https://t.co/8BMuxsp7gf

How does ChatGPT memory work under the hood? I've been interested lately in AI memory. So this post(and the one on Claude memory), by @shloked_ , are super useful.

Looks like it's not?

I'm excited to announce ReqLLM - a new Elixir package that brings Req's middleware power to LLM calls. As I was building @agentjido, I realized I needed a lower level API when calling an LLM Req was the perfect choice

Attending this later today.

I'm hosting "Claude Code Anonymous" in Berlin on Sept 10 with @steipete as special guest A meetup for engineers serious about agentic coding, whether you’re using Claude, Cursor, Codex, Amp or something else. No hype. Just lightning talks, shared learnings, and open discussion

Purchased! Excited to follow along.

The Realtime API is officially out of beta and ready for your production voice agents! We’re also introducing gpt-realtime—our most advanced speech-to-speech model yet—plus new voices and API capabilities: 🔌 Remote MCPs 🖼️ Image input 📞 SIP phone calling ♻️ Reusable prompts

The Expert Language Server is live: https://t.co/Tbqnh50NB3 - nightly builds available for alpha testing and contributors. Thank you to @flydotio, @_tauspace_, and @River for sponsoring the development so far!

Introducing Marvis-TTS 🔥🚀 A new local-first TTS model @lllucas and I built for efficiency, accessibility, and real-time performance right on consumer devices like Apple Silicon, iPhones, iPads, and more. Traditional TTS models often demand full text inputs or sacrifice

https://t.co/i0yNMvGwsg

Attending this later today. Anyone else I know going? Sounds like an exciting set of talks from @steipete and @HHaandr

Something fun we discovered: you can use #Genie3 to step into and explore your favorite paintings. Here's a short visit to Edward Hopper's "Nighthawks".

Fun fact: the show was animated by a team in Nepal(Incessant Rain Studios). I was surprised to see Nepali names in the credits when watching it a few months back.

Full house today at the @PrimeIntellect Berlin meetup

Releasing Shoggoth Mini!🐙 Soft tentacle robot meets GPT-4o & RL. I built it to explore the boundaries of weird: expressiveness, aliveness, and AI embodiment. Blogpost: https://t.co/bNH6b0VEzk

2/ @maciejgryka walked us through an Elixir version of @thorstenball 's coding agent tutorial

Great presentations and conversation at the Elixir Berlin Meetup today 1/ @Overbryd talked about his big bag of Postgres tricks and techniques #myelixirstatus #elixirlang

Now you can add semantic search to your @elixirlang Ecto query in a few lines of code! Check out: https://t.co/aoM4mKDbD1 #myelixirstatus #elixirlang #elixircommunity #elixirprogramming

I've been building voice agents for the last 6mo and I think the chat-supervisor pattern is a game changer. Stitched model (STT-LLM-TTS) is slow, but realtime audio models aren't (yet) as smart as text. This has the best of both worlds. Here's how it works: