Its also a shame because Google has some of the most innovative AI apps, like NotebookLM, but they need smarter brains to power them (as well as the harnesses needed for those brains)
AI
-
LFM2.5-ColBERT-350M reliably selects top 5 tools from 151
By
–
LFM2.5-ColBERT-350M is a surprisingly reliable smart tool selector.
— Maxime Labonne (@maximelabonne) 18 juin 2026
We gave it 151 tools, and it consistently surfaces the 5 most relevant ones based on the user prompt.
This saves tokens and improves accuracy. Ideal for hmmmm agentic edge models? 👀 pic.twitter.com/IPyizctesULFM2.5-ColBERT-350M is a surprisingly reliable smart tool selector. We gave it 151 tools, and it consistently surfaces the 5 most relevant ones based on the user prompt. This saves tokens and improves accuracy. Ideal for hmmmm agentic edge models?
-
Flash models suffice for consumers but not for agentic work
By
–
A very good flash model may be all you need for mass serving consumers, but it doesn't do much for serious or agentic work.
-
Google lacks public frontier model, Gemini 3.1 Pro lagging
By
–
Interestingly, Google no longer has a public frontier model. They have a very good flash model, but a very good flash model can't do frontier work without a good frontier orchestrator. I am sure this will change soon, but Gemini 3.1 Pro is very clearly lagging at this point.
-
Princeton’s Goedel-Architect: AI generates formal theorem proving blueprints for Lean 4
By
–
What if an AI could write its own blueprint to prove math theorems? Princeton researchers introduce Goedel-Architect, a new agentic framework for formal theorem proving in Lean 4. Instead of recursively decomposing lemmas (which can loop on dead ends), it first generates a
-
Causal decoder patched to bidirectional encoder excels in multilingual tasks
By
–
We patched LFM2.5-350M (pre-trained on 28T tokens) to transform a causal decoder into a bidirectional encoder. It worked extremely well: both Embedding and ColBERT models get best-in-class performance, especially for multi/cross-lingual tasks.
-
GLM-5.2: best open-weight model with multi-head latent attention
By
–
Just updated with the recent release of GLM-5.2. The best open-weight model today. Architecture-wise, it is built on the GLM-5 and GLM-5.1 architecture that I covered previously, meaning it reuses the Multi-head Latent Attention mechanisms.
-
Technology integrated into human collaboration via Viktor in Teams
By
–
I like when technology is placed inside the flow of human collaboration, where people already work together every day.
— Antonio Grasso (@antgrasso) 18 juin 2026
Viktor brings this idea into Microsoft Teams as an AI employee designed to support the team in its everyday work.@viktor__com Partner. https://t.co/BpYcuvjRqRI like when technology is integrated into the flow of human collaboration, where people already work together every day. Viktor embodies this idea within Microsoft Teams as an AI employee designed to support the team in its daily work. @viktor__com
-
V can join memory channels and run code
By
–
V can join channels with memory of past tasks and goals, write and execute its own code, and turn conversations into completed tasks. Documented