in 
p/github-copilot
p/github-copilotGitHub Copilot is now usage-based. What's your team doing about it?
Debbie Shapiro:@qifengzheng @fmerian I start tracking my agent usage in a local database last month (when Anthropic sent out their email about shifting automated usage to a different monthly budget). Now I'm tweaking them by leaving very consciously chosen runs on Claude (various models) and routing other work to OpenRouter, Ollama Cloud and local LLMs. I've already reduced their cost substantially.
in p/github-copilot
p/github-copilotGitHub Copilot is now usage-based. What's your team doing about it?
Debbie Shapiro:I'm not using Copilot, but invoices are coming to Claude starting next week as well. I built a tool (I launched here a week or so ago) to support cross-session memory querying. Now it's even more of an imperative to use tools like this. The past month, I've been on a mission to fine-tune my agents to keep them under my monthly budget. It's challenging for sure. I've moved some agents to cheaper or local models, moved as many deterministic things as possible to scripts, etc.
in p/github-copilot
p/github-copilotGitHub Copilot is now usage-based. What's your team doing about it?
xixuan wang:@fmerian Yeah, the vendor lock-in feeling is pretty real in this space. Even when switching models or tools, the GitHub ecosystem (repos, PRs, workflows) still pulls you back in. Curious what specific parts feel most locked-in for you?
in 
p/vcboom
p/vcboomI was a VC for 8 years. The same broken pattern kept showing up.
Yoann Berno:@zhen_han @sabber_ahamed since AI had commoditised product development, what’s left to derisk is marketing and commercial traction. So it’s fair that investors want more proof on this front
in 
p/intrascope-app
p/intrascope-appAI access is easy. AI governance is the next problem.
Ali Umair:Congrats on the launch, AI access is easy. AI governance is the next problem.! — looking forward to trying this.
in 
p/general
p/generalWhat are the 5 tools you simply couldn't do your work without?
sanreds:mine would be, a plain text file with no sync or formatting for zero-friction capture, terminal for anything that actually matters, Claude when i need a draft that still sounds like me, a browser tab i never close, and probably Gmail for the same reason you listed. the tools that don't make these lists are usually the invisible ones you only notice when they're gone.
in 
p/claude
p/claudeClaude Fable 5 just dropped - Anthropic's first Mythos-class model for everyone
sanreds:I am yet to experiment on this, Its free until June 22, then needs usage credits until capacity catches up, that phrasing is doing real work. Anthropic is constrained on this tier right now. The free period isn't generosity, it's a controlled ramp while they build out infrastructure. A new model tier that hits capacity at launch means the per-token compute cost is materially higher than Opus. the $10/$50 pricing reflects that.
in 
p/vibecoding
p/vibecodingDo Vibe coders actually test things?
Andrew Froehlich:Testing has become a serious part of my role as solo founder and developer. We still have a ways to go before I can delegate to a Claude Code testing agent for user ui testing. If anyone has any harnesses that could help with this, it is the final frontier for Claude Code development.
in 
p/producthunt
p/producthunt🚨 Landing Page Roast: 48 Hours Only 🚨
Christopher Alphonse:Would love a roast on Reinvoice: https://reinvoice.co It helps freelancers, contractors, and side-gig workers create invoices, track payments, send reminders, and keep tax-ready records without complicated bookkeeping software. I would love feedback on clarity, CTA strength, and whether the page makes you trust the product enough to try it.
in p/claude
p/claudeClaude Fable 5 just dropped - Anthropic's first Mythos-class model for everyone
Geert Zaal:working on my first few tests with Fable on both MetaScope and Armature. Will know more in the morning, starting a long running loop now with quite a large refactor on Armature. Implementing a medium sized RFC on MetaScope. So far so good, but no clear picture yet. To soon to tell