I'm starting a new meetup called CORE—the Central Ohio Radio Enthusiasts. It's for radio and radio-adjacent interests. It's not a ham radio club, we talk about that, but also GMRS, mesh networks like Meshtastic, Reticulum, and HaLow, software-defined radio, and shortwave listening.

So, I learned a new Internet abbreviation today: ai;dr

A replacement for HamClock, which will shutdown forever come June, has been release. It's called OpenHamClock. Looks pretty swish! openhamclock.com

I registered w8guy.radio yesterday. Nothing there yet but hopefully you'll be able to email [email protected] in the future! Or maybe I should go with [email protected]?

Found this nifty tool to create animated Mermaid diagrams. I've just messed with the free version but it's pretty cheap and even has a pay as you go if you just need a couple. It's at animerm.ai if you think it'd be useful.

Hey @pa-dev.bsky.social, I too am a Full Snack Developer!

Meet "Podbot," an AI chatbot @guy.dev built using Azure OpenAI to recommend podcasts based on your preferences. It uses Redis Memory Server to manage memory, so it remembers what you like across conversations. Build your own Podbot here (and feel free to open a PR): github.com/redis-develo...

6 técnicas japonesas que cumple todo desarrollador de software que tiene éxito. 🧵 👇

"With elevated interiors, bold designs, and specialty drinks paired with your Taco Bell favorites, DRG’s venues are known as some of the most iconic destinations across the country." So you can get drunk and eat Taco Bell at the same time, I guess. Iconic. tacobellcantinas.com

Semantic routing improves predictability and efficiency in AI systems. RedisVL classifies user queries without LLM calls, reducing token usage and latency. Combined with LangGraph, it creates a deterministic flow for routing, execution, and decision logic. Read how here: medium.com/@bhavana0405...

I run a Discord server for my radio club. We have the following rules: 1️⃣ Don't be a jerk. 2️⃣ Don't do anything that makes us add a new rule. 3️⃣ If your not sure if rule 2 will be invoked, see rule 1. This work surprisingly well for our admittedly small scale.

"The legend lives on from the Chippewa on down of the big lake they called Gitche Gumee. The lake, it is said, never gives up her dead when the skies of November turn gloomy."

I like decentralized networks. In particular, I'm a fan of Reticulum. Reticulum is a fully-encrypted, anonymous, and decentralized mesh network that runs over all sorts of hardware—from Ethernet to LoRa radios to serial cables.

The Great State of Ohio has seen fit to pass a law requiring me to perform age verification on this platform. For the children or something. If I don't do this, I won't have access to DMs. I know how to VPN, not a huge problem. But, it does means I won't check my DMs as much. So, email me instead.

I blogged. medium.com/@guyroyse/ma...

Modern problems require…modern memes? At #Devnexus, @guy.dev showed how embeddings + vector databases unlock quirky use cases (like discovering your meme twin) while solving real-world challenges 🦾 Recap here ⤵️ devnexus.org/posts/modern... 🙌 Join us for more inspiring sessions at Devnexus 2026!

So, @dcdietrich.bsky.social will be delighted to know that I got a pair of DMR radios. Once I figure them all out, we’ll have to chat!

Every time I see the letters AI, I have to figure out if it's some dude named Albert, a steak sauce, or hype.

Always a great time chatting with @davidgiard.bsky.social—even on a Monday morning at 9am! I wasn't able to cover nearly as much about ham radio as I would have liked, but if anyone has questions, I'm happy to answer them.

Monday on #TechnologyAndFriends: Guy Royse (@guy.dev) on Ham Radios

Everybody's putting AI in their apps. They're stringing APIs together and sending the results down to the browser. But those APIs are expensive, the round trips are long, and there's a lot of data being exposed. Maybe there's a better way? www.youtube.com/watch?v=nBaG...

Vibe coding browser-based AI stuff. And asking the important questions.

Vector search is cool. Really cool. And, we're all itching a reason to use it. And there are some really good reasons. But there are some, well, not so good ones as well. Know the difference and don't let vector search become your latest Golden Hammer. www.youtube.com/watch?v=Ryme...

Explore the fascinating world of probabilistic data structures with @guy.dev ‘s session from Devoxx UK 2025. Learn how to use Bloom Filters, MinHash, and Top-K to analyse 112,092 UFO sightings. View session at buff.ly/5dcCMaS

These images from the nitric acid spill in Vinton County, Ohio are both disturbing and oddly beautiful. www.10tv.com/article/news...

