DecideAI has rolled out three key enhancements in the second quarter of 2025, reinforcing its ongoing commitment to building decentralised AI infrastructure that values performance, privacy and openness.
The first improvement boosts the Decide ID Model—a facial recognition system that operates across a blockchain network. In head-to-head tests with InsightFace, a widely recognised open‑source benchmark, Decide ID demonstrated up to six times better performance on essential accuracy measures. This places it well ahead in real‑world applications like stopping fraudulent token distribution, combating spam on decentralised social platforms and reinforcing reputational systems in Web3. The system works without any KYC procedures or reliance on central servers—it simply combines artificial intelligence with blockchain verification.
Moving beyond in‑house deployment, DecideAI has integrated Decide ID into the OpenID Connect (OIDC) framework. This allows developers to incorporate Decide ID as an identity provider in applications that already use OAuth2/OIDC. The result: systems that require proof of genuine human users without the need to hold personal data. Integration is designed to be quick—taking minutes rather than weeks. Developers can explore the integration via DecideAI’s online guide.
On the infrastructure side, the company introduced ic-file-uploader, a Rust crate crafted to enhance file transfer speeds to the Internet Computer (IC). With client‑side parallelisation, it delivers around 0.31 MB/s over the IC and up to 3.5 MB/s on local networks. This tool lays the groundwork not only for straightforward file uploads but also for richer services such as video streaming, liveness checks and on‑chain media storage. Future updates are expected as the team continues to expand distributed architectures.
DecideAI has also made strides in computational back‑end performance. By integrating AMD’s AOCL-BLIS into WasmTime on the Internet Computer, the team achieved around fourfold faster matrix operations and up to eightfold improvements in compute‑intensive machine learning workloads. This development should mean reduced costs and lower latency for AI inference systems. In parallel, the company is experimenting with compressing GPT‑2 style chain-of‑thought sequences into continuous space embeddings—an approach that could produce outputs of GPT‑3 calibre using more compact models.
All three releases underscore DecideAI’s opposition to the increasingly centralised and opaque nature of many AI systems. Its strategy remains rooted in transparency, verifiability on‑chain, default open‑source licensing and a privacy‑first design philosophy. These Q2 updates are working examples that decentralised, high‑performance AI is achievable without dependence on large-scale tech platforms.
Dear Reader,
Ledger Life is an independent platform dedicated to covering the Internet Computer (ICP) ecosystem and beyond. We focus on real stories, builder updates, project launches, and the quiet innovations that often get missed.
We’re not backed by sponsors. We rely on readers like you.
If you find value in what we publish—whether it’s deep dives into dApps, explainers on decentralised tech, or just keeping track of what’s moving in Web3—please consider making a donation. It helps us cover costs, stay consistent, and remain truly independent.
Your support goes a long way.
🧠 ICP Principal: ins6i-d53ug-zxmgh-qvum3-r3pvl-ufcvu-bdyon-ovzdy-d26k3-lgq2v-3qe
🧾 ICP Address: f8deb966878f8b83204b251d5d799e0345ea72b8e62e8cf9da8d8830e1b3b05f
🪙 BTC Wallet: bc1pp5kuez9r2atdmrp4jmu6fxersny4uhnaxyrxau4dg7365je8sy2q9zff6p
Every contribution helps keep the lights on, the stories flowing, and the crypto clutter out.
Thank you for reading, sharing, and being part of this experiment in decentralised media.
—Team Ledger Life
