Talk Nerdy to Me!
@talk-nerdyto-me.bsky.social
🚀 Sharing approachable insights on Cloud Tech, AI and FinOps ☁️💰
Subscribe to Talk Nerdy to Me! newsletter here https://talk-nerdy-to-me.com/
Subscribe to Talk Nerdy to Me! newsletter here https://talk-nerdy-to-me.com/
The AI agent wars just got a ceasefire.
OpenAI, Anthropic, and Block donated their agent protocols to the Linux Foundation:
- MCP: 10,000+ servers connecting AI to tools
- AGENTS .md: 60,000+ projects using it
- goose: Local-first agent framework
OpenAI, Anthropic, and Block donated their agent protocols to the Linux Foundation:
- MCP: 10,000+ servers connecting AI to tools
- AGENTS .md: 60,000+ projects using it
- goose: Local-first agent framework
December 12, 2025 at 2:00 PM
The AI agent wars just got a ceasefire.
OpenAI, Anthropic, and Block donated their agent protocols to the Linux Foundation:
- MCP: 10,000+ servers connecting AI to tools
- AGENTS .md: 60,000+ projects using it
- goose: Local-first agent framework
OpenAI, Anthropic, and Block donated their agent protocols to the Linux Foundation:
- MCP: 10,000+ servers connecting AI to tools
- AGENTS .md: 60,000+ projects using it
- goose: Local-first agent framework
Europe just entered the hyperscaler chat.
Mistral Compute is live: tens of thousands of NVIDIA GPUs, European data sovereignty, and a direct challenge to AWS/Azure/GCP.
For teams dealing with GDPR, data residency, or just want infrastructure that isn't US-controlled—this is significant.
Mistral Compute is live: tens of thousands of NVIDIA GPUs, European data sovereignty, and a direct challenge to AWS/Azure/GCP.
For teams dealing with GDPR, data residency, or just want infrastructure that isn't US-controlled—this is significant.
December 11, 2025 at 1:02 PM
Europe just entered the hyperscaler chat.
Mistral Compute is live: tens of thousands of NVIDIA GPUs, European data sovereignty, and a direct challenge to AWS/Azure/GCP.
For teams dealing with GDPR, data residency, or just want infrastructure that isn't US-controlled—this is significant.
Mistral Compute is live: tens of thousands of NVIDIA GPUs, European data sovereignty, and a direct challenge to AWS/Azure/GCP.
For teams dealing with GDPR, data residency, or just want infrastructure that isn't US-controlled—this is significant.
The DevOps paradox of 2025 ☁️
✅ 78% of orgs have implemented DevOps
❌ Developers lose 6-15 hours/week to tool sprawl
📊 Average team uses 7.4 different tools
The solution? Platform engineering.
💡 Gartner predicts: 30% faster delivery for orgs that adopt it.
👉 Learn more at talk-nerdy-to-me.com
✅ 78% of orgs have implemented DevOps
❌ Developers lose 6-15 hours/week to tool sprawl
📊 Average team uses 7.4 different tools
The solution? Platform engineering.
💡 Gartner predicts: 30% faster delivery for orgs that adopt it.
👉 Learn more at talk-nerdy-to-me.com
December 9, 2025 at 5:01 PM
The DevOps paradox of 2025 ☁️
✅ 78% of orgs have implemented DevOps
❌ Developers lose 6-15 hours/week to tool sprawl
📊 Average team uses 7.4 different tools
The solution? Platform engineering.
💡 Gartner predicts: 30% faster delivery for orgs that adopt it.
👉 Learn more at talk-nerdy-to-me.com
✅ 78% of orgs have implemented DevOps
❌ Developers lose 6-15 hours/week to tool sprawl
📊 Average team uses 7.4 different tools
The solution? Platform engineering.
💡 Gartner predicts: 30% faster delivery for orgs that adopt it.
👉 Learn more at talk-nerdy-to-me.com
Weekend Deep Dive: AWS DevOps Agent - your autonomous on-call engineer.
Analyzes CloudWatch, GitHub, ServiceNow to find root causes and coordinate incident response automatically.
No more 3am pages? The future of ops is here.
#DevOps #AWS #AIOps #reInvent2025
Analyzes CloudWatch, GitHub, ServiceNow to find root causes and coordinate incident response automatically.
