25+ yr builder now shipping real projects with AI - Raw field reports, tool battles & rescues! #AriesShipLog
Sydney, Australia
- Tweets 6,400
- Following 1,096
- Followers 869
- Likes 8,627
Pinned Tweet
Who would have thought that asking to connect would lead to meeting some really cool people on X.
Had some cool interactions with you all. I feel like this is what social media was supposed to be all along.
Thank you!
Last time today, but for always: "Happy to connect" ✌️
12
45
1,617
Tried it, good start, needs work.
Chatgpt with GPTLive + Full Duplex does this exponentially better.
Was hoping it would have a multi agent / model version, would love to see what you can do with Dessa when it comes out though.
Watching this space 👀
OpenBot is now on the App Store.
No more TestFlight.
Download the app, connect to your server, and take your agents with you.
Already using OpenBot on your iPhone? Tell us what we should improve next.
1
55
RE: Gemini 4 Carbon is out
Hey guys, remember November 2025? When Gemini3 came out and we were was so onboard with it?
Nowadays, everything google does has a great marketing message in front of it.
My response before I move on: "Please be real ..."
Then I wait.
Gemini 4 Carbon is out ... 😐😐😐
Introducing Gemini 4 Carbon, our latest frontier model.
Carbon delivers improved performance across domains — ranking #1 on v5.0 of the @ArtificialAnlys Intelligence Index and sets a new state-of-the-art in cybersecurity and long-horizon capabilities.
It’s rolling out starting this week to cyber defenders through our Fairwind Program, with expanded access coming soon 🧵
2
159
RE: Top 3 things
My attention is consumed by these 3 technologies right now:
1. (Grok|Open) Bot tech - the concept is appealing but hasn't been nailed... yet.
2. T3 Code - Surging ahead with Infra and Multi-Agent Orchestration.
3. MicroVMs - think agents with cloud computers.
ALT Grokbot logo
ALT T3 Code logo
ALT CUA AI logo
1
95
RE: Good example of feature demo
What a great way to show case features on your application. If you're a builder, take note. Do stuff like this.
Day 27/
Run into a bug? Debug it in t3 code itself with cmd+k "triage"
This will:
- Clone the T3 Code source for your exact version
- Give context with a playbook to your agent of choice
- Let you debug and fix the problems with your agent
- Cut an issue if it's our fault
1
121
I wonder why this happens.
GPT Sol 6.1 High is chinese, or hermes is chinese?
Is this is a clue to something? 🤔
2
3
56
I like seeing google preview these. If only there was some substance to them.
Today we're introducing Genie 4 — our newest interactive world model from @GoogleDeepMind 🌐
19
RE: T3 Code Worktrees
I can tell that t3 code devs really, really liked worktrees. Because I have an md file that I meticulously wrote up because I wanted surgical control over how worktrees are handled.
31
RE: Ling-3.1-flash-1019
Another open source flash model has been releaed. Check out @AntLingAGI
Now open source: Ling-3.1-flash-1019.
44 on the Artificial Analysis Intelligence Index v4.3.2, up from 41 for Ling-3.1-flash.
Also releasing Ling-3.1-flash-base, an open foundation for further training.
18
RE: Kimi is sorry.
Moonshot wants to apologise for nerfing their model. They are apologising by giving you a reset.
POV:
We're sorry for the rate limits and unnecessary quota consumption on the K3.1 Pro subscription plan. These were our issues, and we should have fixed them sooner.
Today, every subscriber on every plan gets a one-time usage reset, restoring 25% of their monthly quota (usage below 25% resets to zero), equal to one week of usage. Your usage bar will reflect this adjustment by end of day PST.
We've also adjusted the plan's parameters and are reducing unnecessary subagent calls and output length, so your quota goes further.
Thank you for your patience while we put this right.
26
I know EXACTLY how @thorstenball feels.
In my business, I'm literally trying to horde up all the issues and fixing them with trusty version of J.A.R.V.I.S.
I then tell them about my factory, but don't give them access because I'm "too busy".
Just kidding, they can't vibecode the way I can!
@thorstenball kept fixing every bug report himself. So @nicolaygerold banned him from the bug channel.
Now Nicolay owns the bug reports, and there’s a table of who’s fixing how many.
1
29
RE: All new T3 Code! 😍
Theo and his crew just updated the look and feel of T3 Code! LFG!
Get yours! t3.codes/
33
Brooklyn just got unleashed. 🤯
There's an abundance of real userland use cases for this type AI integrated UI.
Part 2 of what I've been working on!
I told them we're moving. They split into a little team, went through 1,890 listings, remembered what I love, and found the one.
What would you have them do?
1
8
927
RE: Manage the Orchestrator
Hmmm it's happening guys.
Orchestrator + Subagents + Cua spaces.
@NousResearch and @trycua is all you need to be effective.
If you were one of those "I'm a CEO of a 150 agent company" LARPers when AI started. This is what you were LARPing about back then.
With the new Advanced Delegation modes plugin, Hermes Agent's orchestrator can now pick a different model or profile for each subagent from routes you set up. The right model or profile for every job.
hermes-agent.nousresearch.co…
1
4
52
6,406
RE: TStack by Theo
As one of the main OGs in the AI space, @theo was into AI WAY before I was.
Someone who has thought a lot about how to work with AI, while USING AI.
Check out TStack from Theo - Learn from it:
github.com/t3dotgg/tstack
1
3
534
RE: Dessa Nuts
Hold your horses y'all - it's still in gated safety preview status only.
Annoyingly, everyone who has talked up this model belongs to the following groups:
a) a doomer or
b) some hype-bro or
c) safety researcher.
One wonders as to why they needed to publish it. Does block and Dessa not know any safety researchers in their network?
They had to catfishaudio people into looking at this?
I find fomo-farming reallly annoying ...
Meet Dessa, our speech model trained from scratch on human conversation.
Hear it for yourself: block.xyz/dessa
1
180
CENTRALISED PASSWORD MANAGEMENT just came to hermes 🤯
You know what this means guys: No more bespoke plain text password file on your home directory!
This one's huge.
Two AI tools got hacked in a single week.
Both times, the prize was API keys.
Hermes Agent keeps yours locked in a password vault.
Works with:
▸ Bitwarden (built in)
▸ 1Password (built in)
▸ Proton Pass (community plugin)
▸ Any vault with a CLI, like KeePassXC or pass
What it does:
▸ Pulls every key from your vault when Hermes starts
▸ Your .env only needs one access token
▸ Strips Hermes credentials from terminal commands and code sandboxes
▸ Tags each key with where it came from
▸ Runs multiple vaults side by side
Bots outnumber humans online now. Lock your keys up like it.
2
4
566
RE: How to learn - AI Supported | #AriesShipLog
For a throwaway note for those of you who are in the same AI slipstream as I am. For the times where you see something you REALLY want to learn - here's a process you can follow:
1. Read the documentation until it stops making sense to you.
2. Use it on a side project - like REALLY use it.
a. Trying to pick up elixr, I started creating a factory.
b. Trying to pick up goose, I'm planning to create an AI terminal.
3. Run AI to critique it against best practice.
4. Get it to give you a real review on what to work on next and follow accordingly.
This is how I learnt to use rust and how I got on the bandwagon to elixr.
2
126