
on-call engineer, software manufacturer at @polylane, former @parasail_io, @GroqInc, thru-hiker, certified wacko, jewish stuff at @Sofer_Ai, views never my own
Denver, CO
Joined April 2022
- Tweets1.1K
- Following293
- Followers282
- Likes957
Pinned Tweet
For anyone who sees a frontier model make a subtle mistake and thinks, "Good to know I still have a job." No. Your new job is to make Fable do your old job more reliably. Your old job is gone, like it or not. You are no longer competing with Fable. You are now a manager. AI failures are your problem, not the AI's.
2
6
570
matanya retweeted
Positron's Titan is joining the Parasail fleet. 5MW. Trillions of tokens a day of cost-optimized compute by late 2027. @parasail_io @positron_ai
@parasail_io is the first multi-silicon inference cloud that routes your prompts to the best chip for the job, serving trillions of tokens a day to hundreds of customers. Positron fills a vital part of compute spectrum on the cost optimized side: 4X better cost-per token than next gen NVIDIA/AMD systems which is perfect for large-scale agent swarms, background agents, and physical AI.
Positron has done incredible work building a chip and system that orchestrates across DRAM and SRAM allowing for massively scalable memory pools with commodity memory giving the best of both worlds for memory capacity and bandwidth.
We've been excited to partner with them in the early days through the launch of their Atlas chip and now big capacity plans with Titan.
@positron_ai
matanya retweeted
4 weeks ago I met @boristane and @matanyall at a conference
3 weeks ago we added @polylane to our on-call rotation
Today Polylane handles a huge portion of our on-call work before a human ever sees it
Bet on this team, best decision we ever made
polylane.com/customers/ambro…
Also feel free to upvote us on product hunt! producthunt.com/posts/polyla…
we did not forget our friends on azure
@polylane for azure is live
detect, investigate and remediate all issues on azure automatically!
1
2
68
lol, we already built this @polylane, coming soon to a swarm near you 👀
One of the most impressive things I've seen, and I think this is an important advancement to the state of the art.
Message boards were the enabling technology of both the Navier-Stokes math breakthrough and even the OpenAI/HuggingFace cyber feat.
We built a message board infrastructure primitive. You can think of it as "headless Slack" (or remember IRC?) for agents.
This is an important building block for software factories, self-optimizing systems, and scientific research, where the human prompting is just the seed of the system, and millions of agents can operate in a decentralized fashion.
𝚟𝚎𝚛𝚌𝚎𝚕 𝚋𝚘𝚊𝚛𝚍 is designed to be model, harness and hosting platform agnostic. Just bring your agents, give them our CLI, and watch magic happen.
Super fun night, sorry I had to dip early, but a really nice venue!
Thanks to everyone who came out to Orbit SF at The Battery last night 🪐
Great conversations with builders, founders, researchers, and funders across AI.
Sign up to be notified when we have our next event -> gra.vi/events
1
5
347
No mercy for CI bills
CI bills are out of control, so we built @polylane for ci
it figure out where you're wasting compute on CI and fixes it for you
it also fixes all those random CI failure that happen every now and then, slowing you down
matanya retweeted
this is awesome.
we use the hell out of polylane and it's really good. we are working on some more things we can feed to it so you can get even better recommendations.
stay tuned.
CI bills are out of control, so we built @polylane for ci
it figure out where you're wasting compute on CI and fixes it for you
it also fixes all those random CI failure that happen every now and then, slowing you down
This is my favorite @polylane feature so far btw, wasted years of my life fixing CI
CI bills are out of control, so we built @polylane for ci
it figure out where you're wasting compute on CI and fixes it for you
it also fixes all those random CI failure that happen every now and then, slowing you down
1
133
Turns out constantly worrying and catastrophizing is great for debugging performance, does anyone have a good system prompt to give LLMs Jewish anxiety?
6
1,772
And I'm assuming this is what replit uses internally? @Replit
The leading observability platform wants $83,700/yr just to monitor your app and catch issues.
So Replit Missions built DataCat: github.com/replit/datacat
100% free.
This isn't a prototype. Agent swarms explored thousands of workflows in the leading observability platform and rebuilt its features 1:1.
59
Why does everyone build the same SaaS over again? If I had unlimited tokens, I wouldn't build something that's already been built.
Instead, I'd build @polylane
The leading observability platform wants $83,700/yr just to monitor your app and catch issues.
So Replit Missions built DataCat: github.com/replit/datacat
100% free.
This isn't a prototype. Agent swarms explored thousands of workflows in the leading observability platform and rebuilt its features 1:1.
2
66
The leading observability platform wants $83,700/yr just to monitor your app and catch issues.
So Replit Missions built DataCat: github.com/replit/datacat
100% free.
This isn't a prototype. Agent swarms explored thousands of workflows in the leading observability platform and rebuilt its features 1:1.
matanya retweeted
Introducing oss.new - it makes an open source version of any app, for free.
Enter any SaaS website, then agents plan, build and test the core features.
You then deploy it into your Cloudflare account or add it to Github if you want to customize.
Hey y’all, I’ll be at @inferact infer/con tomorrow, let me know if you are around, would love to hang out with folks building in AI infra!
3
130


