I watched a YouTuber react to The Rocky Horror Picture Show last night and now all the songs are stuck in my head.

RAG !== vectors. The R is for retrieval—i.e. get useful stuff. You can do that lots of ways: 📄 Full-text search 🧾 SQL query ⚡ @redis.io key lookup 🧠 Yes, even vector search It's not how you retrieve—it's what you retrieve. Get the right data. Feed it to your #LLM. That’s the magic of #RAG.

Whenever things get broken and absurd, the circus music starts playing in my head. I sometimes say “Circus Music Intensifies”. From now on, I think I’m gone use the emoji combo of 🤡🎶📈.

Left my hotel in Antwerp at 8:30 am. 20 hours and 6 time zones later, I shall arrive home at about 10:30 pm. It’s gonna be a long day.

I'm trying an experiment on my Renaissance Guy blog on Medium. Here's the first post: medium.com/renaissance-...

Wanna search images? It's all about embeddings! Join my live session—Semantic Image Search with Hugging Face & Redis. → Image embeddings with @huggingface → Fast search with @Redisinc → Step-by-step code walkthrough Register now. redis.io/events/seman...

Y'all know what's gonna happen. But you wanna here the song anyhow!

Missed a session you really wanted to see? Subscribe to our YouTube channel to be notified when the 2025 recordings are published - buff.ly/2qOSCNF

I think we should call the new pontiff Bob Pope.

🎤 Speaking TONIGHT at @dotnetsheff.co.uk (May 5)! "Finding Your Meme Twin with Embeddings & Vector Databases" at Pendo. I'll show how to transform memes into embeddings and then find your meme twin! Doors open at 6:15pm. www.meetup.com/dotnetsheff/... #Embeddings #VectorDatabases

Good news, everyone.

I'm looking to subscribe to cool stuff. Where and by whom do you get your developer news?

I was able to talk @redis.io into putting together a happy hour at Forbidden Root in Easton the night before @stirtrek.com. Just a chance to network, share some food and drink, and talk with nerds.

I keep finding myself having to embed images for testing. So, I wrote this little tool that I imaginatively called the Embeddinator. guy.dev/embeddinator/ I only have two models in the drop down, but it's just using @hf.co's Transformers.js library. Easy enough to add more.

Redis 8 adds a new data type called vector sets. I wrote some code to play with them and shared some my thoughts on when I think they're useful. Check out the code and thoughts at guy.dev/2025/04/15/a.... Or—if you prefer Medium—at medium.com/the-guy-wire....

More ChatGPT images. "Using this image, please render me as a comic book superhero. I am a software developer, public speaker, D&D nerd, and ham radio operator. Come up with a clever superhero name for me and incorporate that into the image. Make it look like a panel from a classic comic book."

Well, all the cool kids are doing it.

Science confirms that I am as white as you can get. But does that Danish bit mean I'm part Viking?

Redis 8 RC1 is out! New features include numerous performance improvements, more commands around hash field expiration, and a new data type written by @antirez.bsky.social—the OG himself—called vector sets. Redis 8 RC 1: hub.docker.com/_/redis. Vector sets: github.com/redis/redis/... @redis.io

I’m leaning about semantic search and vector embeddings with @guy.dev at Orlando Code Camp!

Join @guy.dev at Devoxx UK 2025 to discover your meme twin! Learn how to turn memes into embeddings and search them with a vector database to find your meme lookalike through a live demo. View talk @ buff.ly/MXah4wC

Advancing Semantic Caching for LLMs with Domain-Specific Embeddings and Synthetic Data Redis demonstrates that fine-tuning smaller embedding models on domain-specific data outperforms larger models for semantic caching. 📝 arxiv.org/abs/2504.02268

A key on my mechanical keyboard has been acting up, so I tore it apart to clean it. It was full of whiskers and dandruff. Sorry. But I got it all cleaned up and reassembled. Feeling good about myself. But then...

I wrote a blog post about using async generators in #TypeScript to stream #Bigfoot sightings out of @redis.io. It’s fast, memory-efficient, and weirdly fun. 👣 Read it 👉 guy.dev/2025/04/01/f... Code 👉 github.com/guyroyse/asy...