🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Today we released an open source model, DBRX, that beats all previous open source models on the standard benchmarks. The model itself is a Mixture of Experts (MoE), that's roughly twice the brains (132B) but half the cost (36B) of Llama2-70B. Making it both smart and cheap. Since

Introducing Meta Llama 3: the most capable openly available LLM to date. Today we’re releasing 8B & 70B models that deliver on new capabilities such as improved reasoning and set a new state-of-the-art for models of their sizes. Today's release includes the first two Llama 3

Nice! Structured output is important. Cool to see this deployed at scale

We built a platform to run containers across Cloudflare’s network. It’s different. It’s schedules workloads globally — you don’t deal with regions. It fits with Anycast (of course). And we're integrating it with Workers and Durable Objects — use the right tool for the job, every

We've been shipping y-sweet.dev as an open-source project for over a year. Now you can deploy it on Jamsocket as a natively-supported service, without touching a Dockerfile!

Thanks to a couple contributors, Ghostty has built-in VCS branching characters. Kitty introduced this[1] and I think Ghostty is the second terminal to implement it. The cool detail in this one: for the end markers they're drawn with a true fade (see line 13). [1]:

The secret behind my productivity is that I'm almost exclusively a print-statement debugger

my offering to the llm overlords 491,946 tokens (all 936 of my posts) matt-rickard.com/llms-full.md

Wow there is a *lot* of llms.txt activity happening now. Like… a lot lot… directory.llmstxt.cloud

happy franksgiving

Should check this out! TS as a first class citizen

Interesting approaches on making structured generation more efficient

interesting analogy

Introducing Llama 3.3 – a new 70B model that delivers the performance of our 405B model but is easier & more cost-efficient to run. By leveraging the latest advancements in post-training techniques including online preference optimization, this model improves core performance at

Introducing Veo 2, our new, state-of-the-art video model (with better understanding of real-world physics & movement, up to 4K resolution). You can join the waitlist on VideoFX. Our new and improved Imagen 3 model also achieves SOTA results, and is coming today to 100+ countries

we are working on OpenControl - a library that uses AI to finally solve the headache of internal tooling idea is simple - define a set of tools in code that can reference functions from your codebase or anywhere else these tools get deployed to your infra (runs anywhere,

The distinction between "agent" and "tools" is arbitrary and meaningless. Architectures built on that distinction are dead ends. In the end it's all software: sometimes with LLM, sometimes without. I'm sure someone will soon put an agent inside a MCP and feel like a genius.