The words you use to speak literally tell people where you grew up 🗺️🗣️
Just found Dialectly — an interactive map local slang and hometown sayings. It’s everything from the age-old “Soda vs. Pop” argument to regional infrastructure and weather slang.
https://t.co/Y0en9EMDBC
GPT-6 isn’t out yet.
Instead, OpenAI released GPT-5.6 (Sol / Terra / Luna) on July 9.
Here’s what it actually means for people building real linguistic applications in 2026.
As before, a raw large LLM is only part of the solution.
Hybrid systems (strong base model + lightweight specialized components for memory, style control, and domain logic) remain the most reliable way to get consistent results in production.
GPT-6 isn’t out yet.
Instead, OpenAI released GPT-5.6 (Sol / Terra / Luna) on July 9.
Here’s what it actually means for people building real linguistic applications in 2026.
What’s still problematic:
Cultural nuance, idioms, and tone — progress has been made, but it’s still not production-ready for demanding use cases.
True long-term user memory across sessions and weeks — still not at the level needed for high-quality tutoring or coaching applications.
Low-resource languages and subtle linguistic details — frontier models continue to struggle here.
NVIDIA keeps squeezing beefier mobile GPUs into slimmer laptops… as it wrestles with heat, noise and battery constraints.
But maybe they’re the ones who need to be fixed.
They already have most of the foundation: GeForce NOW and cloud rtx infrastructure.
Instead of running after hotter and hotter hardware, why not get the cloud right?
Fix latency. Erase queues. Speed up the mid-tier subscription so it’s fast and reliable.
That you could be carrying a silent ultrabook — or a tablet — and all the GPU power could be in the cloud.
No overheating.
No loud fans.
No compromise.
A subscription the cost of Netflix might actually be a viable alternative to purchasing a high-end gaming laptop.
The future of performance may not be in your device at all.
NVIDIA, what are you waiting for?
NVIDIA keeps squeezing beefier mobile GPUs into slimmer laptops… as it wrestles with heat, noise and battery constraints.
But maybe they’re the ones who need to be fixed.
They already have most of the foundation: GeForce NOW and cloud rtx infrastructure.
Instead of running after hotter and hotter hardware, why not get the cloud right?
Fix latency. Erase queues. Speed up the mid-tier subscription so it’s fast and reliable.
That you could be carrying a silent ultrabook — or a tablet — and all the GPU power could be in the cloud.
No overheating.
No loud fans.
No compromise.
A subscription the cost of Netflix might actually be a viable alternative to purchasing a high-end gaming laptop.
The future of performance may not be in your device at all.
NVIDIA, what are you waiting for?
NVIDIA is advancing graphics and simulation with agentic and physical AI at #SIGGRAPH2026.
From open models to real-time simulation, AI and graphics breakthroughs are transforming media, content creation, and robotics.
Read the latest announcements. 👇
https://t.co/TuuwW8QdkH
Even the best LLMs in 2026 still struggle with real-world language tasks.
Here’s why building production linguistic applications is much harder than it looks on benchmarks. 🧵