No more 3am pages? The future of ops is here.
#DevOps #AWS #AIOps #reInvent2025
December 6, 2025 at 12:00 PM
Weekend Deep Dive: AWS DevOps Agent - your autonomous on-call engineer.
Analyzes CloudWatch, GitHub, ServiceNow to find root causes and coordinate incident response automatically.
No more 3am pages? The future of ops is here.
#DevOps #AWS #AIOps #reInvent2025
Analyzes CloudWatch, GitHub, ServiceNow to find root causes and coordinate incident response automatically.
No more 3am pages? The future of ops is here.
#DevOps #AWS #AIOps #reInvent2025
🔧 Cloud Security Essentials: Lambda Durable Functions!
Multi-step workflows that:
- Run up to 1 year
- Auto-checkpoint & recover
- Suspend without paying for idle
- Support Python 3.13/3.14 & Node 22/24
Finally, reliable long-running serverless! No Step Functions required ⚡
#AWS #Lambda
Multi-step workflows that:
- Run up to 1 year
- Auto-checkpoint & recover
- Suspend without paying for idle
- Support Python 3.13/3.14 & Node 22/24
Finally, reliable long-running serverless! No Step Functions required ⚡
#AWS #Lambda
December 5, 2025 at 9:00 AM
🚀 Future Forward: Meet AWS Kiro - the autonomous agent that codes for days!
Amazon's internal test:
- 18 months → 76 days
- 30 devs → 6 devs
It learns from PRs, maintains context across sessions, and works your backlog while you sleep 🤯
The future of software development is here.
#AWS
Amazon's internal test:
- 18 months → 76 days
- 30 devs → 6 devs
It learns from PRs, maintains context across sessions, and works your backlog while you sleep 🤯
The future of software development is here.
#AWS
December 4, 2025 at 11:01 AM
🚀 Future Forward: Meet AWS Kiro - the autonomous agent that codes for days!
Amazon's internal test:
- 18 months → 76 days
- 30 devs → 6 devs
It learns from PRs, maintains context across sessions, and works your backlog while you sleep 🤯
The future of software development is here.
#AWS
Amazon's internal test:
- 18 months → 76 days
- 30 devs → 6 devs
It learns from PRs, maintains context across sessions, and works your backlog while you sleep 🤯
The future of software development is here.
#AWS
🤓 Wacky Wednesday: Lambda on EC2? AWS is getting weird!
Lambda Managed Instances = serverless simplicity + EC2 hardware access.
Need GPUs for your function? Custom chips? Now possible while keeping that serverless DX we love! ⚡
The line between serverless and servers just got blurrier 🤔
#AWS
Lambda Managed Instances = serverless simplicity + EC2 hardware access.
Need GPUs for your function? Custom chips? Now possible while keeping that serverless DX we love! ⚡
The line between serverless and servers just got blurrier 🤔
#AWS
December 3, 2025 at 5:04 PM
🤓 Wacky Wednesday: Lambda on EC2? AWS is getting weird!
Lambda Managed Instances = serverless simplicity + EC2 hardware access.
Need GPUs for your function? Custom chips? Now possible while keeping that serverless DX we love! ⚡
The line between serverless and servers just got blurrier 🤔
#AWS
Lambda Managed Instances = serverless simplicity + EC2 hardware access.
Need GPUs for your function? Custom chips? Now possible while keeping that serverless DX we love! ⚡
The line between serverless and servers just got blurrier 🤔
#AWS
☁️ AI in the Cloud: Microsoft Foundry is the agentic platform we needed!
From Ignite 2025:
- MCP tool catalogue (finally standardized!)
- Multi-agent runtime with memory
- Control Plane for enterprise governance
Azure now hosts 11,000+ AI models. Pick your fighter! 🚀
#Azure
👉 talk-nerdy-to-me.com
From Ignite 2025:
- MCP tool catalogue (finally standardized!)
- Multi-agent runtime with memory
- Control Plane for enterprise governance
Azure now hosts 11,000+ AI models. Pick your fighter! 🚀
#Azure
👉 talk-nerdy-to-me.com
December 3, 2025 at 1:45 PM
☁️ AI in the Cloud: Microsoft Foundry is the agentic platform we needed!
From Ignite 2025:
- MCP tool catalogue (finally standardized!)
