Azlan

DevOps engineer using AI agents to build, run and fix real infrastructure.

Building a desktop app for people who use too many AI coding agents (me).

[Azlan @azlan_dev]

I pay for four AI subscriptions and I still hit limits by lunch.

Hi, I'm Azlan 👋

DevOps engineer, using AI agents on real infra every day.

So I post what I learn:
- Which agent wins at what
- Tools worth trying
- Workflows that actually ship
- Making your limits last longer

Let's connect if you're in the same mess ⚡

[Azlan @azlan_dev]

@Da7_Tech I have never used Devin. What does it actually do? Are they a model provider or Harness?

[Azlan @azlan_dev]

@bijanbowen True. Remember it was AGI for a week 😁

[Azlan @azlan_dev]

@pmitu Never mind, I'll ask the other model 😶

[Azlan @azlan_dev]

Gemini 4 Hydrogen didn't beat GPT 6.1 Sol at anything.

- Less capable.
- Slower in my tests.
- 3x more expensive per task.

But that doesn't make it a bad model.

It might be the new winner for the $20 plan.

The usage on Google Pro feels endless, and Argon gets its own separate limits.

Sol for real work. Hydrogen for testing, rough work and quick tasks.

Google Pro is totally worth it now 🎁

[Azlan @azlan_dev]

Built the GPUs for everyone else. Finally gets to use one 🕹️

[Azlan @azlan_dev]

Benchmarkts are from:
Artificial Analysis - Intelligence Index
https://t.co/5MFJSlQGBN

[Azlan @azlan_dev]

Meta's frontier model just lost to a Flash model.

On high, not max.

Muse Wonder Full: 50
Gemini 4 Hydrogen: 51

And you still can't even use Muse 💀

[Azlan @azlan_dev]

@argofowl We stayed up waiting for Astra and got a cyber lecture instead 😴

[Azlan @azlan_dev]

In this industry a month of waiting turns your frontier model into a legacy model 🪦

[Azlan @azlan_dev]

@argofowl @Angaisb_ Wild that it works. Try that on Claude and Anthropic shows up at your door 🚔

[Azlan @azlan_dev]

@kimmonismus Faster than Flash and smarter than Argon. Argon really didn't get to enjoy being the flagship 😅

[Azlan @azlan_dev]

@yigittbuilds Too early to say, testing it right now. 3.8 Flash usage on Pro already feels unlimited, so if Hydrogen on xhigh holds up on bigger refactors that's a big deal. Will share results soon 🧪

[Azlan @azlan_dev]

Gemini 4 Hydrogen just showed up in my CLI.

On the Pro plan.
On launch day.
With xhigh effort.

Google really said no waitlist this time 👀

[Azlan @azlan_dev]

@lyraxana Gemini finally gets xhigh and now I have to decide if my bug deserves it 🤔

[Azlan @azlan_dev]

@kimmonismus A long USB C cable. The only free outlet is always behind the one chair you can't move 🔌

[Azlan @azlan_dev]

@Da7_Tech Consulted the leather jacket in my closet and it said it's retiring too 🧥

[Azlan @azlan_dev]

Nvidia never had a board chair in 33 years.

It got its first one today.

The entire AI boom ran through one guy in a leather jacket.

One resignation and half the AI world is refreshing the news 😬

[Azlan @azlan_dev]

@thdxr Babe I believe you, but why does your browser history say H200 spec sheet pdf 🤨

[Azlan @azlan_dev]

@Georgia_07os So happy for you! Enjoy every bit of this moment ❤️

[Azlan @azlan_dev]

@kaireceipts 5h limits are brutal for heavy users😅

[Azlan @azlan_dev]

I pay for four AI subscriptions and I still hit limits by lunch.

Hi, I'm Azlan 👋

DevOps engineer, using AI agents on real infra every day.

So I post what I learn:
- Which agent wins at what
- Tools worth trying
- Workflows that actually ship
- Making your limits last longer

Let's connect if you're in the same mess ⚡