Please turn JavaScript on
TUN icon

TUN

Want to know the latest news and articles posted on TUN?

Then subscribe to their feed now! You can receive their updates by email, via mobile or on your personal news page on this website.

See what they recently published below.

Website title: TUN Helps Students - TUN

Is this your feed? Claim it!

Publisher:  Unclaimed!
Message frequency:  1.12 / day

Message History

OpenAI published findings from its ChatGPT for Academic Researchers pilot, showing scientists estimated a near-doubling of productivity with AI assistance. The lab is now expanding the free-access program to 200,000 researchers worldwide, with applications opening Nov. 23.

The University Network

OpenAI on Nov.16 released a detailed look at how wor...


Read full story
llama.cpp has shipped its first semantically versioned release after 3.5 years of sequential build numbers. The headline isn’t a flashy feature — it’s a stability promise that benefits every developer building local AI tools, including students running LLMs on a laptop for free.

The University Network

For three and a half years, llama.cpp tracked ...


Read full story
Google DeepMind launched Gemini 4 Hydrogen on Nov.12, 2026, completing its Gemini 4 lineup in just over six weeks. The model is built for speed and everyday use, and it’s rolling out to free-tier users automatically — with API pricing that undercuts GPT-4o and Claude Sonnet.

The University Network

Google DeepMind on Nov. 12, 2026 released Gemini 4...


Read full story
OpenAI launched a science blog on Nov. 15, 2026, designed to share AI-driven discoveries and practical research methods with scientists worldwide. For students, the methods pillar — covering literature review, data analysis and experiment planning — maps directly onto core research tasks.

The University Network

OpenAI on Nov. 15, 2026, launched a ...


Read full story
IBM has released two experimental Granite 4.1 models trained on LLMON, a new markup language that separates instructions from data at the token level. The result: near-zero success rates for indirect prompt injection attacks — a critical security risk in AI apps.

The University Network

IBM has released two experimental language models —


Read full story