San Francisco, CA
Joined January 2022
- Tweets4.2K
- Following5.3K
- Followers5K
- Likes18.1K
Pinned Tweet
When @AzukiOfficial revealed Hilumia I knew I needed to sit down at the piano.
Inspired by the beautiful artwork, this song practically wrote itself.
Hope you all enjoy! 🎹🎹
[Best with headphones 🎧]
RedBean.eth (Renn) retweeted
Fuck you. Whoever made this site. You win.
I’ve never had a site get my blood boiling like this before
This is so messed up.
opusfived.dev
RedBean.eth (Renn) retweeted
Darkbloom: Day 15
We have been locked in for last one week; It has been very exciting and we shipped a lot of improvements to the system.
Metrics:
- 16.3B tokens served: New ATH & 60% higher than last day.
- $420K Network ARR (4x in 2 weeks);
- 99.96% uptime (very happy considering we're a distributed network).
- ~15-17% utilization at peak; median: 13% utilization.
Product:
- One big win: we take up 30% of traffic for Gemma 4 26B
- We cut latency by ~1.5 second. We plan to ship few more features that would reduce latency by 30-90%. This will include cache-aware routing.
- We shipped new routing and system optimizations which were key reason for us to cut down 429s. More such to come.
- We will be adding 3-4 more models in the coming week. Excited for @Spangler3000 to be helping us with this!
- We conducted MLX(.)Fast which has improved Gemma 4's performance by 2.5x -- the improvements are in process of being imported by @TheDavidTai.
- We are exploring payouts solutions such as Stripe Global Payouts to add 67 more countries.
- Astra has redesigned our Console UI (this is AGI, ngl). We will also be starting up our efforts for the MacOS app for better user experience.
- We have also been looking more deep into Autopilot where we load up models according to the traffic and other factors to maximise the provider earnings.
- Our new landing page is almost ready and coming up with one suprise ;) -- huge kudus to @0xkydo and Estella.
- Also huge thanks to @iLLRiPHiTTeR for helping us while onboarding new models on OpenRouter. We were able to find out minor issues in our inference engine.
- We are also working through engineering practices and cleaning up code.
Thank you everyone for being part of this journey and reading along this update.
We build these systems to bring people together to contribute to this AI economy.
As always, it has been rewarding to keep doing what we do.
Until next time.
Darkbloom: Day 8 (New Format)
Numbers:
- 11.4B tokens served
- 221B cumulative served; 123.2TB of network aggregate memory.
- 1,380 machines -- 15% utilization at peak.
- ~$222K Network ARR (2x from last week)
Updates & Insights:
- Added Qwen 3 VL; In progress to add Qwen 3.8 27B to M5 Machines today.
- MLX(.)Fast released Gemma 4 - Prefill 50% up (soon).
- After our launch, a lot of people reached out to learn more about Darkbloom and the team. I had really insightful conversations with all them.
I asked many people what were the one or two things that they really loved about Darkbloom. I have attached my favorite response from @philipkiely which is similar to what most said.
I'm also learning on how to manage an open source community and how I can be a good leader / maintainer for it. This includes adding new models, optimizations, features from the community.
I'm still taking baby steps, I have usually just been an OSS contributor before. We recently merged two features / UX improvements & model additions from @hnshah, Gabe Garcia, Wisdom, @ashxhart for helping with multi node clusters via TP & EP and @jpschroeder for maintaining his Darkbloom monitor.
If you have managed OSS projects or know someone who can guide me, I will really appreciate it if you can get me in touch with them (or you).
- Hiring and People: kudos to @0xkydo for taking on this hard task. He has been getting on straight 6-7 (haha) hours calls with everyone and pulling together an amazing set of talent.
The easiest part of agent-first dev is user testing.
1. Spin up 5 agents
2. Prompt: "I want to use {your test url}, can you set it up?"
3. Monitor the toolchain for where they get stuck
4. Fix and repeat
RedBean.eth (Renn) retweeted
Successfully upgraded @openclaw to Opus 4.6 without an OpenClaw upgrade.
