Cloud Quality Engineer | SDET | Cloud & DevOps Engineer | Automation & Cloud-Native Systems
Joined December 2022
- Tweets692
- Following402
- Followers82
- Likes1.8K
This is a good reminder that more expensive infrastructure doesn't always mean better reliability. Sometimes a simpler setup, managed well, is all you need. The AI handling incidents while the team sleeps is what really caught my attention.
We used to pay AWS close to $8,000 a month.
Today, our entire company runs on 3 rented bare metal servers at Hetzner. That's 4 SaaS products with more than 50,000 users (apps, databases, background jobs and monitoring included).
The bill is around $300 a month.
Everyone told me it would be unstable. It's been the opposite. Our uptime is close to 100%.
The other day, we had about 15 seconds of downtime. An AI agent caught the alarm, fixed it and sent me the report on Telegram. I read it when I woke up, and not a single user noticed 👇
Something worth thinking about. As AI keeps growing, will compute capacity become one of the biggest challenges businesses have to deal with?
AI agents in DevOps sound promising, but production reliability still makes me think.
If one makes a bad call at 2 a.m., who owns the blame? Who approves the fix and hits the rollback?
Curious how teams are handling guardrails and accountability right now.
How is this even possibly
Introducing Griffin, the first model to pass the video Turing test.
48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video.
It’s the first Human Interaction Model (HIM).
Readers added context they thought people might want to know
The 48% figure and "video Turing test" claim are from Tavus's own study of 54 one-minute calls, not independently verified or using a standard protocol. Griffin-Lite leads NVIDIA's VideoFDB benchmark on their public leaderboard.
cellcog.ai/blog/tavus-gri…
research.nvidia.com/labs/amri/proj…
tech-ish.com/2026/10/02/tav…
Daniel retweeted
introducing penna, an open source newsletter platform
it’s MIT licensed, has an API for devs, and was built by one guy who refused to pay for a newsletter tool
penna.dev
Parts of the Bahrain region were physically damaged during the war in March.
Six months later, AWS says it cannot restore the region because the damage affected multiple Availability Zones and exceeded the resilience AWS had designed for.
Cloud has limits too.
1/7 — THREAD
The Bamboo outage got me thinking about how we talk about cloud systems.
An outage happens and suddenly everyone knows what the engineers “should have done.”
Add more servers. Add Redis. Use Kubernetes. Autoscale.
But real systems aren't that simple.
Hello everyone,
First, we’re truly sorry. Yesterday, many of you struggled to access the Bamboo app, fund your accounts or invest.
We are also really disappointed that we couldn't give you a seamless experience to start off your investment journey into the Dangote Refinery IPO. We know we let you down, and we take full responsibility, and we're going to work to make it up to you. Our team worked hard to prepare for the increased demand but ultimately, our efforts fell significantly short.
We wanted to let you know that the Bamboo app is fully back up and running. We restored the systems yesterday but wanted to make sure everything was working perfectly before making an announcement.
We’ve also made changes to ensure our platform is stronger and better prepared moving forward.
We know an apology doesn’t fix the frustration you experienced. Trust is earned through action, and we’re committed to earning yours back.
To start, we’re going to cut all commissions across both US and NGX stocks by 20% starting tomorrow till the end of September.
We also wanted to remind you that the Dangote IPO is running non-stop till the 13th of October. The price doesn’t change and the allocation will not be done on a first-come, first-served basis so there’s still time for you to get in.
Thank you for your patience and for sticking with us. 💚
Richmond,
CEO, Bamboo
6/7
And no system is perfect.
We've seen major platforms go down because of traffic spikes, configuration mistakes, dependency failures, network issues and unexpected bottlenecks.
Cloud gives you better tools to handle failure.
It doesn't remove failure.
