DST Global

OpenAI launches GPT‑Live voice models

OpenAI introduced GPT‑Live, a new full‑duplex family of voice models powering ChatGPT Voice that can listen and speak simultaneously and delegate complex tasks to other models in the background. GPT‑Live (including GPT‑Live‑1 and GPT‑Live‑1 mini) adds live translation, visual answer cards and improved noise suppression; the mini version will be available to free ChatGPT users. OpenAI says GPT‑Live can hand off deeper analyses to larger models such as GPT‑5.5 (and an upcoming GPT‑5.6) and then return results into the ongoing conversation. The models are rolling out on ChatGPT for web, iOS and Android, with API access planned and developer registration open. The article references hands‑on testing by AI expert Jens Polomski and notes more than 150 million people already use ChatGPT’s voice/dictation features.

Rising AI Costs Renew Interest in Open-Source Models

Rising inference costs and geopolitical frictions have revived industry interest in open-source AI models as a lower-cost, more controllable alternative to frontier proprietary systems. The article reports that Chinese labs have pulled ahead since DeepSeek’s 2025 breakout, with open models such as Qwen and Kimi becoming common foundations for startups. Anthropic’s Mythos family faces export-control scrutiny, adding urgency to alternatives. Poolside — a $12 billion model maker backed by Nvidia, Bain Capital and DST — has publicly refocused on open source and released Laguna XS.2 in April, a model tailored for agentic coding, according to co‑founder and co‑CEO Eiso Kant. Investors including Michael Mignano (USV) and Bill Gurley say startups and companies are rethinking costly “tokenmaxxing” and diversifying model usage to control spend.

Nvidia partners with Ineffable to build reinforcement-learning AI

Nvidia announced an engineering partnership with London-based startup Ineffable Intelligence to co-design infrastructure for large-scale reinforcement learning. Ineffable, founded in late 2025 by UCL professor and former DeepMind reinforcement-learning lead David Silver, is pursuing “superlearners” — systems that learn continuously from experience rather than relying primarily on human data. The companies said engineers from both firms will build a pipeline to train reinforcement-learning systems at scale using Nvidia’s Grace Blackwell chips and the Vera Rubin platform. Ineffable raised a record $1.1 billion seed round in April co-led by Sequoia and Lightspeed, with participation from Nvidia, DST Global, Index, Google and the U.K.’s Sovereign AI Fund. Nvidia CEO Jensen Huang described superlearners as the “next frontier of AI.”