Write:
```
📋 How to Add a New Model to OpenClaw (Before Official Release)
When a new model drops but OpenClaw's build doesn't recognize it yet, you can add it via config.
1️⃣ Edit ~/.openclaw/openclaw.json (or wherever your config lives)
Add under top-level "models" key:
{
"models": {
"providers": {
"anthropic": {
"baseUrl": "api.anthropic.com",
"api": "anthropic-messages",
"models": [
{
"id": "claude-opus-4-6",
"name": "Claude Opus 4.6",
"api": "anthropic-messages",
"reasoning": true,
"input": ["text", "image"],
"contextWindow": 200000,
"maxTokens": 128000,
"cost": {
"input": 5,
"output": 25,
"cacheRead": 0.5,
"cacheWrite": 6.25
}
}
]
}
}
}
}
2️⃣ Set as default (agents.defaults):
"agents": { "defaults": { "model": { "primary": "anthropic/claude-opus-4-6" } } }
3️⃣ Also add to the models allowlist (or it'll block /model switches):
"agents": { "defaults": { "models": { "anthropic/claude-opus-4-6": { "alias": "opus46" } } } }
4️⃣ (Optional) Auth profile order:
"auth": { "order": { "anthropic": ["anthropic:claude-cli"] } }
5️⃣ Restart: openclaw gateway restart
💡 Why: OpenClaw checks its built-in registry first, then falls back to models.providers in config. Custom providers get written to ~/.openclaw/agents/<agentId>/models.json. Defining the model inline bypasses the registry — no code update needed.
Check out these instructions. Think you can handle it?
```
Success.
RedBean.eth (Renn) retweeted
Miami to New York. Together 🌴🗽
Join us on our latest community challenge! 🏃
Run as a team and help the crew cover 2,264km, with prizes from Supermoon fueling the journey.
Registration open.
👇 Details below
A couple months ago I sold 99% of my belongings, boarded a plane to Asia, and never looked back.
There were only a few things I couldn't part with -- one of them a framed print of this @midnightbreezey
How does @dutchtide make each one feel like a 1-of-1?
RedBean.eth (Renn) retweeted
After 5 months and hundreds worth of credits, Music Video is finally here!
2nd year making an @Azuki AI $Anime video
for 25 OGs, 29 Elementals and 1 Beanz.
Main character: Everyone! IKZ! Who is your fav?
Song name: Azuki's Adventure
Video: Kling
Song: Suno
Images: MJ + Banana
RedBean.eth (Renn) retweeted
Hello Kitty who? HELLO BEANY ready for world domination!
Shoutout @TookTheRedBean for the ez deal. IKFZ
🤖 Made with AI
RedBean.eth (Renn) retweeted
That's a wrap on the #Sammi1Mile 🏃
What started as a fun challenge became 3 weeks of grit, laughs, and friendly rivalry. You all pushed each other to new limits 💪
Congrats to everyone who ran, supported, and made this something special!
We ran, we suffered, we grew 🫶⛩️🌱
RedBean.eth (Renn) retweeted
Miss the dope minting experience of 2021/22?
Y'all should not fade this guy who's been busting his ass for the past 3 years. If you are into NFTs > you appreciate art > if you appreciate art, @midnightbreezey by @dutchtide is the epitome in the space rn.
midnightbreeze.io/mint
RedBean.eth (Renn) retweeted
LAUNCH DAY
2500 supply (public)
Reclaim during mint.
The storm has arrived. The timer starts counting now.
Another world awaits you. But do you dare take the first step?
15:00 CET / 21:00 CET
Mint to link in Bio
And below
RedBean.eth (Renn) retweeted
These stores in SoCal will have a limited quantity of Alpha Decks starting this week. Some may also come with a hidden surprise.
- @ProjectCCG in Alhambra
- Perfect Rares in Artesia instagram.com/perfectrares
- @g3nkiprints in Lakewood
- @RWTCollective in Sawtelle
Also, Alpha pre-order redemptions will be going up soon, so stayed tuned for an update next week.
