Topic: generative ai models
-
Google's TurboQuant AI Cuts LLM Memory Use by 6x
The high memory demands of large language models (LLMs) are a key factor in current high memory prices, driven by the substantial memory consumption of their key-value caches. Google's new TurboQuant compression technique dramatically shrinks an LLM's memory footprint and accelerates performance ...
Read More » -
ByteDance unveils first AI-designed therapy, enters drug discovery race
ByteDance's drug discovery unit, Anew Labs, unveiled its first AI-designed small molecule at a Boston conference, targeting the "undruggable" IL-17 protein-protein interaction for autoimmune diseases like psoriasis. Anew Labs aims to replace expensive injectable antibody therapies with oral pills...
Read More » -
Amazon AI trained on Twitch streams; users can now opt out
Twitch has launched a new opt-out setting allowing streamers to block Amazon from using their channel content (streams, VODs, clips, chats, and images) for generative AI training, with the option available at twitch.tv/settings/security. All Twitch users are automatically enrolled in AI training ...
Read More » -
Sony sues Udio’s AI over 30,000 songs in copyright case
Sony Music Entertainment has filed a new lawsuit against AI platform Udio, alleging it unlawfully copied over 30,000 copyrighted songs into its generative AI models, including hits from Elvis Presley to Beyoncé. The lawsuit follows a 2024 case where Sony, alongside other major labels, initially s...
Read More » -
Adobe puts 70+ tools into ChatGPT
Adobe launched a unified plugin for ChatGPT, integrating 70+ core tools like Photoshop, Premiere, Lightroom, and Firefly AI, replacing scattered individual integrations. The plugin enables seamless AI-driven workflows within ChatGPT, allowing users to edit, generate assets, and trigger app-specif...
Read More » -
Midjourney demands Hollywood disclose AI usage details
Midjourney is pushing for Disney, Universal, and Warner Bros. to disclose their internal AI practices during the discovery phase of a copyright lawsuit, challenging a court restriction that limits the studios to only providing information about "consumer-facing" AI-generated content. The startup ...
Read More » -
Meta CTO Calls Company’s AI Reorg ‘Atrocious’
Meta CTO Andrew Bosworth acknowledged the company's rollout of its new AI division was "atrocious," citing damaged employee trust and a chaotic restructuring that left teams unstable with rapid strategy shifts. Bosworth announced reforms including capping managers at 20 direct reports, limiting e...
Read More » -
SoundCloud Reverses Controversial AI Terms Update
SoundCloud reversed controversial policy changes that allowed potential AI training on user uploads after facing significant backlash from creators concerned about unauthorized use of their content. The company clarified its terms to explicitly prohibit using uploads for generative AI systems tha...
Read More »