- Multi-agent runtime with memory
- Control Plane for enterprise governance
Azure now hosts 11,000+ AI models. Pick your fighter! 🚀
#Azure
👉 talk-nerdy-to-me.com
From Ignite 2025:
- MCP tool catalogue (finally standardized!)
- Multi-agent runtime with memory
- Control Plane for enterprise governance
Azure now hosts 11,000+ AI models. Pick your fighter! 🚀
#Azure
👉 talk-nerdy-to-me.com
What if you don't need Kubernetes? 🤔
Only 30% of devs still use K8s directly. The rest? Choosing simplicity:
☁️ Docker Swarm (easy mode)
🚀 Nomad (flexibility)
⚡ Cycle.io's LowOps (zero cluster headaches)
Sometimes less complexity = more productivity.
👉 Learn more at talk-nerdy-to-me.com
Only 30% of devs still use K8s directly. The rest? Choosing simplicity:
☁️ Docker Swarm (easy mode)
🚀 Nomad (flexibility)
⚡ Cycle.io's LowOps (zero cluster headaches)
Sometimes less complexity = more productivity.
👉 Learn more at talk-nerdy-to-me.com
November 27, 2025 at 7:04 PM
What if you don't need Kubernetes? 🤔
Only 30% of devs still use K8s directly. The rest? Choosing simplicity:
☁️ Docker Swarm (easy mode)
🚀 Nomad (flexibility)
⚡ Cycle.io's LowOps (zero cluster headaches)
Sometimes less complexity = more productivity.
👉 Learn more at talk-nerdy-to-me.com
Only 30% of devs still use K8s directly. The rest? Choosing simplicity:
☁️ Docker Swarm (easy mode)
🚀 Nomad (flexibility)
⚡ Cycle.io's LowOps (zero cluster headaches)
Sometimes less complexity = more productivity.
👉 Learn more at talk-nerdy-to-me.com
AI workloads are making platform engineering cool again!
🚀 KubeCon NA just dropped new CNCF certs as platform teams master the art of hiding Kubernetes complexity.
Plot twist: Only 30% of backend devs directly use K8s now. The abstraction is working! 🤓☁️
#KubeCon2025
👉 talk-nerdy-to-me.com
🚀 KubeCon NA just dropped new CNCF certs as platform teams master the art of hiding Kubernetes complexity.
Plot twist: Only 30% of backend devs directly use K8s now. The abstraction is working! 🤓☁️
#KubeCon2025
👉 talk-nerdy-to-me.com
November 27, 2025 at 3:03 PM
AI workloads are making platform engineering cool again!
🚀 KubeCon NA just dropped new CNCF certs as platform teams master the art of hiding Kubernetes complexity.
Plot twist: Only 30% of backend devs directly use K8s now. The abstraction is working! 🤓☁️
#KubeCon2025
👉 talk-nerdy-to-me.com
🚀 KubeCon NA just dropped new CNCF certs as platform teams master the art of hiding Kubernetes complexity.
Plot twist: Only 30% of backend devs directly use K8s now. The abstraction is working! 🤓☁️
#KubeCon2025
👉 talk-nerdy-to-me.com
Terraform + watsonx AI Integration ☁️
Post-IBM acquisition, Terraform is getting an AI upgrade. watsonx-powered provisioning means natural language → infrastructure code.
Post-IBM acquisition, Terraform is getting an AI upgrade. watsonx-powered provisioning means natural language → infrastructure code.
November 26, 2025 at 4:00 PM
Terraform + watsonx AI Integration ☁️
Post-IBM acquisition, Terraform is getting an AI upgrade. watsonx-powered provisioning means natural language → infrastructure code.
Post-IBM acquisition, Terraform is getting an AI upgrade. watsonx-powered provisioning means natural language → infrastructure code.
Opus 4.5 peak performance: 4 iterations vs 10 🎯
Anthropic's new model reaches optimal solutions for complex multi-system bugs in just 4 attempts—while competing LLMs require 10 iterations.
Anthropic's new model reaches optimal solutions for complex multi-system bugs in just 4 attempts—while competing LLMs require 10 iterations.
