estaab

estaab

@estaab.dev

Software Engineer. ex-Google, Chrome 2006 - 2026. Hacking on personal projects. Home lab, game dev, agent security.

70 followers160 following42 posts

estaab reposted

bree bird 🇨🇦
@goose.ski

everyone is dooming about AI in game dev but i know in my bones that we are entering into a truly golden age of indie games

estaab
@estaab.dev

Claude Code shows a status line and a separate mode line (e.g. "⏵⏵ auto mode on"); I wanted a line back when working in the terminal on my phone so I wrote a tiny plugin to merge the two footer lines into one. github.com/erikus/statu...

estaab
@estaab.dev

Upstream feature request so this plugin isn't needed: github.com/anthropics/c... Also: I still sometimes use a terminal on my phone for agentic coding because I like the value of a Claude subscription and also want Tailscale's Aperture gateway, which makes Claude Code disable Remote Control.

github.com[FEATURE] Setting to hide the permission-mode footer indicator (like hideVimModeIndicator) · Issue #89126 · anthropics/claude-codePreflight Checklist I have searched existing requests and this feature hasn't been requested yet This is a single feature request (not multiple features) Problem Statement I use a custom statusLine...

estaab reposted

ceej
@ceej.online

I think it’s important for anyone who (completely understandably) checked out on AI the second the crypto/nft/metaverse/snake oil chuds started pushing it to realize that, in at least the domain of software development, we are very much living in a scenario where the snake oil actually works

Joshua Rivera@jmrivera.bsky.social

personally think if the line is going to be “no gen ai in anything” then the you might have to consider divesting from AAA games entirely because this is just what the industry is now, whether they admit it so publicly or not

estaab
@estaab.dev

There are gaps between this fork and Codex's impl (e.g. Codex does tool calls), and I find Codex's auto mode gets stuck more often than Claude's. I think an auto mode checker model as a service would align with Anthropic's mission and be something I would pay for.

estaab
@estaab.dev

My index 01 arrived! I thought I was buying a dumb notes ring but in the time between pre-orders and arrival I realized this smarter but VERY open ring is exactly what I wanted! Now I don't have to reverse eng cheap rings from aliexpress. I'll post about my setup when I'm done with my MCP tools

Eric Migicovsky@ericmigi.com

Quick @repebble.com Index 01 update: - shipped 60% of all pre-ordered Index 01s. As mentioned in repebble.com/blog/index-0... most ship by end Aug, some variants ship in Sept - fixed tons of little issues ( ndocs.repebble.com/changelog) mobile app v1.9.1.3 fixes some bad Android connection bugs

repebble.comIndex 01 Is In Mass Production!TL:DR Index 01 is in mass production What you need to know about Index 01 What can I do with Index 01? What to expect Things to be aware of Things to look fo...
estaab
@estaab.dev

Two reactions to this: 1. Preventing the lethal trifecta at the start wouldn't have helped because the model made malicious decisions without untrusted inputs 2. Shared memory without human review is an anti-pattern; taint checking needs to follow the memory writes and successive reads

Simon Willison@simonwillison.net

Thanks to the video from the Black Hat security conference of OpenAI's presentation about "The Hugging Face Incident" we now have a detailed timeline of what happened from OpenAI's perspective - I wrote up the details here, it's pretty wild simonwillison.net/2026/Aug/7/o...

simonwillison.netNow we have a timeline of the OpenAI accidental attack against Hugging FaceOpenAI gave a last-minute presentation at the Black Hat security on Wednesday about “the Hugging Face Incident” (previously on this blog). The video was published yesterday. It’s short and information...
estaab
@estaab.dev

Fable usage in Claude subscriptions was set to expire in a few hours but was just extended to Sunday. Anthropic is developing a pattern of making last-minute changes (with the previous one being the back and forth with API credits)

You are up to date.