ContentPosts from @babually..
Link
@faun shared a link, 1 month ago

Are OpenAI and Anthropic Really Losing Money on Inference?

DeepSeek R1 running on H100s puts input-token costs near$0.003 per million—while output tokens still punch in north of$3. That’s a 1,000x spread. So if a job leans heavy on input—think code linting or parsing big docs—those margins stay fat, even with cautious compute. System shift:This lop-sided ..

Are OpenAI and Anthropic Really Losing Money on Inference?
Link
@faun shared a link, 1 month ago

Some thoughts on LLMs and Software Development

Most LLMs still play autocomplete sidekick. But seasoned devs? They get better results when the model reads and rewrites actual source files. That gap—between how LLMs are designed to work and how prosactuallyuse them—messes with survey data and muddies the picture on real gains in code quality and..

Link
@faun shared a link, 1 month ago

How to prepare for the Bitnami Changes coming soon

The Bitnami team has delayed the deletion of the Bitnami public catalog until September 29th. They will conduct a series of brownouts to prepare users for the upcoming changes, with the affected applications list being published on the day of each brownout. Users are advised to switch to Bitnami Sec..

Link
@faun shared a link, 1 month ago

ECScape: Understanding IAM Privilege Boundaries in Amazon ECS

A new ECS security mess—ECScape—lets low-privileged tasks on EC2 act like the ECS agent. That’s bad. Real bad. Why? Because it opens the door to stealing IAM credentials from other ECS tasks sharing the same host. Here’s the trick: The attacker hits the instance metadata service (IMDS) and fakes a ..

ECScape: Understanding IAM Privilege Boundaries in Amazon ECS
Link
@faun shared a link, 1 month ago

Observability in Go: What Real Engineers Are Saying in 2025

Go observability still feels like pulling teeth. Manual instrumentation? Tedious. Span coverage? Spotty. Telemetry volume? Totally out of hand. Even with OpenTelemetry gaining traction, Go lags behind Java and Python when it comes to auto-instrumentation and clean context propagation. Devs are hunt..

Observability in Go: What Real Engineers Are Saying in 2025
Link
@faun shared a link, 1 month ago

Availability Models: Because “Highly Available” Isn’t Saying Much

Antithesis and Jepsen want to kill hand-wavy "high availability" talk. Instead, they push for clearavailability models—majority,total,sticky, etc.—that spell out when an operationactuallyworks during failures. It's about precision, not platitudes. Why it matters:This reframes availability from a va..

Availability Models: Because “Highly Available” Isn’t Saying Much
Link
@faun shared a link, 1 month ago

Google Develops KFuzzTest For Fuzzing Internal Linux Kernel Functions

Google droppedKFuzzTest, a lean fuzzing tool built to hit Linux kernel internals—way past just syscalls. It brings a clean API, docs, and sample targets to get fuzzing fast. Why it matters:KFuzzTest marks a shift. Kernel fuzzing’s no longer just about hammering syscalls—it’s going deeper into the g..

Link
@faun shared a link, 1 month ago

v1.34: User preferences (kuberc) are available for testing in kubectl 1.34

Kubernetes v1.34 pusheskubectlinto the future with a betauser preferencessystem. Drop a.kubercfile in place, and you can bake in default flags, toggle features likeinteractive deleteorServer-Side Apply, and wire up custom aliases—including pre- and post-args...

Link
@faun shared a link, 1 month ago

Battle for Resources or the SSA Path to Kubernetes Diplomacy

A full-stack engineer and systems architect with hands-on time incloudandIoT, building real-world tools for theoil and gas sector. Think connected rigs, smart pipelines, and infrastructure that doesn’t flinch at scale. Market signal:Industrial tech’s going deep. Cloud and IoT aren’t side projects a..

Link
@faun shared a link, 1 month ago

Kubernetes in an AI-Native World: Can It Stay Relevant?

At KubeCon + CloudNativeCon Hyderabad 2025, CNCF leads made it clear:cloud-native infraisn’t just supporting AI—it’s becoming its backbone. The conversation’s moved on from“Can Kubernetes run AI?”to“How does it evolve for AI-first everything?”..

Kubernetes in an AI-Native World: Can It Stay Relevant?