Perplexity AI Open-Sources Unigram Tokenizer That Achieves 5x Lower p50 Latency Than Hugging Face tokenizers Crate

Perplexity AI’s research team reimplemented their Unigram tokenizer from scratch in Rust and open-sourced the code in pplx-garden, their inference technology repository. At production input lengths, the new encoder cuts p50 latency by roughly 5x versus the Hugging Face tokenizers crate, ~2x versus SentencePiece (C++), and ~1.5x versus IREE’s tokenizer (C), with zero steady-state heap…

Read More

Vertu wants CEOs to run companies from an AI foldable starting at $6,880

Luxury smartphone brand Vertu on Thursday unveiled a foldable phone powered by an AI agent that connects with enterprise software and coordinates workflows. The company is targeting executives who manage business operations and communications on the move. Called the Alphafold, the foldable smartphone starts at $6,880 for the calfskin version. Higher-end models feature bespoke finishes…

Read More

CrowdStrike and Google take down botnet used by hackers to target open source software developers

CrowdStrike, working with Google and Shadowserver, a nonprofit organization that scans and monitors the internet for cyberattacks, took down a botnet that cybercriminals used to push malware and steal passwords from open source software developers. The takedown operation had the goal of disrupting the activities of the cybercriminals behind the so-called Glassworm botnet, who have…

Read More
With a brand new 0M increase, Princeton’s Thea Vitality is now a top-funded fusion startup

With a brand new $100M increase, Princeton’s Thea Vitality is now a top-funded fusion startup

Thea Vitality has raised an oversubscribed $100 million Sequence B led by U.S. Modern Expertise Fund, the fusion startup informed TechCrunch. The sum locations the corporate among the many better-funded fusion startups, giving it an improved probability at attaining a industrial reactor. The brand new funding will assist Thea develop manufacturing for its uniquely designed…

Read More

Sakana AI Proposes DiffusionBlocks: a Block-wise Training Framework That Converts Residual Networks into Independently Trainable Denoising Modules

Researchers from Sakana AI and the University of Tokyo propose DiffusionBlocks. It trains transformer-based networks one block at a time. Training memory is reduced by a factor of B, where B is the number of blocks. Performance is maintained across diverse architectures. The Memory Problem in Neural Network Training End-to-end backpropagation requires storing intermediate activations…

Read More