𝗮𝘁𝘁𝗲𝗻𝘁𝗶𝗼𝗻 𝗶𝘀 𝗮𝗹𝗹 𝘆𝗼𝘂 𝗻𝗲𝗲𝗱

Main

Two new stealth models. 2 Million Context window. Possibly new Grok

Introducing Sonoma Alpha, two new stealth models 🥷

Context: 2 million tokens
Price: Free https://t.co/WfxeMdNMs3

Two new stealth models. 2 Million Context window. Possibly new Grok https://x.com/OpenRouterAI/status/1964128504670540264?t=M3fcGYgCEEoVTdr9hDxeQA&s=19

Super interesting research on hallucinations and an interesting thread by Adam with conclusions

New research explains why LLMs hallucinate, through a connection between supervised and self-supervised learning. We also describe a key obstacle that can be removed to reduce them. 🧵https://t.co/6Lb6xlg0SZ

Super interesting research on hallucinations and an interesting thread by Adam with conclusions https://openai.com/index/why-language-models-hallucinate/ 
https://x.com/adamfungi/status/1964040819196752312?t=ekqICNqeOHHURJCNNSrqWw&s=19

Introducing DeepSeek-V3.1: our first step toward the agent era! 🚀

🧠 Hybrid inference: Think &amp; Non-Think — one model, two modes
⚡️ Faster thinking: DeepSeek-V3.1-Think reaches answers in less time vs. DeepSeek-R1-0528
🛠️ Stronger agent skills: Post-training boosts tool use and

Today we're releasing Claude Opus 4.1, an upgrade to Claude Opus 4 on agentic tasks, real-world coding, and reasoning. https://t.co/25vh0b3FsX

Opus 4.1 https://x.com/AnthropicAI/status/1952768432027431127

GPT OSS from OpenAI https://openai.com/index/introducing-gpt-oss/

Introducing Genie 3, the most advanced world simulator ever created, enabled by numerous research breakthroughs. 🤯

Featuring high fidelity visuals, 20-24 fps, prompting on the go, world memory, and more. https://t.co/aTVguwTkSJ

wooah
https://x.com/OfficialLoganK/status/1952732206176112915

TL;DW Context Rules Everything Around Me
https://youtu.be/nyvmYnz6EAg

Phind launched an prompt based image editor 👀

Phind launched an prompt based image editor 👀
https://phind.design/

&gt;&gt;&gt; Qwen3-Coder is here! ✅

We’re releasing Qwen3-Coder-480B-A35B-Instruct, our most powerful open agentic code model to date. This 480B-parameter Mixture-of-Experts model (35B active) natively supports 256K context and scales to 1M context with extrapolation. It achieves https://t.co/Z8HfyrVScE

https://x.com/Alibaba_Qwen/status/1947766835023335516?t=q68BFI9K6C5gcVhVfij4Dg&s=19

Oh wow https://x.com/theo/status/1943910421859381320

🚀 Hello, Kimi K2!  Open-Source Agentic Model!
🔹 1T total / 32B active MoE model
🔹 SOTA on SWE Bench Verified, Tau2 &amp; AceBench among open models
🔹Strong in coding and agentic tasks
🐤 Multimodal &amp; thought-mode not supported for now

With Kimi K2, advanced agentic intelligence https://t.co/PlRQNrg9JL

https://x.com/Kimi_Moonshot/status/1943687594560332025

We just added Kimi K2 to T3 Chat!

It's the first open weight model that's actually good at tools/function calls before. The potential is nuts. It meets or beats Anthropic's best in tool calling! https://t.co/SxaYj2uHMx

https://x.com/Yuchenj_UW/status/1943005122793214267

Interesting to see that EE is looking to get massive wins in AI

PCB layout will be one of the first EE roles to be taken by AI.

Someone is going to combine a physics simulator, a DRC implementation, and a relatively small AI model (possibly diffusion?)

Low hanging fruit because the output is easily verifiable with enough compute

Generative AI meets RF circuit design = game changer

• Passive networks tailored by diffusion models.
• Specify stop-band/pass-band; AI does the rest.
• Pixel patterns are not intuitive to electrical response.

Designs getting more abstract.
Prepare for a cognitive shift. https://t.co/80ks0CBHB9

Interesting to see that EE is looking to get massive wins in AI https://x.com/EERandomness/status/1943394728336040027  https://x.com/vikramskr/status/1937658826414784589

From the Framework team but also applicable in general to Windows and Linux.

From the Framework team but also applicable in general to Windows and Linux. 

https://frame.work/au/en/blog/using-a-framework-desktop-for-local-ai