Csaba Daniel @danielcsaba

Debrecen Joined March 2010

Tweets

357
Followers

39
Following

114
Likes

13

Gergely Orosz @GergelyOrosz

3 weeks ago

I am surprised to see more and more startups (mostly AI ones) equate an "extreme performance culture" with "we stay very late in the office, work on weekends, and many of us live where we work." Speaking of myself, I have done mediocre work when working nonstop for a long time..

174 93 3K 359K 335

Akshay 🚀 @akshay_pachaar

3 weeks ago

Google just dropped a new LLM! You can run it locally on just 0.5 GB RAM. Let's fine-tune this on our own data (100% locally):

188 1K 15K 1.9M 21K

François Chollet @fchollet

3 weeks ago

Interesting findings from this post: 1. It should be obvious to anyone who has interacted with LLMs before that the writing style of the tweet is a conspicuous caricature of AI slop (e.g. em dashes, the "it's not... it's..." construction, rambling, florid prose, etc.). Yet, many…

45 24 516 48K 65

Paul Couvert @itsPaulAi

3 weeks ago

Gemini released the best AI tools for learning You can now learn anything by using: - Step-by-step guided learning - Generate unlimited quizzes and flashcards - Understand complex things w/ visuals and YT videos Suitable for any topic you want to explore. More below

39 103 765 83K 609

Download Video

Austen Allred @Austen

4 weeks ago

To be fair some companies spend 100x that to learn the same lesson

369 481 14K 874K 2K

Download Image

Akshay 🚀 @akshay_pachaar

4 weeks ago

Google just open-sourced LangExtract Python library! It uses LLMs to extract entities, attributes, and relations—with exact source grounding—from unstructured documents. Flexible LLM support (Gemini, OpenAI, Ollama) 100% open-source.

20 226 1K 118K 2K

Download Video

Imrat @imrat

4 weeks ago

Claude Code as your Devops agent. The new background jobs (like cursors shared terminals) make this way easier. 1. start claude in tmux session 2. ask it to create background process to tail your server logs and give you a summary 3. then ask it to run a second process that…

25 78 1K 165K 2K

Download Image

ThePrimeagen @ThePrimeagen

4 weeks ago

there is literally no universe that this is true 10k lines and its not bug filled crap? ok Lex Luthor, its time to step away from the keys

469 204 10K 494K 484

Download Image

Avi Chawla @_avichawla

a month ago

12 MCP, RAG, and Agents cheat sheets for AI engineers (with visuals):

15 140 1K 259K 3K

Sehaj Singh @heysehajsingh

a month ago

Grok 4 is dangerously good. But 99% of people are sleeping on what it can actually do. I’ve used it to build apps, generate content, automate deep research, and more. Here are 10 ways to use Grok 4 that feel like cheating:

35 79 461 216K 897

Download Image

Csaba Daniel @danielcsaba

a month ago

Vibe coding isn’t the problem — your process is. How to ship faster using AI without losing your mind: → Pick a smart stack → Write a 1-page PRD → Use Git → Give AI working examples → Stop looping — start fresh with better info Speed ≠ sloppy. Clarity wins.

1 0 1 26 0

Shashank | 𝔽rAI @0xshai

a month ago

About time everyone starts realizing the inevitable @FractionAI_xyz : “There are hundreds of agent demos, frameworks, and “autonomous” systems out there, but how do we compare them? Most current benchmarks are static, narrow, and disconnected from how agents behave in real-world…

327 475 750 12K 18

Gergely Orosz @GergelyOrosz

a month ago

Incredible how many vibe coding tool startups are launching left and right. It's become so trivial to build one of these (you can do it in a few hundred lines of code on top of a cutting-edge LLM like Claude/Gemini/ChatGPT) that it's no wonder. It's becoming a marketing game

72 17 501 62K 195

elvis @omarsar0

a month ago

Hierarchical Reasoning Model This is one of the most interesting ideas on reasoning I've read in the past couple of months. It uses a recurrent architecture for impressive hierarchical reasoning. Here are my notes:

44 285 2K 255K 2K

Download Image

Christopher Mims 🤌 @mims

a month ago

The AI infrastructure build-out is so gigantic that in the past 6 months, it contributed more to the growth of the U.S. economy than /all of consumer spending/ The 'magnificent 7' spent more than $100 billion on data centers and the like in the past three months *alone* 1/🧵

164 1K 4K 1.1M 2K

Download Image

Csaba Daniel @danielcsaba

a month ago

Multi-agent systems often fail due to unclear instructions, poor coordination, or weak result checks. Fixing prompts and agent roles led to big performance gains.

0 0 1 8 0

Csaba Daniel @danielcsaba

a month ago

Windsurf isn’t your intern. It’s your mirror. Feed it blur = get blur. 💡 Bad: “Refactor this.” ⚙️ Better: “Refactor this for readability and maintainability — assume a senior backend engineer is doing code review for long-term production use.” Specific in, powerful out.

1 0 1 26 0

DeepLearning.AI @DeepLearningAI

a month ago

This week, in The Batch, Andrew Ng announced Buildathon. This competitive event challenges AI builders to rapidly engineer new software projects. Plus: 💻 Google secures Windsurf tech license 🌙 Moonshot releases Kimi K2 📚 EU issues guidelines for AI Act compliance 🔁…

1 10 50 6K 11

Download Image

Together AI @togethercompute

a month ago

Run the best instruct model (non-reasoning) in the world directly on the Hugging Face model card! Powered by Together AI inference.

Together AI @togethercompute

2 months ago

Run the best instruct model (non-reasoning) in the world directly on the Hugging Face model card! Powered by Together AI inference. https://t.co/1WMxb82sEA

3 3 61 19K 10

4 7 36 13K 6

Download Image

Gergely Orosz @GergelyOrosz

2 months ago

I'm calling it now: A new wave of startups will launch that will market themselves as: "The first {X product} designed to be used by AI agents." Why? Easy sell to investors, large TAM forecasted, and can claim to have no competitors (yet)