Wanna speak about AI software factories? I will be in Hamburg next Tue at the AI-SUMMIT 2026 organized by @hamburg_ai
Let's talk about AI software factories, their impact (we have some awesome benchmarks), and how they could transform your engineering work.
You can find our booth in the AI startup area hosted by @Oracle .
@ivanburazin That makes total sense to me. Tbh, nowadays, with the fast development of AI, I feel that any roadmap longer than a few weeks is a waste of time.
I shared yesterday at our agents in prod meetup by @glassflowdev the impact of our software factory:
_3x more shipped changes
_34% shipped autonomously overnight, with higher-quality output (see the results in the photos).
The best part came after the talk. Sharp questions, honest debates about when a factory makes sense and when it doesn't, and a room full of people already running agents in prod.
New 2h crash course from Divesh Jadhwani: modern #AIagents & multi-agent workflows with #LangChain/#LangGraph, traced end to end with Rius.
"This mapping-level breakdown is what makes it really different."
https://t.co/heSpJx6lcI
#AgenticAI#AI
Frankfurt, thank you 🙌
Great builders and even better conversations at #GrokBot x Build Days AI Week. We built an #AISRE and a #softwarefactory live and left excited about what's next.
Build your own: https://t.co/PIFmnPILZg
@businessbarista I am sure there's a massive wave of shadow AI happening in these orgs right now.
IMO, it's only a matter of time until the business units move so fast that IT is forced to remove its approval process.
In my experience, founder-friends are the most loyal, supportive friends you can imagine.
_Need an intro? Done.
_Want me to beta test your product before launch? Done.
_Want me to drop your name on an event/conversation? Done.
Unconditional support because they actually know the pain.
When an agent crashes with an error, that’s easy.
The real threat is the agent that delivers a wildly wrong answer with absolute confidence 😅. No warning. No trace. Just a lost user who closes the tab.
That’s a silent failure. More logs won't fix it. You need outcome verification.
Hundreds of developers in one room building with GrokBot? Count us in! 🚀
Excited to announce that @glassflowdev is sponsoring the Build Days w/ GrokBot this Friday in Frankfurt.
Join us to build, innovate, and meet the GlassFlow team. Looking forward to seeing everyone there!
Register here:
https://t.co/8vs4KogNub
@cjc IMO, it depends on who the recipient is. Investor-to-investor intros, then yeah, they work best when mainly written by my investor who knows their peer's thesis.
Customer intros need the founder's pitch + a warm personal note from the connector.
@coryalthoff Yeah, that's only because we don't catch them. There are so many silent failures that don't throw loud errors and we usually only notice them after the harm is done, or not at all because the receiving agent simply accepts the bad output as fact.
Wanna build your own software factory? Super excited to run a workshop with the @claude_code community in Munich.
📅 October 29, 18:00 / JetBrains Munich office
Register here:
https://t.co/Ot9cBLJaUs
@Trace_Cohen Totally agree! Too many founders put way too much weight on fundraising, but if you can figure out a working product in a good category, you don't even need to be amazing at raising money. VCs will find you.
@FilipPanoski IMO, you can have a semi-good product and use distribution (first a bit smaller) to iterate fast on the product before you step on the gas for full distribution.
Agent observability was built for a world where humans fix everything.
Traces and evals point backward. They explain what already went wrong.
Self-improving agents move forward. They need context for the next step.
Two different products.
It was such a pleasure to host the Claude Code community Berlin. Awesome to see people building on Tares by GlassFlow. The participants have built their own AI SRE and a software factory running locally.
Wanna try Tares yourself? Here you go:
https://t.co/CaAR5dfwmY