Introducing the next generation: Claude Opus 4 and Claude Sonnet 4.
Claude Opus 4 is our most powerful model yet, and the world’s best coding model.
Claude Sonnet 4 is a significant upgrade from its predecessor, delivering superior coding and reasoning.
Love this paper on AI agents vs Agentic AI. It’s summarized the difference between working with one highly specialized person vs having a team of committed people working towards the same goal! #ai#aiagent#agenticai
AI Agents vs. Agentic AI
Interesting paper summarizing distinctions between AI Agents and Agentic AI.
It also talks about the key ideas, solutions, and the future.
Here are my notes:
Using SWE-1 on Windsurf to code and create web apps is like enslaving a team of #software engineers working for you…yes, it still does mess things up and forgets some rules, but it’s just so much faster! @windsurf_ai thank you for creating an addictive software!
While I am still impressed with #Gemini 2.5 pro, I get a lot of errors when I ask it to build an app. Some are clear #coding mistakes, like adding closing HTML tags inside a .js file with no HTML #code, or breaking while generating an answer. Ok, it's a Preview model, yet...
How we "guessed" the Pope using network science: inside the cardinal network. A study by me, Beppe Soda and Alessandro Iorio. Article: https://t.co/xQ0fTmpVxb @Unibocconi
@giffmana Reasoning not only uses more tokens but it uses more computational resources. As more tokens are generated internally the demand increases, and so the cost. Yet, it’s cheaper than o3!
This is how Gemini helped my 5yo daughter fall asleep. Every night she wants me to tell her what she is going to dream about, and I was running short of ideas…so, hiding my phone with #Gemini open and the mic on, I let her tell me the topic she picked…and Gemini returned this..
The latest Gemini 2.5 is a pure blast. It seems fast, accurate, and able to tackle full blown coding tasks better than any of its predecessors. API cost too is better, cheaper than o3 for 1M tokens. It reasons, it codes…and it just works!!! Hats off to you Google!
This is pretty cool. One of the biggest challenges for me has always been the hardware and cost limitations .
Considering how Qwen3 performs, this will be my next try
You can now fine-tune Qwen3 (14B) for free with our notebook!
Unsloth makes Qwen3 finetuning 2x faster with 70% less VRAM and 8x longer context lengths - with no accuracy loss.
Guide: https://t.co/KfgCo4iQkL
GitHub: https://t.co/2kXqhhvLsb
Colab: https://t.co/S6y9vCuIx5
Over the next 20 years, the likelihood of age- and lifestyle-related diseases is expected to increase by about 20 percent. However, chronic respiratory diseases are likely to drop. #ChartOfTheDay https://t.co/QzrG4yENTB
obtaining #grantfunding for #research is like rolling down a 2000ft hill full of rocks with no protection. You might land on a patch with soft leaves or grass once in awhile, but otherwise it’s a real pain and on the way down you ask yourself WTH you are doing there!
The Immuno-Oncology Translational Network (IOTN) of @theNCI is a Cancer Moonshot Initiative to improve immunotherapy outcomes for both “hot” and “cold” cancers - and to prevent cancers before they occur. Read more at https://t.co/OstP8iyT6i and https://t.co/Hwk1z0Luy5