Gopinath Varadharajan retweeted
Iroh v1.4.0 got released!
> iroh v1.4.0 fixes critical rustls security issues, adds Endpoint-Id header, improves LAN networking, and boosts relay performance with several bug fixes.
ALT Image: Iroh v1.4.0 got released! > iroh v1.4.0 fixes critical rustls security issues, adds Endpoint-Id header, improves LAN networking, and boosts relay performance with several bug fixes.
Gopinath Varadharajan retweeted
SSH Tunnels: An age-old trick that's still widely used today
- Expose a local service to the Internet
- Map a remote service to a local port
- Query an AWS RDS database with a local GUI client
- Access a server in your private VPC from a dev machine
...and many more cases
Learn how to use SSH tunnels at labs.iximiuz.com/tutorials/s…. Hands-on exercises included.
Gopinath Varadharajan retweeted
Decider V1.2 answered 99.7% of claims alone and sent the rest to Claude Haiku 5.5, for $0.11 per 1,000 claims against $1.19 for Haiku 5.5 alone, at 99.9% accuracy.
For the full results, visit vals.ai/benchmarks/claimbenc…
Gopinath Varadharajan retweeted
Perplexity’s Decider V1.2 is the best decision model on the benchmark and gets within 0.3 points of a perfect score at $0.10 per 1,000 claims. Mercury Decide is second at 99.1%, then Decider V1.1 27B at 98.8% and Liquid AI’s d1 at 98.7%.
The five language models scored between 99.2% and 100%, at $1.37 to $91.83 per 1,000 claims.
Gopinath Varadharajan retweeted
Today we’re releasing ClaimBench, our new benchmark targeted towards decision models. It tests whether a model can tell if a financial document supports, contradicts, or doesn’t address a claim.
We evaluated 15 models on 1,960 claims paired with 10-K excerpts, with Decider V1.2 leading the pack of decision models.
Gopinath Varadharajan retweeted
Perplexity Decider pareto-optimal on Financial ClaimBench of Vals AI
Today we’re releasing ClaimBench, our new benchmark targeted towards decision models. It tests whether a model can tell if a financial document supports, contradicts, or doesn’t address a claim.
We evaluated 15 models on 1,960 claims paired with 10-K excerpts, with Decider V1.2 leading the pack of decision models.
Gopinath Varadharajan retweeted
Big Boy rendered by Computer
Computer built Big Boy No. 4014, the world's largest operating steam locomotive, as a live isometric SVG with its own renderer and a 7-step operating sequence from cold boiler to 60 mph.
The live model has 1,324 parts and 15,312 convex faces, built from @UnionPacific's published figures.
Try it: big-boy-4014.pplx.app
Gopinath Varadharajan retweeted
Borrow Checking Zero-Cost Cyclic Mutability
> Ante introduces a static, zero-cost borrow checker using places and instability markers, enabling safer shared mutability, cyclic references, and fewer restrictions.
ALT Image: Borrow Checking Zero-Cost Cyclic Mutability > Ante introduces a static, zero-cost borrow checker using places and instability markers, enabling safer shared mutability, cyclic references, and fewer restrictions.
Gopinath Varadharajan retweeted
A Mac with at least 36GB of RAM can now run Perplexity AI workflows locally. engadget.com/2291754/perplex…
Gopinath Varadharajan retweeted
Indra OCR 1.1 is out now!
- 1M param main OCR model, 44 pages/s on CPU
- Within 4 points of 3.4B param DeepSeek OCR 2 on OmniParseBench
- 15x faster than Tesseract, and more accurate
- 80+ languages, plus handwriting
Gopinath Varadharajan retweeted
24x faster Rust compilation powered by rust-analyzer.
nix run github:indexable-inc/ra-jit
Gopinath Varadharajan retweeted
LayerFS is a Rust-powered filesystem for AI agents, enabling zero-copy workspace branching, deduplicated snapshots, parallel rollouts, and efficient state recovery.
ALT Image: LayerFS is a Rust-powered filesystem for AI agents, enabling zero-copy workspace branching, deduplicated snapshots, parallel rollouts, and efficient state recovery.
Gopinath Varadharajan retweeted
Crusoe is extremely well positioned to become a full-stack cloud: power, compute, and services running on top. Perplexity’s open source sandbox platform SPACE will become part of Crusoe’s Managed AI products. Look forward to more collaborations with them!
Agents need secure sandboxes as much as they need GPUs.
We're partnering with @perplexity_ai to bring SPACE to our Managed AI portfolio on Crusoe Cloud.
Gopinath Varadharajan retweeted
Using state machines for agent logic is extremely slept-on
Don't think of them as static workflows. They're like LEGO bricks that the agents can rearrange + improve on autonomously, in a human-readable way
Otherwise it's like shaping markdown Play-Doh and hoping it holds
Gopinath Varadharajan retweeted
a brand new model paradigm! introducing 𝚌𝚕𝚊𝚒𝚛𝚒𝚌: a system one model for extracting, computing, estimating, and predicting values from unstructured inputs.
blog.cloudflare.com/clairic-…
