Language

Say hello to Gemini 3.5 Transcribe! - Build apps that understand user speech / intent, even w/ multiple speakers! - Auto-detection of 85+ l...

Ox Alpha revealed: @Zai_org’s GLM-5.3-Flash, the first native multimodal model in the GLM-5 series. Ox Alpha was the biggest model ever on ...

This finally stopped Claude's crazy pursuit of talking more and more unintelligible every passing day > Save in memory to always use ASD-ST...

Ox Alpha is actually GLM-5.3-Flash, and it’s one of my favorite models ever released. https://x.com/theo/status/2093078228491731177/video/1 ...

We need to get away from software that we can’t change with a prompt. https://twitter.com/dhh/status/2091841106480627871

New Anthropic research: A global workspace in language models. Of everything happening in your brain right now, only a tiny fraction is con...

Few people care more about software performance than @cmuratori. Give him a few minutes of your attention with this episode, and he'll convi...

In contrast to some other frontier models, GPT 5.6 has quickly consumed share within the OpenAI model family. 5.6 Sol now sees 50%+ of spen...

I was thinking about how I recognize AI writing, and one big tell is excessively colorful verbs. A handful of journalists might write that a...

this is good i always thought a minimal harness should be all native they'll probably add an embedded scripting language for extensibility ...

I don’t like ELI5 as default for “make this easy to understand,” a great thing about LLMs is the ability to work as universal translators ac...

SL2T is our amazing sign-language-to-text model allows users to sign directly to their phones for the first time. Built in close collaborati...

I have tried this a few times But the overhead of understanding the diagram (and fixing slop) is huge for complex systems Consistent domai...

New inference provider: @makora_ai is now live on OpenRouter. Starting with: • DeepSeek V4 Flash • GLM 5.2 FP8 • GLM 5.2 NVFP4 • Kimi K3 R...

Deft, a new model focused on “better writing,” is a promising demonstration against “not X, but Y,” the patterns of three, and the words “re...

One way you can recognize AI slop is when the ideas and the diction don't match — when the ideas are completely ordinary, but the diction is...

Eventually, all UI will be chat and gestures in, chat and video out.

SL2T is our breakthrough sign language-to-text model powering new features for Deaf and hard of hearing users on @Android. Starting with Am...

.@lfschiavo reveals the recurring religiosity patterns she's observed in AI models, from Opus 4's Zen koan cycles to Fable's fascination wit...

While we’re at it can we get trends.claude & trends.chatgpt Would be fascinating https://twitter.com/jeff_weinstein/status/209191366796526...