Iain Freestone

Whenever I show dev tools tricks in a video, I get lots of "I had no idea that was a thing!" What are your favourite or most under-utilized dev tools tips? There are so many nuggets in there, I'm sure I'm even missing some

@langfuse Prompt management for your @vercel AI SDK App 🚢 With Langfuse prompt management, your prompts benefit from - versioning (clear history with what changed) - labeling (serve different prompts in dev vs prod) - ability for rapid iterations - 100% availability in prod with zero latency - rollbacks and much more. 🍬Bonus: if you are on `ai` >=3.4.29 you will get the list of provided tools to the LLM and the toolChoice setting traced in Langfuse as well ✅ As always, thanks @lgrammel and @nicoalbanese10 for making it easy to integrate with the AI SDK 🥹🙏🏾

Finally, a NO CODE Editor for WebGL & Shader Effects!? @hiunicornstudio

⚡️Did you know that you can use Gemini 1.5 Flash 8B + function calling as a model router, to pick the best open-source or proprietary model based on your data + use case? This example uses function calling to select the most appropriate @HuggingFace model - then downloads the model, uses the open-source model with the image, and returns the result, all with a single Gemini inference call. Since Gemini 1.5's context window is ~1M tokens, you could register *thousands* of different OSS or proprietary models as potential tools. "Is this question challenging, or does it need powerful performance on a given eval?" -> automatically route it to the best model option "Is this question easier, or could it be solved with a specialized OSS model?" -> automatically have that function selected, and that model called Mind = blown. ‼️

Yeah, I'm migrating to this youtube.com/watch?v=AuHqwQ…

Hey Laravel devs 👋 I’m just reaching out to find out if any of you use AI in your Laravel projects? I’ve built a few AI projects using Laravel for some of my clients with has been a struggle due to the lack of AI packages. As we know python is the most supported language in the AI space so most of the AI tools out there have python support out of the box. I’ve used tools such as LangChain and Pinecone as was wondering if anyone would use these tools if I created a Laravel client for them? Let me know what you think… #Laravel #php #AI

Introducing sidebar.tsx—25 components to help you build all kinds of sidebars. I don't like building sidebars. So I built 30+ of them. All types. Then simplified the core into sidebar.tsx—a strong foundation to build on top of. It works with Next.js, Remix, Vite & Laravel.

Pan v0.1 is out today: an open-source, simple, lightweight, and privacy-focused product analytics PHP package 🎉

Anyone love their home WiFi mesh setup? Need to cover about 5500sqft and they don't sell the satellites for my existing Orbi setup anymore so looking for something new 🤔 I don't like tinkering with this stuff so hoping for something plug and play 🙏🏻

📹 Self-Hosting Next.js Learn how to deploy Next.js, Postgres, and Nginx to a $4 VPS with Docker. I'll explain how to use and configure Next.js features like image optimization, caching & ISR, streaming, middleware, server components, and more. Demo app is open source!

GM. I just deployed Next.js, Postgres, and Nginx to a $4 VPS. Time for a tutorial?

you DO need LoRAs! I had an idea for this vibe that I dont have time to make so just tweeting it out for others to take: Other Lives: a full 1990s picturebook of -*YOU*- going thru an alternate universe history, taking the other fork in the road, playing thru the consequences of one major event in your life ending differently. Snapshots from across the multiverse. could be therapeutic, could be horrifying, could be lifechanging. maybe @nutlope or @RobertHaisfield thing?

@DynamicWebPaige I did not realize the Gemini API had code execution support! ai.google.dev/gemini-api/doc…

🤯⚡️ Gemini 1.5 Flash-8B + code execution is insane. I just uploaded the text for a short story - Harlan Ellison's "Repent, Harlequin!" - and it was able to generate Python not only to tell me how many times Harlequin occurred in the text (✅ correct) but also to generate a dictionary for how many times each word occurred.

aimg - A Node.js CLI for generating AI images with Replicate and saving them to disk. zeke/aimg: 🧑‍🎨 A Node.js CLI for generating AI images with Replicate and saving them to disk. (github.com)

Fine-tuning Flux is pretty easy, but sometimes you need to do a little more work to get the best results. We just published a blog post with 3 techniques for using synthetic training data to improve your model outputs. 👇 Link in thread...

Let me show you... Inertia Modal! 🎉 🚀 Open any route in a modal or slideover 🔄 Supports nested/stacked modals 🎨 Customize position, size, and styling 🔌 Built-in event system for communication between nested modals 🆓 Free package 🔜 Coming soon to Inertia UI! Video ⬇️

It's easy to create your own Flux fine-tunes on the web, but you can also make them programmatically using our API. Here's a new guide to get you going: replicate.com/blog/fine-tune…

@fofrAI @mattrothenberg This is a great additional but it has just confused me a while as it is showing the the triggerword of a different model. I do have another trained model that has the triggerword FRDARTS. Can send more information if it might help.

We now show a flux lora’s trigger word on models to help you with your prompts. Also dark mode, coming soon. @mattrothenberg 🔥

Fine-tuning Flux with your own styles and characters is fun, but what's even more fun is combining multiple LoRA fine-tunes in a single image. Here's an example that combines two Flux fine-tunes to render Mona Lisa in a softserve anime style.