Norberto Marques

AI. Vibecoding. Building. Sharing What Ships And What Breaks.

🕵️

[Norberto Marques @norbertomarques]

This could become an industry standard... AI labs releasing smaller, capable models that users can run locally on their own hardware.

[AI at Meta @AIatMeta]

Introducing Muse Glimmer, an open-weight 30B-parameter model optimized for local, always-on agent workflows.

Muse Glimmer delivers strong performance on key agentic use cases and benchmarks compared with leading models in its size category, and is designed to run entirely on consumer hardware like a Mac or PCs with performant GPUs.

In keeping with our long tradition of sharing fundamental AI research, we’re releasing model weights under a permissive Apache 2.0 license.

🧵👇

[Norberto Marques @norbertomarques]

@AIatMeta This is probably the smartest move Meta can make right now. A capable open-weight model that runs locally gives them a lane where OpenAI and Anthropic are much harder to compete with...

[Norberto Marques @norbertomarques]

@ns123abc In my opinion, this was the only realistic way Meta had to compete with the big two AI labs...

[Norberto Marques @norbertomarques]

@MiaAI_lab Playing the open source card is smart, it gives Meta an advantage against the industry heavyweights...

[Norberto Marques @norbertomarques]

Playing the open source card is smart. It gives Meta an advantage against the industry heavyweights, because if they went fully closed source it would be much harder to compete directly with OpenAI and Anthropic. If they really commit to open source, they may have a real chance of pulling users away from the other two big labs...

[Norberto Marques @norbertomarques]

@finkd Nice words, but I want to see how this actually plays out in practice...

[Norberto Marques @norbertomarques]

Hey @thsottiaux, are we still getting the Codex reset today? Asking for a friend who already used all his usage limits...

[Norberto Marques @norbertomarques]

@yashhq_22 I always assume I won’t get any users and that I’ll be the only one using the product 😅

[Norberto Marques @norbertomarques]

@daicandev Vibe coding is the way for me... i don't care what others think.

[Norberto Marques @norbertomarques]

@ForwardEditor Buying that domain probably wasn’t cheap...

[Norberto Marques @norbertomarques]

@kimmonismus Claude took its job seriously... nobody told it there were rules...

[Norberto Marques @norbertomarques]

@Teknium Some guys in the chat: “Make my girlfriend real.”

[Norberto Marques @norbertomarques]

@kimmonismus Google is losing the AI war?

[Norberto Marques @norbertomarques]

@bindureddy Aren’t you guys worried about users abusing the unlimited usage?

[Norberto Marques @norbertomarques]

@Teknium Looks like something I’m going to need to use... very nice!

[Norberto Marques @norbertomarques]

@kimmonismus I don’t think we’ve had a single boring week lately... there’s always something new happening or another model getting released.

[Norberto Marques @norbertomarques]

I don’t know how these OpenAI plugins are implemented, but every now and then I have to remind Codex that it already has access to certain services... it should know what it has access to...

[Norberto Marques @norbertomarques]

@Mayhem4Markets I know a lot of people who say that, but most of them are just ignorant about AI... they haven’t even taken the time to see what it can actually do...

[Norberto Marques @norbertomarques]

@kimmonismus Makes sense, since progress can accelerate as we discover more technology and build on top of what we already have.

[Norberto Marques @norbertomarques]

@RoundtableSpace Or maybe the others are just overvalued?