👾 https://nitter.gq/t.co/5OFfOQQtaV 🥚 https://nitter.gq/t.co/rxAeU2D9pO Previously @vercel AI DX The OG egghead
home
Joined July 2008
- Tweets27K
- Following2.3K
- Followers30.3K
- Likes17K
Pinned Tweet
Grab an early-bird 🐥 ticket to my "Agentic Software Factory" workshop on July 17th.
This workshop supports users of Claude Code, Codex, and every other other harness.
egghead.io/workshop/software…
Remember how Cursor re-invented "Tab, Tab, Tab" for generating code?
How come we haven't seen that predictive experience for building apps?
I'm oddly excited for Carbon.
I find chatting with Argon really fast and enjoyable.
I have high hopes that Carbon delegating to Astra might exceed Opus.
Anyone have a guide for codex?
Do you just ask for the Flare model?
Is there any way to tell which model it used?
I've been doing a massive amount of research and analysis around how I teach using the 100s of lessons and workshop I've recorded.
Seeking to understand myself and how I approach teaching all the way down to sentence structure.
Absolutely fascinating to see how my mind works.
We need to mentally prepare ourselves for more "predictive" experiences.
Everyone is familiar with autocomplete and ghost text.
But how does that transfer to other forms of I/O? What if your AI glasses predict what's going happen next?
Black Mirror, where you at?
I wired up a button on my watch:
- Instant record mic while holding down
- Auto-upload recording to Google drive
- Ping Grok Bot webhook
- Grok Bot transcribes and delegates whatever I said to the appropriate bot
So good.
I _love_ that grok bot can create webhooks for the agents.
Wiring the plumbing was always the most annoying part of automations.
Absolutely in love with @herdrdev 's new desktop app.
Already modding the living crap out of it.
Agent-first EVERYTHING
I don't use Dots, doubt I ever will.
I wish we could just remove it from the ChatGPT sub and save 25% or something.
Okay, Anthropic. You can release Fable 5.5 now.
Introducing Gemini 4 Carbon, it's a big jump over Argon which we released just 2 weeks ago.
Gemini 4 Carbon is SOTA on cyber, long horizon tasks, ranks #1 on Artificial Analysis V5, significantly improves on usability in domains like coding, and is more aligned.
Just did a 5 minute brain dump talking to an agent on my phone while driving.
It wasn't listening. 💩
I'm honestly really upset about this.
If you haven't explored the depths of what ChatGPT Pro is capable of, then you just don't understand.
ChatGPT Pro users, don't forget that your usage limits are getting cut in half in just 3 days.
Terrible news to be reminded of, of course. And the only way OpenAI can make this whole situation any less awful is by releasing a model that's as good as or better than Opus 5.5, while being way more efficient than the current GPT-6 Astra.
Searching for a memory leak and Opus says:
"Sorry, it retried every session 20.6 million times"
I know the models are amazing, but sometimes they're just not :)
