Own code vs agent frameworks?
3 years running agents in production taught us this:
We started with LangChain. Its loop didn't fit, so we built our own.
Even the LLM integration layer needed custom code.
A dependency can save code and still cost you control. π§΅
Improvements in caching and inference let us deliver GPT-6 Sol and GPT-6 Luna at lower cost.
Weβre passing those savings on through lower API prices. Sol and Luna deliver more intelligence at half of the previous generationβs API prices, making more complex work practical to automate at scale.
GPT-6 Sol and Luna just landed in Astraβs orbit.
Both launch today with API prices 50% lower than GPT-5.6.
Build with Sol. Scale with Luna. To production and beyond.
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family.
It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
@buildwtim Plug in an analytics tool like Umami and replay the user journey on your landing page.
Track where they come from, time spent, what they click, user agent, bots, drop-off points, etc.
@IamMysticc Weeklies are enough? Doing twice a week and constant chatting on slack but even though it's harder than being on the same office I feel like.
Anyone here building remotely with other cofounders ?
How do you stay aligned/focused when everyone works on their own stuff , without constantly chasing each other for updates ?
@itsalicesoul Na I don't let my agents post at all on X they suck at human interactions π
Other than that I use agents for anything else code , monitoring, debug, log review and so on.
@roudyhermez Ahah I'm the opposite. Over the years I built so many things without ever thinking they were worth productizing , until someone launched the same thing as a SaaS and made money π
@prasenx Agreed, but most vibecoders are just integrating HTML sites with external APIs, so the risk and the bar are pretty low.
We rarely see genuinely complex products being built by vibecoders.