Minha gente, fiz um site aqui pra visualizar os casos de COVID-19 por munícipio: https://t.co/Kwk3GxBZXR
Completamente open source, subindo nos ombros de gigantes como next.js (@zeithq), chakra-ui (do @thesegunadebayo), vx, d3.js entre outros ❤️
Just wanted to defend the term "Software Factory" for a second.
First, the caveats:
It's bullshit. It sounds fancier than it really is.
It's a bandwagon. People are using it to get you and their VC's hyped.
But it's also referring to a real change.
Earlier this year, most SOTA teams were using agents to work on the product. Shipping features, fixing bugs.
But now, they're building automations to help agents work on the product for them. They've moved up a layer.
What do we call the thing they're building? Loops? Workflows? Some other generic label?
"Software Factory" is as good a term as any other.
- "We're not working on the product, we're working on the factory that builds the product."
- "We're not product engineers, we're factory engineers."
I like it. I'll continue using it.
I have thoughts on the recent ai bot craze:
- Grok Bot
- OpenAI Dot
- Muse
- OpenMuse
- Cue
- Fo
- Instinct
- Gemini Spark
- Claude (agent mode)
- Microsoft Copilot Autopilot / Scout
They're cool. It's like hiring an employee, and firing them will be likewise painful unless you set yourself up right.
Each one of these products (and the others inevitably on their way) are going to do their best to make theirs the winner. But the fact is that they will continue to leap frog each other in capabilities. Not to mention some will be locked down to specific models.
There will be plenty of reasons you want to fire one and hire another. Or even hire multiple of them. But the thing is, they all want to drive their claws so deep into the flesh of your business (to be useful of course), that it will be a huge chore to get rid of them.
The integrations, automations, memories, secret tokens, etc etc etc. All of these will be locked into the assistant and they are not incentivized to make it easy for you to transfer these things to your "new hire."
In fact, they're incentivized to do the opposite. To become so core to the way you do business that you could never replace them.
This is why I think now more than ever https://t.co/HBYOHHeqJi is filling in a critical gap. Instead of your agent putting everything inside its own home for the work it does for you, connect it to Kody and put it there. Then when you hire another agent, it can work off the same platform.
I think a lot of people are going to feel this pain in the coming months. When they do, Kody will be ready!
you now have the ability to
- play with every possible solution to a problem
- refactor everything when you think of better patterns
so many people complaining about the code the LLMs produce, if you're not producing the best software of your life right now something is wrong
I've been using @kentcdodds kody mcp and this is possibly the best usage I've seen from an mcp. you can set it up once in any of your harnesses and all your integrations become available safely through kody. no need to reconfigure any oauth app again or share secrets yet again
Take a look at your favourite skill.
Go on, take a look.
Check for lines like:
- "Make the commit message very detailed"
- "Be thorough"
- "Make the implementation easy to read"
What do these lines have in common?
They're no-ops. They do nothing to change the agent's behavior.
Agents always write good commit messages. They always try to be thorough. They always try to create easy-to-read implementations.
Try removing them. Does the output change? No? Then the line was a no-op.
Agent-authored skills are LITTERED with no-ops. Entire paragraphs and sections of text that does nothing to change the agents' behavior.
No-ops make skills harder to evaluate, harder to maintain, and burn tokens uselessly.
Learn to spot them, test them, and remove them. Your skills will thank you.
Announcing mattpocock/skills v1
- Achieved a 63% reduction in token cost for skill descriptions
- Split skills into model-invocable and user-invocable skills, adding /codebase-design, /domain-modeling, and /grilling
- (UPDATED) /writing-great-skills - rewritten from the ground up, encoding my skill-writing best practices
- (UPDATED) /diagnose -> /diagnosing-bugs - now model-invocable, awesome for fixing hard bugs
- (NEW) /ask-matt: a router skill that teaches you how all the engineering skills work together
/teach is live
Learn anything, from rubik's cube to vocal harmonies to software fundamentals.
npx skills add mattpocock/skills --skill teach
Best skill I've ever built, video coming soon
https://t.co/GAv1jBFwsX
@ebrionesm@xNavalha This is the classic confusion of damage and might. Might is not reduced when the unit has damage on it. A 7 might unit with 6 damage on it is still a 7 might unit.
@adamwathan It’s about time we all agree to let go of Claude Code. A harness for developers must be open source, otherwise it serves the provider and not the developer.
@daniloromano@matheuslf Rapaz, apertado essa conexão. Eu procuraria a companhia aérea só pra ter certeza que ela estão cientes da sua situação. Eu acho que dá, geralmente a staff da própria companhia aérea já te encaminha direto pra conexão.
4 of the most confusing terms in AI, defined:
Model: a blob of parameters, written during training. Does next-token prediction and nothing else. Stateless.
Harness: everything around the model that turns it into an agent: tools, system prompt, context window management, etc.
Environment: the world the agent acts on. Anything outside the harness that the agent perceives and acts on via tools.
Agent: a model, harnessed, in an environment.
---
Opus is a model.
Claude Code and Claude Web are different agents, because their harnesses differ - even though the models are the same.
The file system is an environment. MCP servers add tools to the environment.