from what I can tell, ChatGPT agent mode = browser control (operator) + tool use* + web search + code interpreter** in a deep research type loop. *tool use has been available in deep research for a bit, but limited to read (i believe), and if I had to guess, it would stay that
AGENTS
-
SLMs and Mini-Agents Power Reliable Agentic AI Workflows
By
–
At #DataAISummit, @LaminiAI CEO @realSharonZhou explored how SLMs and mini-agents can power reliable agentic workflows. Learn about memory RAG, MoME-based memory tuning, and real-world use cases like text-to-SQL and code analysis: https://
databricks.com/dataaisummit/s
ession/composing-high-accuracy-ai-systems-slms-and-mini-agents?utm_source=twitter&utm_medium=organic-social&utm_scid=701vp00000d7qrxia3
… -
Manus AI surpasses OpenAI with a new feature
By
–
Manus AI released "Data Visualization" feature, where users can upload raw data and let Manus analyse and generate comprehensive reports.
— 🚨 AI News | TestingCatalog (@testingcatalog) 17 juillet 2025
Overtaking OpenAI close to the final turn 👀 https://t.co/Sw9dM2nLM3 pic.twitter.com/mDMluy02C3Manus AI released "Data Visualization" feature, where users can upload raw data and let Manus analyse and generate comprehensive reports. Overtaking OpenAI close to the final turn.
-
WebSailor paper uses agentic RL post-training to boost Deep Research scores
By
–
Recent WebSailor paper by Alibaba-NLP, shows how to post-train models for Deep Research – good insights in there, about creating a dataset then training recipe. I particularly like how the agentic RL at the end of post-training improves scores by ~4 p.p. across the board: RL +
-
Rork Agent Margins Improve With Mobile App Launch
By
–
This basically becomes a forcing function for Rork to become the best agent in the space. The better the agent gets, the better their margins get 🙂 Plus they launched a lot of other cool stuff, including a mobile app that builds mobile apps! Scroll the thread to see.
-
Rork launches free AI agent error correction standard
By
–
Rork is the first agent to implement a new standard: when your AI agent makes a mistake, it's free to fix it!
— Matt Shumer (@mattshumer_) 17 juillet 2025
This is really powerful for two reasons:
1. Users worry less about wasting $ on failed outputs
2. Rork now has an insane incentive to improve their agent! https://t.co/Keurdkb7qzRork is the first agent to implement a new standard: when your AI agent makes a mistake, it's free to fix it! This is really powerful for two reasons: 1. Users worry less about wasting $ on failed outputs
2. Rork now has an insane incentive to improve their agent! -
Memory-Native Language Models: The Missing Architecture Innovation
By
–
again, the AI labs are obsessed with building reasoning-native language models when they need to be building *memory-native* language models – this is possible (the techniques exist)
– no one has done it yet (no popular LLM has a built in memory module)
– door = wide open -
OpenAI Agent Mode combines Operator and Deep Research capabilities
By
–
Pues OpenAI confirma el contenido del streaming de hoy y pinta interesante. Es lo que algunas filtraciones han identificado como el Agent Mode, la combinación de Operator (uso agéntico del navegador) con Deep Research (razonamiento de largo alcance) para tareas más complejas.
-
FakePsyho Wins AtCoder World Tour Finals 2025 Heuristic
By
–
Congrats to @FakePsyho for winning AtCoder World Tour Finals 2025 Heuristic Humanity has prevailed (for now!) Thanks OpenAI for sponsoring #AWTF2025, and getting #2 on this grand challenge. Proud of @SakanaAILabs & @AtCoder
’s ALE-Agent for reaching #5, on a shoestring budget!