Just shipped stup
Scaffold production-ready Python projects in seconds
Every new project used to mean the same setup:
uv init, pin Python, create venv, wire requirements, configure folders… and a lot
pip install stup
github.com/muhdaliyan/s...
Want to unlock your inner genius? This series on Polymaths explore a non-zero-sum approach to innovation. The third and final part dropped today! #Polymath #Genius #Descartes #Newton #Musk #CrossPollination #TransferableSkills #Fails #Success #Obsidian
Anthropic just killed OpenClaw.
Using these 4 native features inside Claude
1. Channels
2. Remote Control
3. Dispatch
4. Scheduling
What do you think? Are you gonna shift from OpenClaw to Anthropic?
#openclaw #claude #anthropic #agenticai
3/4
It feels like a shift from:
asking AI for answers
to
letting AI simulate possibilities and social dynamics
And the fact that it’s open-source makes it even more exciting you can explore how it works under the hood and even build on top of it.
2/4
Here’s what’s really cool about it:
1️⃣ Agents act like real users they don’t just “talk,” they post, quote, comment, like each other’s posts, follow and interact.
2️⃣There’s memory + knowledge graphs.
3️⃣Results come from the emergent behavior of these interactions.
1/4
Realistically, I tried it myself, and it actually feels different. For the future, I can see how it could help people like me I’ve built a lot of projects that didn’t work out, and something like this could really help simulate outcomes before investing time and effort.
I just found this and honestly, no one is talking about.
It’s called MiroFish🐟, an open-source project by a Chinese student.
It’s more like a simulation engine where multiple AI agents interact, argue, and behave like real people to explore scenarios and generate outcomes.
read thread...
2/2
6. It’s fully open source, so you can check it out, learn from it, and use it in your projects.
7. Can operate via Telegram
Source code: github.com/muhdaliyan/z...
Pypl: pypi.org/project/zeer/
Just prompting, learning, experimenting, and building with hashtag AI.
1/2
It can:
1. Read and list your directories, files, and contents
2. Write or change files, run code, and manage tasks in the background
3. Start and manage dev servers so you can test live
4. Add custom skills to extend its behaviour
5. Explore multiple AI providers and tool-calling workflows
I built Zeer
An open-source agentic AI CLI. Think of it like a combination of Claude Code and OpenClaw. Built for learning, experimenting with AI workflows, calling tools/skills, and trying out multiple providers.
Try:
pip install zeer
#opensource #buildinpublic #AgenticAI #AI
OpenAI + Oracle = $300B GPU deal.
Larry makes $101B in a day and tops the rich list.
Fun fact: neither Sam nor Larry has $300B worth of GPUs.
We must be in a simulation
#larry #SamAltman #oracle
Here’s why it felt so different from Cursor:
✅ Stays on track, doesn’t drift
✅ Rarely throws errors
✅ When errors happen, fixed in 1–3 prompts
✅ Tries different techniques until it works
Honestly feels way smoother for debugging & problem-solving than Cursor.
I got early access to Kiro AI by Amazon and burned through all 200 bonus credits of Vibe and Spec in a single day 😅
Asked the team for more → they reset my bonus credits… and I finished them in just 3 hours.
Check comments for how is better then cusror
#AI #Coding #kiro #amazon #vibecoding
2/2
This is just the beginning, and I’m excited about how AI can change the way we create, share, and scale content online.
✨ Would love to hear your thoughts — how would you use an AI news agent like this?
1/2
To test it, I created a page called TechGlich where I’ve already published 120+ posts (Insta + facebook).
Even with only 80 followers, TechGlitch reached 10,000+ views this month — all completely automated 🤯
I just vibe coded an AI Agent!
I’ve been working with Cursor and ChatGPT on an AI-powered news publishing agent that can:
✅ Fetch news
✅ Rewrite
✅ Generate images
✅ review, edit, publish
Check comments for more.
#AI #vibecoding #cursor #chatgpt #agent #aiagent #coding
Hello bluesky fam 👋
What are u cooking?
Just finished testing out a new feature we’ve been building.
1. Generate full landing pages from a simple prompt.
2. Edit the landing page live, right in your browser.
3. Copy the code instantly.
4. Coming soon: Deploy directly from Taskra — no setup needed.
#videcoading #saas #ai #buildipublic
Reality called - I hit decline.
#vibecoding
Just checked out LLaMA 4—initial vibes: solid upgrade. Feels sharper, more grounded. Curious to see how it holds up in the wild.
Anyone else diving in?
#llama4 #buildinpublic #ai
Watching AI move from reactive tools to agentic systems feels like watching a kid learn to ride a bike wobbly, exciting, and a little terrifying. These systems aren’t just answering questions—they’re setting goals, making plans, and taking action. Wild times ahead.
#AgenticAI #buildinpublic
Day 1: Excited to bring my ideas to Life.
.
.
.
Day 7: "what is vibe coding?" 😭
#vibecoding #ai
vibe coding some days is just… staring at the screen,
waiting for inspiration that never shows up.
AI's throwing ideas, none of them click.
it's not always flow — sometimes it's just fog.
#vibecoding #devlife #realtalk
honestly? not really.
feels fun in the moment, but the burnout creeps in when vibes run out and structure’s missing.
trying to find that balance now. you?
Vibe coding these days feels like…
half coffee, half chaos.
trusting the AI, ignoring half its suggestions,
figuring out the rest by vibes and instinct.
somehow… it still works.
#ai #vibecoding #buildinpublic
Quantum AI blows my mind. The idea that AI models could run on quantum processors, solving problems in seconds that would take regular computers years, feels like sci-fi—but it's happening. We’re on the edge of something huge. Who else is keeping an eye on this?
#QuantumAI #FutureIsNow
Ideas are worthless without execution.
AI is no different—it's not about what it can do, but what we make it do.
#AI #Tech #ideas
I listed and here are some suggestions.
1. Add a "Continue with Google" option.
2. Remove the image URL field; instead, provide an image upload option or make it optional.
3. The available categories are insufficient. When selecting "Other," there should be a field to add a custom category.
