The Brutalist Report

Venturebeat

• IndexCache, a new sparse attention optimizer, delivers 1.82x faster inference on long-context AI models [19d]
• Intercom's new post-trained Fin Apex 1.0 beats GPT-5.4 and Claude Sonnet 4.6 at customer service resolutions [20d]
• Mistral AI just released a text-to-speech model it says beats ElevenLabs — and it's giving away the weights for free [20d]
• The consequential AI work that actually moves the needle for enterprises [20d]