November 25, 2025 at 7:02 PM
Opus 4.5 peak performance: 4 iterations vs 10 🎯
Anthropic's new model reaches optimal solutions for complex multi-system bugs in just 4 attempts—while competing LLMs require 10 iterations.
Anthropic's new model reaches optimal solutions for complex multi-system bugs in just 4 attempts—while competing LLMs require 10 iterations.
Claude Code expands to Mac desktop app! 💻
Anthropic just made terminal-native AI development available in the desktop app for the first time—previously limited to web and mobile.
Anthropic just made terminal-native AI development available in the desktop app for the first time—previously limited to web and mobile.
November 25, 2025 at 10:00 AM
Claude Code expands to Mac desktop app! 💻
Anthropic just made terminal-native AI development available in the desktop app for the first time—previously limited to web and mobile.
Anthropic just made terminal-native AI development available in the desktop app for the first time—previously limited to web and mobile.
Platform Engineering is now a boardroom priority in 2025 🚀
What started as DevOps evolution is now business-critical:
• GitOps for deployment automation
• Policy-as-code for governance
• Internal developer platforms for self-service
What started as DevOps evolution is now business-critical:
• GitOps for deployment automation
• Policy-as-code for governance
• Internal developer platforms for self-service
November 24, 2025 at 7:00 PM
Platform Engineering is now a boardroom priority in 2025 🚀
What started as DevOps evolution is now business-critical:
• GitOps for deployment automation
• Policy-as-code for governance
• Internal developer platforms for self-service
What started as DevOps evolution is now business-critical:
• GitOps for deployment automation
• Policy-as-code for governance
• Internal developer platforms for self-service
Kubernetes becomes THE standard for AI/ML workloads in 2025 ☁️
With 90%+ adoption, K8s is now managing:
• GPU resource scheduling
• Dynamic allocation for ML training
• Enterprise AI infrastructure at scale
It's not just container orchestration anymore—it's the foundation of AI infrastructure. 🤓
With 90%+ adoption, K8s is now managing:
• GPU resource scheduling
• Dynamic allocation for ML training
• Enterprise AI infrastructure at scale
It's not just container orchestration anymore—it's the foundation of AI infrastructure. 🤓
November 24, 2025 at 3:01 PM
Kubernetes becomes THE standard for AI/ML workloads in 2025 ☁️
With 90%+ adoption, K8s is now managing:
• GPU resource scheduling
• Dynamic allocation for ML training
• Enterprise AI infrastructure at scale
It's not just container orchestration anymore—it's the foundation of AI infrastructure. 🤓
With 90%+ adoption, K8s is now managing:
• GPU resource scheduling
• Dynamic allocation for ML training
• Enterprise AI infrastructure at scale
It's not just container orchestration anymore—it's the foundation of AI infrastructure. 🤓
🌀 Breaking: Azure AI Foundry Agent Service reaches GA at Microsoft Ignite 2025!
Design, deploy, and scale enterprise AI agents with multi-agent orchestration. MCP and Agent2Agent protocols ensure your bots play nice together.
Design, deploy, and scale enterprise AI agents with multi-agent orchestration. MCP and Agent2Agent protocols ensure your bots play nice together.
November 19, 2025 at 4:02 PM
🌀 Breaking: Azure AI Foundry Agent Service reaches GA at Microsoft Ignite 2025!
Design, deploy, and scale enterprise AI agents with multi-agent orchestration. MCP and Agent2Agent protocols ensure your bots play nice together.
Design, deploy, and scale enterprise AI agents with multi-agent orchestration. MCP and Agent2Agent protocols ensure your bots play nice together.
🚀 The Rise of TACOS in 2025 (and no, not the food kind).
TACOS = Terraform Automation and Collaboration Software. Think Atlantis (open-source), Spacelift (multi-IaC), Scalr (enterprise).
TACOS = Terraform Automation and Collaboration Software. Think Atlantis (open-source), Spacelift (multi-IaC), Scalr (enterprise).
November 18, 2025 at 3:03 PM
🚀 The Rise of TACOS in 2025 (and no, not the food kind).
TACOS = Terraform Automation and Collaboration Software. Think Atlantis (open-source), Spacelift (multi-IaC), Scalr (enterprise).
TACOS = Terraform Automation and Collaboration Software. Think Atlantis (open-source), Spacelift (multi-IaC), Scalr (enterprise).
