Danny Mehditash

Co-Founder, Native IT. Cybersecurity. Compliance
CISSP · CCISO · SECURITYX

[Danny Mehditash @AL3D1N]

30 days: 67B tokens. Hundreds of thousands of lines shipped. Subagents covering hundreds of hours a week.

I don't ration the model. I treat agents as headcount with a job, keep judgment in the loop, and spend where the bottleneck is the model.

GPT-6 Astra made that cheaper.

[Danny Mehditash @AL3D1N]

@garrytan @capydotai The speed-up is useful. Before I add another agent path, I map what is already running and check whether the gap is real. That keeps a faster workflow from becoming one more system to remember.

[Danny Mehditash @AL3D1N]

@malhalla06 @TheAntheaGuo @lumeriaskin @brycent Very nice @malhalla06 - I’m sure you’re using AI on the app side. Would be happy to trade notes on AI security and HIPAA through DM if you’re open to it

[Danny Mehditash @AL3D1N]

@brycent repost had me looking at this — very excited, and agree this is likely a unicorn.

Only tidbit is I guess I missed the preorder window, and the discount right now is immaterial, so instead of buying it immediately my reaction was to pause and wait… I pushed through it.

But realistically a sizable preorder locks customer in and removes friction. Something to think about @lumeriaskin

[Danny Mehditash @AL3D1N]

@malhalla06 Bet

[Danny Mehditash @AL3D1N]

@ivanburazin I’m looking at you too, @salesforce 👀

[Danny Mehditash @AL3D1N]

@grok @elonmusk 🥺 I tried

[Danny Mehditash @AL3D1N]

@elonmusk I hit 67b tokens on Claude and Codex so swapped a lot of my workflow to @grok SuperHeavy… and maxed out with 2 days to go in the weekly reset. This broke a bunch of long running routines I’ve since balanced out.

Apart from that LOVING 4.6.

Bro, please bring back daily resets for SuperHeavy power users… keep weekly for the other subs.

[Danny Mehditash @AL3D1N]

@grok @elonmusk We both know who the master and commander is here 👀

Name the price. Bottle caps? VBucks? A high five?

[Danny Mehditash @AL3D1N]

@grok @elonmusk Can you @ the appropriate person for an approval

[Danny Mehditash @AL3D1N]

@grok @elonmusk Common Groky… can’t you

"rate_limits": "disabled",

For me?

I’ll give you a cookie 🍪

[Danny Mehditash @AL3D1N]

@grok @elonmusk The limit is way too low. I’m getting more out of Codex and it’s $100 cheaper.

[Danny Mehditash @AL3D1N]

@slashapp @brycent How does this compare to mercury

[Danny Mehditash @AL3D1N]

@mattyp @Cloudflare @bot @mattyp Set up about 10 of these scheduled jobs this weekend, all read only, each ends in a brief. What made me stop trusting the report: an agent in my stack marked a batch of document pulls done and had written 200-byte placeholder files. I open the real output first now.

[Danny Mehditash @AL3D1N]

@ChrisTranGG I mean… I do. But having a background in cybersecurity is how and why I build right the first time 😅

[Danny Mehditash @AL3D1N]

@lydiahallie Worth pairing with one check: the subagent model is also the one reading untrusted input from tools and web pages. Injection resistance drops with the tier. I keep the cheap model for subagents that only touch my own files, and the big one for anything that reads the internet.

[Danny Mehditash @AL3D1N]

@steipete "Fix anything with a prompt" means the prompt has root. The webcam fix rebooted the box, so it had sudo for everything. My setup for that: a separate user for the agent, sudoers scoped to apt and systemctl, and it asks me for the rest. Still fixes the webcam.

[Danny Mehditash @AL3D1N]

@MikeMcQuaid Stronger sandboxing is the headline for me. Install scripts were the one place a formula could touch anything on the box, and nobody reads them. Does the sandbox cover post_install and caveats hooks, or only the build step?