Browse Tag

AI backlash

AI Breakthroughs, Billion-Dollar Bets & Backlash – The Global AI News Roundup (Aug 11–12, 2025)

AI Breakthroughs, Billion-Dollar Bets & Backlash – The Global AI News Roundup (Aug 11–12, 2025)

OpenAI released GPT-5 on August 7, 2025, but by August 11 users reported sluggish responses and a less personal feel, prompting Sam Altman to admit the rollout was bumpy and to keep GPT-4 available for paying users while fixes are rolled out. OpenAI released GPT-OSS models with 20 billion and 120 billion parameters, the first freely downloadable large language models since 2019, available via AWS and excelling at coding, science, and math. Caltech researchers used reinforcement learning with a two-agent system called “player” and “observer” to crack the 60-year Andrews–Curtis Conjecture, showing AI can find solutions to long-standing group-theory problems.
AI’s Frenzy: Billion-Dollar Bets, Breakthroughs & Backlash – Global AI News Roundup (Aug 5–6, 2025)

AI’s Frenzy: Billion-Dollar Bets, Breakthroughs & Backlash – Global AI News Roundup (Aug 5–6, 2025)

MIT and Duke University researchers used AI to design tougher plastics, identifying iron-based ferrocene additives and mechanophores that make polymer networks absorb stress, with results published in ACS Central Science. Profluent Bio used generative AI to create OpenCRISPR-1, the first AI-designed CRISPR enzyme, which edited human genome cells and was trained on 500 million sequences, with the work published in Nature and open-sourced. OpenAI unveiled two open-weight LLMs with 120 billion and 20 billion parameters that can run on a single GPU or a laptop, with open weights for fine-tuning and availability on AWS Bedrock. Thomson Reuters launched CoCounsel Legal,
Go toTop