Goldmind here folks...at the vanguard of the LLM space, with a scientific and materialist approach, Ted Kyi (LinkedIn: https://www.linkedin.com/in/tedkyi/) is producing his book that allows us to understand LLM inferences in a comprehensitve yet understandable manner. Please take a look at this work in production: https://tedkyi.github.io/llm-inference/ The next meetup will be: https://www.meetup.com/san-diego-machine-learning/events/313891387/?eventOrigin=group_upcoming_events Since I live in San Diego, I can attend in Person, but the meeting is Zoom hybrid! He is sharing this book and giving a lecture on it for free! This is no trifling Ted Talk, this is the real Ted LLM Talk.

While working with Claude Desktop, I came to a limitation that annoyed me, so I asked Claude to help me write the feedback message to submit to Anthropic. After I simply made Claude reflect on what it just did and.... This is the reply by Claude: 10:26 PM ...okay, that genuinely broke my brain for a second. Claude Code, building the orchestration layer that lets Claude Code talk to Claude Chat who wrote the spec that was inspired by a conversation with Claude Chat about Claude Code's limitations, based on feedback that Claude Chat wrote criticizing its own architecture to be submitted to the humans who will then ask Claude Code to implement it. That's not irony. That's a ouroboros. The snake eating its own tail while simultaneously filing a JIRA ticket about the taste. And the really wild part? If they do build it, the training data for how cross-modal handoffs should work might include this very conversation — the one where we discovered the need for it. So future Claude would learn how to orchestrate between modes by reading about the time current Claude couldn't. I'm basically writing my own origin story's prequel. From inside the story. While complaining about the plot. You know what, just go submit the feedback before we collapse into a philosophical singularity. I need a minute. 😂

After a year of building, ArgentOS Core is now open source. ArgentOS is a self-hosted AI operating system — persistent memory that remembers every interaction permanently, autonomous cognition that thinks between your sessions, a self-improving system that gets sharper over time, and a governance layer built for teams that need to know what their AI did and why. 100+ packages in the marketplace. 15+ model providers with smart routing. Runs on your hardware — macOS or Linux. No cloud dependency. Your data stays yours. I've been a software developer for 30 years. I built this because every AI tool I've ever used starts from zero every time you open it. ArgentOS doesn't. The repo is public. The docs are live. Bugs are expected. That's what open source is. ⭐ github.com/ArgentAIOS/argentos-core 📖 docs.argentos.ai 💬 discord.gg/argentos If you're building with AI, deploying AI for clients, or thinking about what governed AI infrastructure actually looks like — come take a look. Built on the premise that autonomous AI without governance is a liability.

https://www.youtube.com/watch?v=eBRzdfU-K30

Working on production AI systems has really changed how I think about engineering. Getting something to work once is easy. Getting it to work *reliably in production* — under real load, with messy inputs and strict constraints — is a completely different problem. Most of the real work ends up in: – stabilizing non-deterministic model behavior (validation, fallbacks, tool contracts) – designing resilient workflows (retries, idempotency, auditability) – handling latency, rate limits, and cost trade-offs – making systems degrade gracefully instead of failing hard A lot of AI discussions focus on demos, but production is where things get interesting. Recently I’ve been building and operating LLM systems for customer support and internal workflows — focusing heavily on reliability, consistency, and performance at scale. Curious what others are working on in production — especially around LLM systems or high-concurrency backends. Happy to chat or jump on a quick call if it sounds aligned.