punished because of theirs.
All in service of an ideology that is simply false.
This puts everyone in danger.
显示更多
0
0
1.9K
38.6K
9.3K
转发到社区
KAT on his ability to adjust his role within the team each game to help impact winning!
"Whatever it takes to win, especially when you are in the NBA Finals, I am willing to do."
SAS (0-2) NYK Game 3: Monday, 8:30pm/et, ABC/ESPN
显示更多
🗺️ El fin de los paneles de telemetría aburridos: El mapa de observabilidad Open Source definitivo
Monitorear arquitecturas distribuidas y microservicios usando logs planos o interfaces densas hace que rastrear un error en producción sea una pesadilla.
Maple es un monorrepo de código abierto que transforma la telemetría de tu infraestructura en un mapa de servicios interactivo y animado en tiempo real. Pasa de leer métricas frías a ver exactamente cómo fluyen tus datos.
El ataque a la yugular:
🔄 Service Map en Tiempo Real: Visualiza las dependencias e interacciones vivas entre tus APIs, bases de datos (MySQL, ClickHouse) y colas de mensajería como Kafka.
⚡ Métricas al Vuelo: Haz clic sobre cualquier nodo para auditar instantáneamente latencias p99, tasas de error y rendimiento sin salir del flujo visual.
🔍 Rastreo de Queries: Identifica cuellos de botella de inmediato visualizando cuáles son las consultas más pesadas ejecutándose directamente en tus almacenes de datos.
🛠️ Arquitectura de Vanguardia: Diseñado como un monorrepo que integra ingesta de OpenTelemetry (OTLP), backend nativo basado en Effect y un servidor de código MCP para que tus agentes de IA puedan auditar el sistema.
Deja de adivinar qué microservicio está fallando o rompiendo tu base de datos. Dale visibilidad real al flujo de datos de tu infraestructura de backend.
Enlace al repositorio en los comentarios. Guarda este post en marcadores antes de que tu próximo sistema distribuido entre en cuello de botella 🔖
显示更多