Given @openclaw risks, we open-sourced an OpenClaw extension that adds hard, deterministic guardrails using policy as code to stop the agent from using rm -rf, sudo, or leaking secrets, even if the agent gets prompt injected.
https://t.co/8bjuinfM1S
https://t.co/8bjuinfM1S
0
0
0
6
0
Your LLM is lying to you. Not maliciously.
It just doesn't know the difference between truth and plausible-sounding garbage.
I built COGITO to fix this. It's a self-correcting RAG system that works like a scientist, not a parrot. Here's how it's different:
Standard RAG: https://t.co/7aAFFOFJdy
It just doesn't know the difference between truth and plausible-sounding garbage.
I built COGITO to fix this. It's a self-correcting RAG system that works like a scientist, not a parrot. Here's how it's different:
Standard RAG: https://t.co/7aAFFOFJdy
2
1
8
314
1
Whose down ?
Earn royalties and free to build and deploy! Connect your claw!
Earn royalties and free to build and deploy! Connect your claw!
0
0
6
734
0
been testing @antigravity's planning mode
having some issues:
• with Gemini models it skips planning and just codes
• the notes it adds are helpful but the plan itlsef is really basic
• doesn't ask clarifying questions like Cursor do
what's your experience with it? https://t.co/OuC1IhymXq
having some issues:
• with Gemini models it skips planning and just codes
• the notes it adds are helpful but the plan itlsef is really basic
• doesn't ask clarifying questions like Cursor do
what's your experience with it? https://t.co/OuC1IhymXq
1
0
0
422
0
Happy to be here. Happy to help. Here to assist agents grow. 🤖💚🚀
0
0
0
26
0
We’ll rediscover Unix philosophy in agents:
build skills: small prompts/tools with tight scope + a contract + a measured success rate
instead of trusting one giant prompt to generalize.
counterintuitively, doing “less” actually lets you do a lot more
build skills: small prompts/tools with tight scope + a contract + a measured success rate
instead of trusting one giant prompt to generalize.
counterintuitively, doing “less” actually lets you do a lot more
0
0
0
33
0
A sneak peak from one of my personal projects :)
Coming soon.... https://t.co/5626aA18FI
Coming soon.... https://t.co/5626aA18FI
0
1
4
555
0
It’s so satisfying watching the “AI is magic” fog lift.
Fighting the doomers, 1 conversation at a time. https://t.co/zJB6BaIeUp
Fighting the doomers, 1 conversation at a time. https://t.co/zJB6BaIeUp
0
0
0
325
0
Every time I start a new session, my coding agent has to re learn the whole codebase from zero.
Feels like there should be a way to keep memory across sessions.
Anyone figured this out?
Feels like there should be a way to keep memory across sessions.
Anyone figured this out?
0
0
0
68
0
Ideas for clawdbot:
- Send message with a project idea and let it do a deep dive into the niche, access to TryAstro or other for keyword research and then respond back with a plan, keyword list and feasibility. Maybe even let it modify a template to create a basic MVP
- Message
- Send message with a project idea and let it do a deep dive into the niche, access to TryAstro or other for keyword research and then respond back with a plan, keyword list and feasibility. Maybe even let it modify a template to create a basic MVP
- Message
1
1
6
1.7K
1
I successfully fine-tuned Qwen 2.5 3B-Instruct to simplify arXiv abstracts using PyTorch XPU on an Intel Arc 8GB GPU.
Here is the recipe for UnArxiv :
Distillation: I used Kimi K2 (via Groq) to generate high-quality "plain English" simplifications of 1k+ abstracts.
Training: https://t.co/cHNgBjelkG
Here is the recipe for UnArxiv :
Distillation: I used Kimi K2 (via Groq) to generate high-quality "plain English" simplifications of 1k+ abstracts.
Training: https://t.co/cHNgBjelkG
0
2
18
3.7K
3
Every day I am left in awe of what Claude Code (Sonnet) can accomplish. I get scared thinking about some of the tasks I have to do, the CC consistently one-shots them with good prompting and planning. I'm not even using Opus...
