TechVibe

Question: How living becomes expensive Me: My PC price just doubled from what it was when I got it😭(2.5 years ago btw) @devwitheyob #random #pc

Wtf😭😂 @devwitheyob #random

I wanted to say 21 things about my 21 ግን ገና ሳስበው ደከመኝ but few lessons : - the most important decision of your life is the most uncomfortable one - choosing and wanting are completely different things, don't fool yourself - have the uncomfortable conversations early (it may make any of you feel bad, but trust me, it gets worse) - if you want to help, first help yourself out, then you can help others ....

I'm sure you guys realized this. Things aren't normal (they used to be) after corona and TikTok, yemr. 😳

Fayda verify please fix this, this is crazy🙄 @devwitheyob #fayda #bugs

Ene non cafe ሁኜ 1 እየበላሁ መልክ እሰራለሁ😎😭 @devwitheyob #random

What do you think ppl😂 @devwitheyob #random #gbi

async function measureDownload(durationMs = 10000, streams = 4) {
  const controller = new AbortController();
  let bytes = 0;
  const start = performance.now();
  let warmupDone = 0;

  const runStream = async () => {
    const res = await fetch(
      `/api/speedtest/download?bytes=100000000&cb=${crypto.randomUUID()}`,
      { signal: controller.signal, cache: 'no-store' }
    );

    const reader = res.body.getReader();

    while (true) {
      const { done, value } = await reader.read();
      if (done) break;

      const t = performance.now() - start;

      if (t < 1500) {
        warmupDone = bytes;
        continue;
      }

      bytes += value.length;
    }
  };

  const tasks = Array.from({ length: streams }, runStream);

  setTimeout(() => controller.abort(), durationMs);

  await Promise.allSettled(tasks);

  const elapsedSec =
    (performance.now() - start - 1500) / 1000;

  return (bytes * 8) / elapsedSec / 1e6; // Mbps
}

We don't really realize how bad AI-generated code can be if we're not careful. This code was generated by Fable 5.1, one of the most capable models out there, with good context and everything. I simply asked it to write one function because I wanted to understand how it works. I went through the code to understand each flow and found 5+ bugs, 4 bad implementations, and 1 serious breaking issue that only shows up in an edge case. #AIGenerated #Code #AIGenerated #Code

`async function measureDownload(durationMs = 10000, streams = 4) { const controller = new AbortController(); let bytes = 0; const start = performance.now(); let warmupDone = 0; const runStream = async () => { const res = await fetch(/api/speedtest/download?bytes=100000000&cb=${crypto.randomUUID()}, { signal: controller.signal, cache: 'no-store' }); const reader = res.body.getReader(); while (true) { const { done, value } = await reader.read(); if (done) break; const t = performance.now() - start; if (t < 1500) { warmupDone = bytes; continue; } // discard TCP slow-start bytes += value.length; } }; const tasks = Array.from({ length: streams }, runStream); setTimeout(() => controller.abort(), durationMs); await Promise.allSettled(tasks); const elapsedSec = (performance.now() - start - 1500) / 1000; return (bytes * 8) / elapsedSec / 1e6; // Mbps }` We don't really realize how bad AI-generated code can be if we're not careful. This code was generated by Fable 5.1, one of the most capable models out there, with good context and everything. I simply asked it to write one function because I wanted to understand how it works. I went through the code to understand each flow and found 5+ bugs, 4 bad implementations, and 1 serious breaking issue that only shows up in an edge case. #AIGenerated #Code

