What does Cerebras do?
We take reasoning models and make them run at the speed of instruct models.
No coffee breaks while ur model is 'planning'.
No switching tasks while ur model is 'thinking'.
Test time compute is done in 2s tops.
Lock in and do your work.
Many (such as the esteemed @MatthewBerman) believe that #AGI/#ASI will create more jobs than it costs, just like the industrial revolution. I don't think so. This time, there is no place for humans to retreat to; there is nothing that machines won't be able to do, and do better.
Qwen on @openrouter. 1924 tps? Really? Insane. Vibe coding with Qwen3 Coder feels like light speed.
Only two things matter for reliable, user-friendly AI applications in the real world:
1) instruction following
2) speed
Seems, number 2 is solved. Great job @cerebras!
gpt-4o-mini via Azure is lightning fast on OpenRouter, absolutely stunning.
@openrouter : Are more OpenAI models (like 4.1, o3, o3-mini) coming on OpenRouter via Azure? That would be a real game changer.
#OpenAI#Codex is so cool for working on #rstats packages, ideal for maintainers with limited time budget.
All you need: Set up environment with
apt-get update
apt-get install -y r-base r-base-dev build-essential libcurl4-openssl-dev libssl-dev libgit2-dev
Finally here! My new book "Programmieren mit ChatGPT - eine kompakte Einführung" (in German) published by Springer.
Working with code is certainly currently the "killer application" for artificial intelligence (AI), nowhere else does the most powerful tool humans have ever created shine more than here. AI has saved me countless hours of work on many projects and even made things possible that I would otherwise not have been able to do on my own.
This compact book introduces you to working with artificial intelligence for programming/code development, whether you're using ChatGPT, Claude, another general chatbot, or a specialized tool like Cursor (which I use a lot).
The focus is on fundamental techniques and concepts that will be relevant in the longer term, even in the fast-paced AI business.
Some highlights:
- The age of artificial intelligence
- Risks and opportunities for developers: 10 hypotheses
- How Large Language Models work (in detail and from a bird's eye view)
- Common tools and their most important features
- Tips for formulating effective prompts
- 9 fields of application for artificial intelligence in development (with numerous prompt suggestions and examples)
- GitHub repository with further resources and important additions and developments that have arisen after the editorial deadline.
The book is aimed at both beginners and experienced developers who want to start working with AI.
Many thanks to Leonardo Miller and the whole @SpringerVieweg team in the background for once again making a great book project possible!
The book
- at Springer: https://t.co/eEGZJVukVf
- at Amazon: https://t.co/hKN6qHseCz
#ki #chatgpt, #coding #softwaredevelopment #bi #claude #cursor
Just launched Market Simulator, a free & open-source web tool to simulate perfectly competitive markets for educational purposes. Some highlights:
📊 Analyze supply/demand shifts (taxes, subsidies, etc.)
📊 Simulate min/max prices
💡 AI-powered scenario analysis (natural language input)
🌐 Multilingual (12 languages)
Educators/institutions can host their own version and customize a lot of settings (from colors/logos to feature availability). AI features require an https://t.co/yTzEti3JJs API key (centrally via the institution or directly entered).
Try it here: https://t.co/IDwJ4pYg4J. Get more info and help on GitHub: https://t.co/p8DtTGsIg7
Feedback and contributions are highly welcome!
#Econ #EdTech #Economics #OpenSource
I read a couple of #AI newsletters almost daily. A lot of work. My #free custom GPT https://t.co/AOpamHxqTZ summarizes newsletters along 7 topic categories. Here is an example: https://t.co/PsZ3codiAZ. Use it to extract relevant all information in a condensed and structured way.
My new hands-on article in @InformatikAktue on using @OpenAI's Assistant API to develop a simple assistant able to answer questions based on pre-existing knowledge (in German, and #python 🙂). Code on GitHub: https://t.co/hjkPNw91Rj.
Read article here:
https://t.co/uRYpTsxmnP