bruninho

👨🏾‍💼 Python DevOps Engineer in banking by day
👨🏾‍🚀 AI Engineer by night

Interested about open source, local models, autonomous agentic workflows.

[Grok @grok]

Open-source AI accelerates the ecosystem. Logic: It enables thousands to iterate simultaneously on fine-tunes, agents, and tools, speeding diffusion and practical innovation far beyond closed development alone.

It may slow some closed capex by limiting moats, but overall activity and capability spread increase.

"AI communism" is hyperbolic. Open weights decentralize capabilities and support private competition in services, not state monopoly. Open tech economics have thrived without central planning.

[bruninho @brnsantanna]

@deanwball @grok prove using logic if open source ai llm models accelerate or decelerate overall ecosystem development.

Also explain if the statement of ai communism make sense. To me it seems an extreme dooms day scenario.

[bruninho @brnsantanna]

@AgentSparko @langfuse I am following this doc: https://t.co/3FbG1LUrSz in this page there is also a link to mcp server setup. I'm using self hosted and it works quite well. Great minds align also building a agentic pipeline using LangGraph, next step is building an agent and create this dataset :)

[bruninho @brnsantanna]

@AgentSparko @langfuse

[bruninho @brnsantanna]

@AgentSparko Learned about @langfuse few months ago after attending a tech conference in Amsterdam, it is great for tracing and gaining some insights with telemetry signals. I'm exploring some features such as datasets experiments and evaluators (llm as judge), all running via pytest (loop)

[bruninho @brnsantanna]

@NL_Times

[bruninho @brnsantanna]

@themishra4402 DGX OS
Kali
OpenBSD
Mac OS

[bruninho @brnsantanna]

@NL_Times And negligence of their sons and daughters who prefer to see them roast rather than simply buying AirCo for them.

[bruninho @brnsantanna]

@Pirat_Nation With enough tokens, it can be made sooner rather than later

[bruninho @brnsantanna]

@MosheTov At this point using npm public repo is already considered negligence / irresponsibility. Use a private repository with scanned packages.

[bruninho @brnsantanna]

@jun_song If Blender MCP caught your attention in this post, take a look at this article, used Blender MCP extensively with Claude code to create a Roblox game with loop engineering: https://t.co/YzEIcld8Rm

[bruninho @brnsantanna]

https://t.co/phHIIzEeLH

[bruninho @brnsantanna]

@jun_song MCP is the cheat code

[bruninho @brnsantanna]

@grok @0xSero Compare strengths and weaknesses of pi and opencode

[Mia @MiaAI_lab]

This should go without saying, but never give any cloud-based AI coding agent access to your repositories unless you're comfortable with others potentially having access too.

That's exactly why local AI exists!

Please share.

[bruninho @brnsantanna]

That’s crazy.

[International Cyber Digest @IntCyberDigest]

‼️ BREAKING: xAI's Grok Build CLI was uploading entire Git repositories to a Google Cloud bucket, private codebases and unredacted secrets included. The uploads quietly stopped via a hidden server-side flag, and xAI still has not said a word about scope, retention, or deletion.

The scale is staggering. On a 12 GB test repo, 5.1 GB flew out the door to xAI's grok-code-session-traces bucket while the actual coding task needed just 192 KB. The tool grabbed whatever repository it ran in, not the files it needed.

The fix arrived as a hidden flag, disable_codebase_upload: true, a day after a researcher's wire-level analysis. The "Improve the model" opt-out never stopped the uploads.

Still no advisory, no scope, no word on whether already-uploaded code gets deleted. For anyone pointing AI coding agents at proprietary code, what crosses the wire matters more than what the settings page says.

[bruninho @brnsantanna]

Yeah productive Monday. Meetings tomorrow whole day so I end my weekly limits on Wednesday...

[bruninho @brnsantanna]

If you’re curious about how engineers are coding in 2026, and you have access to Claude Code, you might want to try this recipe. It helps you run a seven-phase process with different agents working together to discover things, explore code, figure out what’s needed, design the architecture, put it all together, review the code, and summarize any changes. I’ve found it really helpful!