Announcing Vortex v1.0.0: Cloud-Native Distributed Video Infrastructure Engine I built Vortex, an open-source distributed video transcoding and streaming platform written in pure Go. It ingests raw media streams, coordinates workers via asynchronous message brokers, segments video into multi-bitrate HLS playlists, tracks real-time encoding progress, and executes durable post-processing workflows. Key Features - Multi-Protocol Ingestion: Streaming HTTP multipart upload (32KB constant memory footprint) and high-throughput gRPC client streaming via Protocol Buffers with server reflection. - Distributed Transcoding: Parallel FFmpeg processing pipeline chunking video into 360p, 720p, and 1080p Apple HLS master playlists and transport segments (.ts). - Real-Time Progress: Live WebSocket broadcast streaming millisecond FFmpeg progress directly to web players via Redis Pub/Sub. - Event-Driven Architecture: NATS JetStream 2.10 at-least-once message delivery with durable worker consumer groups. - Durable Workflows: Inngest step-based execution engine with automatic checkpointing for poster extraction, animated GIF preview generation, and webhook delivery. - Defense & Rate Limiting: Caddy edge reverse proxy with unbuffered 500MB payload streaming paired with an atomic Redis sliding-window token bucket rate limiter. - Full-Stack Observability: OpenTelemetry distributed tracing with W3C context propagation across NATS, Prometheus RED metrics, and pre-provisioned Grafana monitoring cockpits. - Cloud-Native Orchestration: Parameterized Helm 3 charts on Kubernetes with non-root multi-stage containers and CPU-driven Horizontal Pod Autoscaling (HPA). Core Tech Stack Go, PostgreSQL (pgx/v5), Redis 7, NATS JetStream, MinIO S3, FFmpeg, Gorilla WebSockets, gRPC, Inngest, OpenTelemetry, Prometheus, Tempo, Caddy, Kubernetes, Helm 3. GitHub Repository: https://github.com/ruhamabek/vortex License: MIT

We did it again❤ Remember we started a highschool reunion last time and this year we had an amazing event once again I had a chance to manage the whole thing and it was a really great time we talk play games, answer some Q&A, and music. @devwitheyob #friends #highschool #memory

This is what I going to do😁

Another small GitHub Actions setting that’s worth knowing

permissions:
  contents: read

This basically tells GitHub “Give this workflow a read-only token.” So even if something bad runs inside your CI(like a bad npm package that excute a backdoor commit), it can't use the GITHUB_TOKEN to push code, modify your repo, or mess with things. It's a simple example of least privilege, only give a workflow the permissions it actually needs. Small line, big security win #TechVibe #GitHubActions #DevOps

Bro is abusing fable

One tiny GitHub Actions config that can save you a lot of CI time and headaches

concurrency:
  group: ci-${{ github.ref }}
  cancel-in-progress: ${{ github.event_name == 'pull_request' }}

Basically, if you push new changes to the same PR while the old CI is still running, GitHub cancels the old run and only tests your latest code. But for deployments, it doesn't cancel the current one, so you don't end up killing a deployment halfway through 💀 Small config, but really useful if your CI takes a while to run. @devwitheyob #TechVibe #DevOps #GitHubActions

After a very long time of procrastination, I fixed the AI model problem on my portfolio. Check it out: eyobsimachew.vercel.app @devwitheyob #portdolio

Dear AAU Community, The Huawei ICT Competition 2026–2027 – Innovation Track is now officially open for registration. This track focuses on developing innovative technology solutions to address real-world challenges. 📅 Registration: August–December 2026 🏆 National Final: January 2027 🌍 Regional Final: February–March 2027 🌐 Global Final: May 2027 🔗 Register here: https://e.huawei.com/en/talent/#/ict/innovation-details?zoneCode=026902&zoneId=98269713&compId=85132021&divisionName=Northern%20Africa&type=C002&isCollectGender=N&enrollmentDeadline=undefined&compTotalApplicantCount=6 Eligible AAU students are encouraged to review the guidelines and apply. @AAUMEREJA

I almost lost everything this year. Not “stressed about deadlines” losing everything. My dad. My family. Everything I’ve been building. There were nights I genuinely didn’t know if RAVN would survive the month. But my cofounder and I kept building anyway. Through the doubt, uncertainty, and nights when nothing seemed to work. Today, I’m incredibly grateful to share that RAVN has been accepted into Antler. $400K in funding, backed by one of the world’s leading venture builders, alongside companies like Lovable, Airalo, Reebelo, and Tactiq. To everyone who watched us build this from the beginning, thank you. This wasn’t luck or a guarantee. We kept building long before there was anything to celebrate. Most importantly, I’m grateful to God. He never left us. We’ve come a long way. We’re just getting started.