Joseph Chalabi retweeted
New load balancer went live for AkashML. It keeps capacity saturated across all available GPUs hosting the model.
Yesterday: AkashML served 25.6B tokens as a dominant provider for gpt-oss-120b.
Traffic on OpenRouter is earned. The more performant endpoint serves the tokens.
Joseph Chalabi retweeted
Next Akash Fireside Chat: Thursday with @chalabi__, @ovrclk_ Sr. Core Engineer.
We will walk through the recently upgraded CLI as well as what the future of @akashnet provider audits will look like in our monthly sit down with various $AKT Core Team.
RSVP Now:🔗👇
Joseph Chalabi retweeted
new $akt ci is out, try it:
$ brew tap akash-network/tap
$ brew trust akash-network/tap
$ brew install akt
Joseph Chalabi retweeted
Introducing MIRA.
A playable, multiplayer world model.
A dream of Rocket League.
Trained on 10k hours of data collected with publicly available bots, MIRA learns the dynamics of a four-player game. The model runs in real time at 20 fps, based on the keys you and the other players press.
Built by General Intuition and @kyutai_labs, in collaboration with Epic Games. Not used to develop Rocket League.
▶️ Play the demo, read the technical report, and explore the open-source code at generalintuition.com/mira
At ICML? Find us at Booth 111 to try it yourself and dig into the results with the team.
Joseph Chalabi retweeted
Grassroots initiatives like this are how local AI will be mass adopted.
I am raising 50,000$ in donations this time all funds will go towards donations to 3 community centers in the United States.
- A school
- A church
- A recovery center
The funds will go towards purchasing 6 DGX sparks and travel to these locations.
A curriculum will be developed to help each of these locations make use of the 2 sparks for their digital needs.
Why I am doing. There are increasing risks related to losing access to AI. Most Americans dislike AI, and have never used anything like Claude Code/Codex
By educating and spreading the love we will get boots on the ground and figure out how to onboard the next 1 billion people to local AI.
Help me make this dream a reality, 0% of this money will go to me. All receipts will be shared and the journey will be documented.
donate.sybilsolutions.ai
Joseph Chalabi retweeted
The @ovrclk_ AI team has been hard at work making @akashnetAI the best inference offering in the market and its starting to pay off.
A week ago we set a new record with 10 billion tokens processed per day, today we're about to nearly double that to 17 billion tokens.
We've also been reaching new ATH every day since.
Got stopped at SFO for the AGI Pills heading home from @aiDotEngineer
“Sir are these prescription pills?”
Joseph Chalabi retweeted
a new ATH for @akashnetAI yesterday -- and most likely today too -- closing on 8 billion tokens a day
.@akashnetAI processed over 7 billion tokens yesterday, a new ATH and has been consistently processing over 5 billion tokens everyday on an average.
Another proof point that Inference on decentralized compute is works great at scale.
.invoice-generator.com is full of adds now :(
so I built invoicegenerator.xyz and it will be add-less for as long as I can afford the 2$ domain
Real-time 3D-to-ASCII renderer. @threejs → hidden canvas → luminance sampling → character mapping. Slide transitions melt the text forward or backward. Hover triggers a ripple from your cursor.
What an unbelievably cool company!
. @big_duca tasked me with reviewing and selecting the bounty winners for building the hyper liquid TWAP api for @AwakenTax.
And the $5k goes to @cryptodevbrian!
- very clean code
- great documentation
- made a website showing it off
We'll give $250 to the other contenders.
@0xBrandon
@joeyroth
@davidrsdlife
@chalabi__
We appreciate the effort you put in!
Joseph Chalabi retweeted
Most blockchain API providers use a weighted credit model: 1 credit ≠ 1 request. Credits depend on request type, duration, and response size. At Chandra Station, we chose simplicity: 1 request = 1 credit. Transparent and predictable.
Weighted credits are practical for large-scale providers handling billions of app requests. Our boutique approach focuses on fewer, direct customer relationships, ensuring clarity, fairness, and straightforward support without complex pricing.
Our model works because we deploy our own servers in data centers across the US, with cloud backups. When major infrastructure providers went down yesterday, we stayed live. This balance of owned infrastructure and cloud ensures transparency and reliability.
We're here to scale responsibly by adding more nodes, not hidden multipliers. If you value transparency, resilience, and 24/7 support for your blockchain app, visit us @ chandrastation.com/chandra-a…
If you're interested in our blockchain API infrastructure, read our detailed blog post: chandrastation.com/blog/tran…
Joseph Chalabi retweeted
If you were affected by the recent AWS outage, check out our new Blockchain API.
We offer both on-prem and cloud deployments for real redundancy and uptime.
Unlike others, our pricing is simple:
1 credit = 1 request.
No hidden multipliers, no surprises.
Reliable & Transparent.