2024 was a great year for building general-purpose LLMs (with some special-purpose finetunes for math & code etc). Looks like 2025 is the year of diverging more into (1) reasoning models (for math and code) and (2) agents (for workflow automation). Exciting times!
AGENTS
-
OpenAI’s Operator Agent Code Now Open Source
By
–
Le code de Operator de OpenAI est open source ! https://
github.com/browserbase/st
agehand
… () -
The Evolution of AI from Tool to Autonomous Teammate
By
–
We’re moving from “AI as a tool” → AI as a teammate. • AI that doesn’t just answer questions but takes action.
• Imagine a future where you delegate 90% of your digital tasks to an AI agent. This could be just the start of something huge. -
OpenAI’s New Computer-Using Agent Model Explained
By
–
Operator is powered by OpenAI’s new Computer-Using Agent model (CUA): • Combines GPT-4o (vision) + advanced reasoning models.
• Interacts with websites like a human: clicking buttons, navigating menus, and filling forms.
• No APIs needed it uses the actual front-end. It’s -
OpenAI’s Operator AI Agent Capabilities and Use Cases
By
–
Operator is OpenAI’s new AI agent that can take over tasks autonomously, using a built-in browser to complete actions like:
— God of Prompt (@godofprompt) 24 janvier 2025
• Booking travel
• Making reservations
• Shopping online
It’s the next step in AI moving from "thinking" to doing.pic.twitter.com/zRbxi2QwsQOperator is OpenAI’s new AI agent that can take over tasks autonomously, using a built-in browser to complete actions like: • Booking travel
• Making reservations
• Shopping online It’s the next step in AI moving from "thinking" to doing. -
Operator AI: Building Agentic Workflows Made Simple
By
–
Operator is here, and it’s still in its infancy. I’ve been experimenting with building agentic workflows alongside my team, but the process feels annoying, cumbersome, and slow—especially for impatient non-techies. Then it hits you: in just a few months (weeks?), all that
-
Essential Skills for Modern Multimodal AI Projects
By
–
Surely, the ability to inspire/motivate others is also a very important merit, as well as adaptability to unexpected and unknown thing. I guess all these will also be appreciated in big multimodal and multitask AI projects nowadays
-
OpenAI Operator Tests DeepSeek AI with Question Challenge
By
–
Just got access to OpenAI’s new Operator: so I put it to the test with DeepSeek ⚔️. I asked Operator to keep asking DeepSeek questions until it got one wrong.
— Jiquan Ngiam (@JiquanNgiam) 24 janvier 2025
Operator grilled DeepSeek for 5 solid minutes, until it stumbled on "Which is the longest river?"
DeepSeek went with… pic.twitter.com/tmsDOMdAzJJust got access to OpenAI’s new Operator: so I put it to the test with DeepSeek . I asked Operator to keep asking DeepSeek questions until it got one wrong. Operator grilled DeepSeek for 5 solid minutes, until it stumbled on "Which is the longest river?" DeepSeek went with
-
DeepSeek Operator Passes Question Test With Impressive Accuracy
By
–
Did a test to have Operator ask DeepSeek questions until it got one wrong, and it did so really well. You guys shipped this one well.https://t.co/RBtNZxU4EN
— Jiquan Ngiam (@JiquanNgiam) 24 janvier 2025Did a test to have Operator ask DeepSeek questions until it got one wrong, and it did so really well. You guys shipped this one well.
-
Marek Rosa’s Dev Diary: Space Engineers 2 and AI People
By
–
I just published my weekly dev diary where I share my work on Space Engineers 2 and AI People. https://
blog.marekrosa.org/2025/01/mareks
-dev-diary-january-23-2025.html
… @SpaceEngineersG