🔋 TechStack Deep-Dive: KEDA (Kubernetes Event-Driven Autoscaling) reached CNCF graduation in 2025.
What makes it special? HPA scales on CPU/memory. KEDA scales on ANYTHING - Kafka lag, Redis queue depth, PostgreSQL connections, HTTP requests.
What makes it special? HPA scales on CPU/memory. KEDA scales on ANYTHING - Kafka lag, Redis queue depth, PostgreSQL connections, HTTP requests.
November 18, 2025 at 10:01 AM
🔋 TechStack Deep-Dive: KEDA (Kubernetes Event-Driven Autoscaling) reached CNCF graduation in 2025.
What makes it special? HPA scales on CPU/memory. KEDA scales on ANYTHING - Kafka lag, Redis queue depth, PostgreSQL connections, HTTP requests.
What makes it special? HPA scales on CPU/memory. KEDA scales on ANYTHING - Kafka lag, Redis queue depth, PostgreSQL connections, HTTP requests.
🚀 2025's DevOps revolution? AI integration everywhere.
What does this mean? Automated optimization, intelligent resource allocation, and AI-powered tools enhancing every aspect of operations.
What does this mean? Automated optimization, intelligent resource allocation, and AI-powered tools enhancing every aspect of operations.
November 17, 2025 at 10:00 AM
🚀 2025's DevOps revolution? AI integration everywhere.
What does this mean? Automated optimization, intelligent resource allocation, and AI-powered tools enhancing every aspect of operations.
What does this mean? Automated optimization, intelligent resource allocation, and AI-powered tools enhancing every aspect of operations.
GitOps is becoming the Kubernetes standard in 2025 🚀
ArgoCD automatically syncs your clusters with Git repos.
What's in Git = what's running in production. Period.
✅ Automated rollbacks
✅ Self-healing deployments
✅ Complete audit trails
#GitOps #Kubernetes
ArgoCD automatically syncs your clusters with Git repos.
What's in Git = what's running in production. Period.
✅ Automated rollbacks
✅ Self-healing deployments
✅ Complete audit trails
#GitOps #Kubernetes
Demystify Cloud, FinOps & AI with Talk Nerdy To Me
Step-by-step tutorials, real-world cost playbooks and witty analogies for Azure, OCI and multi-cloud pros.
talk-nerdy-to-me.com
November 15, 2025 at 1:01 PM
GitOps is becoming the Kubernetes standard in 2025 🚀
ArgoCD automatically syncs your clusters with Git repos.
What's in Git = what's running in production. Period.
✅ Automated rollbacks
✅ Self-healing deployments
✅ Complete audit trails
#GitOps #Kubernetes
ArgoCD automatically syncs your clusters with Git repos.
What's in Git = what's running in production. Period.
✅ Automated rollbacks
✅ Self-healing deployments
✅ Complete audit trails
#GitOps #Kubernetes
Mistral just solved Europe's AI catch-22: self-hosted deployment with full data sovereignty. 🚀
Run Medium 3 (matches Claude Sonnet 3.7) entirely on your infrastructure. Finance, healthcare, government - keep data in YOUR data center while using cutting-edge models. ☁️
Run Medium 3 (matches Claude Sonnet 3.7) entirely on your infrastructure. Finance, healthcare, government - keep data in YOUR data center while using cutting-edge models. ☁️
November 13, 2025 at 2:02 PM
Mistral just solved Europe's AI catch-22: self-hosted deployment with full data sovereignty. 🚀
Run Medium 3 (matches Claude Sonnet 3.7) entirely on your infrastructure. Finance, healthcare, government - keep data in YOUR data center while using cutting-edge models. ☁️
Run Medium 3 (matches Claude Sonnet 3.7) entirely on your infrastructure. Finance, healthcare, government - keep data in YOUR data center while using cutting-edge models. ☁️
Wacky Wednesday Cloud Fact ☁️🌀
Did you know? AWS S3 stores more data than every library on Earth combined. We're talking exabytes of cat videos, forgotten backups, and buckets named "test-delete-later" from 2018.
💡 Pro Tip: Use lifecycle policies or you're paying to store digital junk forever.
Did you know? AWS S3 stores more data than every library on Earth combined. We're talking exabytes of cat videos, forgotten backups, and buckets named "test-delete-later" from 2018.
