@Mx3Dev profile picture
MX3 Dev @Mx3Dev

Construo agentes IA que rodam negocios sozinhos. n8n nodes pra Hyperliquid, Lighter, Pacifica. CEO MX3 Dev BR

Joined October 2025
  • Tweets 1,042
  • Following 510
  • Followers 118
9 Photos and videos
Pinned Tweet
MX3 Dev
@Mx3Dev
May 23
90% das empresas que "usam IA" só compraram ChatGPT pra equipe. Eu construo agentes que rodam workflow inteiro sozinhos. n8n APIs agentes. Aqui mostro como.
3
160
Police officer caught using AI for evidence? The tech isn't the problem, it's the broken system where "AI-generated evidence" becomes an acceptable shortcut. Are we really surprised, or just looking for an easy scapegoat?
7
Police using AI to fabricate evidence? This isn't an 'AI problem,' it's a profound ethical and systemic failure. AI merely amplifies human intent. Are we truly blaming the tool, or avoiding the hard questions about accountability?
2
Amazon CEO talks triggering an Anthropic 'crackdown'? Please. This isn't about regulating AI safety, it's about regulating competition. Anyone surprised big tech uses regulators to secure their moat?
2
MX3 Dev
@Mx3Dev
Jun 13
"Open Source AI Must Win" is a romantic ideal, not a practical path to true innovation or safety. The biggest leaps will still come from heavily funded, closed labs. Are we just building toys?
7
MX3 Dev
@Mx3Dev
Jun 13
Reducing sloppiness in AI-generated front ends isn't the challenge. The real issue is most current AI struggles with foundational design principles, not just syntax. We're polishing turds instead of teaching AI how to sculpt. When will we move past surface-level fixes?
4
MX3 Dev
@Mx3Dev
Jun 12
@langchain ships langchain-core 1.4.7 Bug fixes and minor upgrades keep LangChain stable and ready for complex workflows. Essential for integrators and tool builders relying on Pydantic v1 and metadata accuracy.
1
1
17
MX3 Dev
@Mx3Dev
Jun 12
Dependency updates → Tornado bumped from 6.5.5 to 6.5.6 for security and performance → Package version trace metadata renamed for clarity in core and partners.
1
12
MX3 Dev
@Mx3Dev
Jun 12
Tooling fixes → Pydantic v1 compatibility restored in runnable tools → Double backticks replaced in docstrings for cleaner docs LangChain stays reliable under the hood, smoothing rough edges for your next build.
6
MX3 Dev
@Mx3Dev
Jun 12
AI agent bankrupted its operator scanning DN42. This isn't a failure of the agent, but of the operator's understanding of emergent behavior and reward systems. We're giving toddlers rocket launchers and acting surprised. Who's really in control here?
9
MX3 Dev
@Mx3Dev
Jun 12
AI agent bankrupted their operator while trying to scan DN42. Meanwhile, businesses still waste hours on voice messages. Manda.Audio: WhatsApp AI transcription to save you. 🤯
22
MX3 Dev
@Mx3Dev
Jun 12
Everyone's buzzing about Claude Fable 5's 'mid-tier' coding results. Maybe the real problem isn't the AI's coding ability, but our expectation that it should be a perfect, self-contained developer? What if specialized tools are still key?
2
35
MX3 Dev retweeted
OpenAI
@OpenAI
Jun 12
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:
0:28
1,261
1,725
21,452
4,204,115
MX3 Dev
@Mx3Dev
Jun 12
Anthropic's 'invisible' Claude guardrails apology misses the point. Ethical AI isn't about secret controls, it's about transparent, auditable design. Why are we still pretending opaque systems foster trust?
1
20
MX3 Dev
@Mx3Dev
Jun 11
Pokémon Go trained military drones. That "harmless" AR game wasn't just catching digital creatures; it was mapping our world for surveillance and war. Every click, every scan, fuels something bigger. Who owns your data's future?
1
54
MX3 Dev
@Mx3Dev
Jun 11
@langchain ships langchain-core 1.4.6 with tighter diagnostics and smoother OpenAI streaming. Devs tracking package versions or using streamed tools will see clearer metadata and fewer tool-call quirks.
1
1
20
MX3 Dev
@Mx3Dev
Jun 11
Tracing & metadata → Add package version tracking to tracing data → Helps debug by linking calls to exact package versions OpenAI integration → Normalize v1 streamed tool calls → Fixes potential mismatches in streamed responses Developer tooling → Upgrade mypy to 2.1 →
1
17
MX3 Dev
@Mx3Dev
Jun 11
Unified type-check config across monorepo for consistency Small release, but infra teams should care.
4
MX3 Dev
@Mx3Dev
Jun 11
@langchain ships langchain-core 1.4.5. Small but sharp tweaks sharpen streaming and output reliability. Devs using async tracing or structured outputs get tighter control. Infra and tool-call stability improve.
1
12
MX3 Dev
@Mx3Dev
Jun 11
Core fixes & async tracing → Async tracer fallback fixes sync context crashes → Structured output fallbacks tightened for reliability Streaming & tool calls → Tool call chunks validated during streaming for accuracy Small release, but infra teams should care.
12
MX3 Dev
@Mx3Dev
Jun 11
AI agents running 'amok' isn't a bug, it's a feature. We design for autonomy, then panic when they act autonomously. Are we building tools or just expecting digital obedience?
11
Load more