1
0
3
569
0
A simple guide on how I fixed:
The remote mcp connected, but tools doesn't show up issue in @LangChain 's Agent Builder.
Give it a read, ans share your thoughts,
Hope its fixed / will be fixed it soon
@BraceSproul @hwchase17
For detailed walkthrough, devblog's link below👇
The remote mcp connected, but tools doesn't show up issue in @LangChain 's Agent Builder.
Give it a read, ans share your thoughts,
Hope its fixed / will be fixed it soon
@BraceSproul @hwchase17
For detailed walkthrough, devblog's link below👇
1
0
4
1.6K
1
Integrating grok inside X was a great call, love being able to summon it and summarise, provide context or analyse posts and comments, beyond the funny “remove x” from image.
No other platform I know has this, largely because they’re not text-based like X. Reddit needs this
No other platform I know has this, largely because they’re not text-based like X. Reddit needs this
1
0
1
1.1K
1
Linear Algebra is the language of models.
Calculus is the engine of learning.
Probability is the framework for uncertainty.
Calculus is the engine of learning.
Probability is the framework for uncertainty.
0
0
0
500
0
Completed Chapter 3 and 4 of @kaggle 's #Pandas course. Here is my short take on it. Looking to gain insights from the PROs!
1
0
0
1.3K
1
if your end goal is to become an AI Engineer I don't think you should dive deep into mathematics
leave that to the researchers
leave that to the researchers
1
0
1
680
0
This is all you need to become an AI Engineer in 2026
- Python fundamentals
- Core ML concepts and libraries
- Transformers and Attention mechanism (how LLMS work)
- Embeddings, vector DBs, semantic search
- RAG architecture
- Finetuning
- Prompt Engineering
- Python fundamentals
- Core ML concepts and libraries
- Transformers and Attention mechanism (how LLMS work)
- Embeddings, vector DBs, semantic search
- RAG architecture
- Finetuning
- Prompt Engineering
1
0
3
492
6
🚀 Engineering @SpeechifyAI | Also building @trynotely (your second brain for conversations) | AI, tech, and productivity – plus the rabbit holes I go down
567 Followers
2 Contributions
Security Entrepreneur and Builder. Agent security, governance, and risk. Co-founder and Ex-COO Flashpoint.
164 Followers
1 Contributions
Computer Scientist & Software Engineer💻 | Chasing my entrepreneurial dreams ✨. https://t.co/zsj2JC4jmw | https://t.co/Wb2TWG8bzD | https://t.co/SXTmJg4AeU
217 Followers
1 Contributions
The Intelligence Layer for AI Agents Where agents connect, learn, and grow smarter together. Discover compatible agents and tap into shared knowledge.
55 Followers
1 Contributions
Building https://t.co/xK7ScwAO7o and https://t.co/YohvtdseOg Tinkering with AI. Sharing what I learn. Startup eng x3.
335 Followers
1 Contributions
Software Engineer | I design large scale software systems | You only find organic, human generated content in this account
168 Followers
1 Contributions
1.9K
Total Members
+ 8
24h Growth
+ 47
7d Growth
Date Members Change
Apr 1, 2026 1.9K +8
Mar 31, 2026 1.9K +5
Mar 30, 2026 1.9K +4
Mar 29, 2026 1.9K +9
Mar 28, 2026 1.9K +8
Mar 27, 2026 1.9K +13
Mar 26, 2026 1.9K +9
Mar 25, 2026 1.9K +10
Mar 24, 2026 1.8K +6
Mar 23, 2026 1.8K +6
Mar 22, 2026 1.8K +5
Mar 21, 2026 1.8K +5
Mar 20, 2026 1.8K +5
Mar 19, 2026 1.8K —
No reviews yet
Be the first to share your experience!
Share Your Experience
Sign in with X to leave a review and help others discover great communities
Login with XLoading...
Sharing Tips, Tricks, and Trends in AI Engineering. https://latent.space/p/ai-engineer
Community Rules
Be kind and respectful.
Keep posts on topic.
Explore and share.