John in 't Hout

@vcluster @hrittikhere @kunalD_official Argo + FinOps automation is powerful. Yasu.cloud layers autonomous prevention on top: AI catches wasteful patterns in IaC/PRs before they deploy into your K8s environments. We’ve seen 30%+ savings with zero dev friction.

Second session of the day at #KubeCon EU. @Hrittikhere and @Kunald_official as from CastAI on automating FinOps through Argo Workflows. Teams reduced waste by 45% with these patterns. 1:35 PM CET, G102-103. #CloudNativeCon

Transparency is table stakes. The real ROI comes from prevention. Yasu.cloud acts as your 24/7 AI Cloud Engineer: it doesn’t just show the waste, it stops costly misconfigs and sprawl in PRs/Slack before the bill hits. Perfect for controlling AI-driven cloud spend proactively.

#AI investments are everywhere—but do you know where the money is going? Cost transparency helps leaders see the full picture, optimize spend, cut waste, and turn tech investments into real business outcomes. Learn more here: ibm.co/6010EHyU4 #FinOps #CIO

Exactly, usage patterns are the real driver. Yasu.cloud turns this into prevention: our AI Cloud Engineer scans IaC and configs in GitHub PRs to stop linear cost growth before it happens (works across AWS/GCP/Azure). Our customers are seeing 30-40% savings with zero extra headcount.

Cloud costs aren’t rising because of scale. They’re rising because of how data is accessed and these costs continue to scale linearly with usage. Strategy Mosaic changes that. Learn how to take control ↓ ow.ly/PjJI50YvWNF #CloudCost #DataStrategy #FinOps #SemanticLayer #AIInfrastructure

@abhitwt Spin up more agents. at least 3 at the same time?

POV: you’re a developer in 2026😂

@thenowhereway No winner, different audiences.

Who will be the winner of the AI race? - Google? - Anthropic? - OpenAI?

@justbyte_ Damn.... unpickable...

Developers, choose one forever

@NaveenS16 100% agree, ‘cost in PR’ is the ultimate behavior changer. We at Yasu see some amazing clarity for our customers in there CI and devs now optimize proactively. As a bonus we now started showing the sustainability info as wel. Magical to see engineers becoming more aware.

IMHO #FinOps shouldn’t be a scary monthly bill. The best way to cut cloud waste is showing the estimated cost directly in the PR. If a dev sees that their one-line change adds a few dollars a month, they’ll optimize it before it even hits prod. #Kubernetes #PlatformEngineering

@osodevops OpenCost is fantastic for K8s visibility. Pairing it with automated PR cost gates (via Yasu) has prevented so many ‘oops’ commits our customers. Great guide btw!

Kubernetes FinOps with OpenCost: A Comprehensive Cost Optimization Guide @thamizhelango/kubernetes-finops-with-opencost-a-comprehensive-cost-optimization-guide-53bcf26dc266" target="_blank" rel="nofollow noopener">medium.com/@thamizhelango…

@Apptio FOCUS is a game-changer for multi-cloud + SaaS normalization. We’ve started extending it to AI inference costs too, huge visibility win. And combining it in our platform at Yasu it shines with shift left + Agent combo! 🫰 it!

How are #FinOps teams using FOCUS beyond #cloud cost normalization? Think SaaS, data centers, #AI spend—and a spec that travels with you from org to org. Check out the new episode of FinOps Fridays with Matt Cowsert from the FinOps Foundation: ibm.co/6014EH0pu

@subdupes Exactly, SaaS sprawl is a silent killer. We see it so often at Yasu

Every team solves problems by adding tools. Nobody solves it by removing them. #SaaS #FinOps #TechDebt

@InstaDevops Love the spot instances + VPA combo! We’ve seen another 15-20% by shifting left with our customers at Yasu in PRs so devs see the bill before merge. Anyone tried combining this with OpenCost dashboards?

💰 Kubernetes Cost Optimization Stop wasting 30%+ of your K8s budget: ✅ Right-sizing with VPA ✅ Demand-based HPA scaling ✅ Spot instances for 60-90% savings ✅ Namespace quotas Master FinOps for K8s: dev.to/instadevops/ku… #Kubernetes #FinOps #DevOps

@justbyte_ :D

Be honest, is this relatable??

6.8 billion people have never used AI. Not once. My feed has me convinced I’m already behind. Then I show a mate something basic and he thinks I’m a wizard. The bar to be in the top 1% globally? We are so early it’s almost embarrassing

Seedance 2.0 Prompt: Sum up the AI discourse in a meme - make sure it’s retarded and gets 50 likes.

When your token limit has been reached!

@justbyte_ I have been trying codex last few weeks, feels pretty good to be honest. Claude Code was prompting me alot of obvious questions. Codex since (5.2) just understands the code base better (sometime reads for 10 mins straight but does not matter (I am doing other stuff)

Be honest, is there any replacement for this.

@ionleu 👋

if you're in tech say hi

So my current setup for testing while agentic coding: Pre commit hook: - EsLint - Typecheck - Vitest unit test Than in my CI/CD github pipeline: - Above checks - Full e2e checks suite (playwright). Only deploying to Vercel of all checks complete. Any other suggestions?

Finaly got all the tests fully covered / working! I started focusing more and more and extensive testing (Test in this list is e2e tests btw.) to be more confident when vibe coding / deploying! When you know all critical paths are covered vibe code easier! What do you guys use?