💡 Pro Tip: Use lifecycle policies or you're paying to store digital junk forever.
November 12, 2025 at 7:02 PM
Wacky Wednesday Cloud Fact ☁️🌀
Did you know? AWS S3 stores more data than every library on Earth combined. We're talking exabytes of cat videos, forgotten backups, and buckets named "test-delete-later" from 2018.
💡 Pro Tip: Use lifecycle policies or you're paying to store digital junk forever.
Did you know? AWS S3 stores more data than every library on Earth combined. We're talking exabytes of cat videos, forgotten backups, and buckets named "test-delete-later" from 2018.
💡 Pro Tip: Use lifecycle policies or you're paying to store digital junk forever.
When Your AI Model Calls in Sick ☁️
AI models don't get PTO, but they DO crash. Auto-scaling = infinite backup singers—one goes down, another steps up.
💡 Configure health checks BEFORE your model naps during peak traffic.
👉 talk-nerdy-to-me.com/news
#CloudEngineering #AI #TalkNerdyToMe
AI models don't get PTO, but they DO crash. Auto-scaling = infinite backup singers—one goes down, another steps up.
💡 Configure health checks BEFORE your model naps during peak traffic.
👉 talk-nerdy-to-me.com/news
#CloudEngineering #AI #TalkNerdyToMe
November 12, 2025 at 3:02 PM
When Your AI Model Calls in Sick ☁️
AI models don't get PTO, but they DO crash. Auto-scaling = infinite backup singers—one goes down, another steps up.
💡 Configure health checks BEFORE your model naps during peak traffic.
👉 talk-nerdy-to-me.com/news
#CloudEngineering #AI #TalkNerdyToMe
AI models don't get PTO, but they DO crash. Auto-scaling = infinite backup singers—one goes down, another steps up.
💡 Configure health checks BEFORE your model naps during peak traffic.
👉 talk-nerdy-to-me.com/news
#CloudEngineering #AI #TalkNerdyToMe
Reserved Instances vs. Savings Plans
Reserved Instances lock in discounts for predictable workloads. Savings Plans give flexibility across compute families. Use RIs for databases, Savings Plans for variable loads. Stack both for maximum AWS/Azure savings 🤓 #FinOps #CostOptimization
Reserved Instances lock in discounts for predictable workloads. Savings Plans give flexibility across compute families. Use RIs for databases, Savings Plans for variable loads. Stack both for maximum AWS/Azure savings 🤓 #FinOps #CostOptimization
November 11, 2025 at 7:08 PM
Reserved Instances vs. Savings Plans
Reserved Instances lock in discounts for predictable workloads. Savings Plans give flexibility across compute families. Use RIs for databases, Savings Plans for variable loads. Stack both for maximum AWS/Azure savings 🤓 #FinOps #CostOptimization
Reserved Instances lock in discounts for predictable workloads. Savings Plans give flexibility across compute families. Use RIs for databases, Savings Plans for variable loads. Stack both for maximum AWS/Azure savings 🤓 #FinOps #CostOptimization
TechStack Tuesday: Terraform Modules ⚡
Copying the same VPC config for the 10th time? Stop. Build reusable Terraform modules instead.
Write once → deploy everywhere → maintain in one place.
💡 Pro Tip: Version your modules with semantic versioning (v1.0.0, v1.1.0) to track changes safely.
#Terraform
Copying the same VPC config for the 10th time? Stop. Build reusable Terraform modules instead.
Write once → deploy everywhere → maintain in one place.
💡 Pro Tip: Version your modules with semantic versioning (v1.0.0, v1.1.0) to track changes safely.
#Terraform
November 11, 2025 at 3:03 PM
TechStack Tuesday: Terraform Modules ⚡
Copying the same VPC config for the 10th time? Stop. Build reusable Terraform modules instead.
Write once → deploy everywhere → maintain in one place.
💡 Pro Tip: Version your modules with semantic versioning (v1.0.0, v1.1.0) to track changes safely.
#Terraform
Copying the same VPC config for the 10th time? Stop. Build reusable Terraform modules instead.
Write once → deploy everywhere → maintain in one place.
💡 Pro Tip: Version your modules with semantic versioning (v1.0.0, v1.1.0) to track changes safely.
#Terraform