Dinesh Pandiyan

Architect @thethinkmill. I moonlight as a homelabber.

[Dinesh Pandiyan @dpandiyan]

The things that are worth doing are the things that you suck at.

Kubernetes has been a humbling year-long journey. Embracing the suck was not fun, but I'm now prepared to ship anything anywhere. The role SRE sounds a lot more appealing now than a year before.

[Dinesh Pandiyan @dpandiyan]

@siddharthkp this is how most companies (at least the ones that do it right) run interviews these days

[Malte Ubl @cramforce]

@0xBunny Candidates need to demonstrate that they
- Could write code without agentic help
- Can write code with agentic help

[Dinesh Pandiyan @dpandiyan]

@DanielW_Kiwi ansible is delightful to work with when there's an agent help build playbooks. i am having a blast with ansible this year compared to the terrible experience i had 3 years ago.

[Simon Willison @simonw]

Uber reportedly now caps coding agents at $1,500/month per employee per tool - seems sensible to me, but it's also an interesting hint at the value Uber thinks these tools are providing
https://t.co/6YT0lCzPml

[Dinesh Pandiyan @dpandiyan]

github repo here
https://t.co/ZApPbxIKgX

[Dinesh Pandiyan @dpandiyan]

I work on the same set of repos on multiple machines and constantly have to pull multiple times a day. So I built a cli to automate it. I've name it gitsy.

gitsy —sync

automatically pulls all git repos and worktrees in a dir.

[Dinesh Pandiyan @dpandiyan]

I've written about it here
https://t.co/Ro2OK5MAbX

[Dinesh Pandiyan @dpandiyan]

@burkeholland I am trying to find this info online but couldn’t. Is it allowed or against the terms to use copilot subscription in opencode or pi?

[Gergely Orosz @GergelyOrosz]

It’s wild how Meta - a company going all-in on AI - somehow missed the memo on how AI can generate images and videos that renders “take a selfie of yourself” verifications utterly useless

So now Instagram accounts hacked at scale. 2FA also fully bypassed - by Meta’s own design

[Dinesh Pandiyan @dpandiyan]

@thsottiaux mini on bedrock when?

[Dinesh Pandiyan @dpandiyan]

@thsottiaux Excellent timing. We've just been exploring bedrock and not having GPT models on bedrock was a bummer. Looks like that's not a problem anymore.

[Tibo @thsottiaux]

Heard that AWS is where the cool kids are. Hello. We have GPT-5.5.

[Corey Quinn @QuinnyPig]

Hello and welcome to #MicrosoftBuild, friends.

[TANSTACK @tan_stack]

New @Lovable projects now ship on TanStack Start.

That means full-stack React with SSR/SSG/CSR per route, server functions colocated with components, deploy-anywhere architecture, and type safety that helps both humans and AI generate better apps.

1/2

https://t.co/Vb962C0h3P

[fks @FredKSchott]

Huge win of support for @tannerlinsley and the entire Tanstack ecosystem.

Tanner is one of the few people I know who watches npm download charts even more closely than I do, so I know he's got to be feeling good about this rn.

Well deserved!

[Dinesh Pandiyan @dpandiyan]

@SashaRakovets @thdxr Oh wow it’s right there. Thanks.

[Dinesh Pandiyan @dpandiyan]

@thdxr is there a way to debug what goes into my context whenever i fire up an opencode session? it seems to start at 15k+ tokens for me and I am not sure how much comes from tool call definitions, system prompts, my custom AGENTS.md and skills metadata.

[Dinesh Pandiyan @dpandiyan]

yeehaw

[Tibo @thsottiaux]

The Codex usage limits have been reset for all paid ChatGPT subscriptions. You should be back to 100% weekly and 100% hourly limits.

Let the tokens do incredible things today and have fun.

[Tibo @thsottiaux]

Five million users would agree. Resetting the limits tomorrow morning to celebrate.

Time to go /fast

[Tibo @thsottiaux]

The Codex usage limits have been reset for all paid ChatGPT subscriptions. You should be back to 100% weekly and 100% hourly limits.

Let the tokens do incredible things today and have fun.

[Dinesh Pandiyan @dpandiyan]

@dexhorthy within days? try hours.

[David K 🎹 @DavidKPiano]

Unpopular opinion: just because companies can ship features faster w/ AI doesn't mean they should.

Slow down.

Use the saved to actually test the feature: go through the edge cases, put it in front of real humans, refine it, and actually consider whether it's even a good idea.