🤖AI Newsletter
September 24, 2026 · 04:45 Uhr
1ChatGPT Voice now accesses email, calendar, and Slack and completes tasks by voice
THE DECODER OpenAI expands ChatGPT Voice with agent-like capabilities that directly access business tools (email, calendar, Slack) and execute tasks fully automated by voice. This positions ChatGPT as a productive enterprise solution and significantly intensifies competition with Microsoft Copilot and Google Workspace. The move targets monetization through B2B integration and dependency building among enterprises.
2Gemini 3.8 Flash TTS: Google wants to transform audiobooks, games, and podcasts with AI voices
THE DECODER Google expands its AI portfolio with highly flexible text-to-speech technology that, with 100+ languages, voice replication, and granular control, could democratize the production of audiobooks, games, and podcasts. This addresses a rapidly growing market and could pressure established providers (ElevenLabs, Microsoft) while creating new business opportunities in the creator ecosystem.
3YouTube upgrades Creator Studio with AI: script tips, smart thumbnails, and Gemini editing
THE DECODER YouTube integrates AI tools to boost content creator productivity (script analysis, intelligent thumbnails, automated editing with Gemini, live translation). This lowers production barriers, increases upload frequency, and ties creators more closely to the platform – a direct competitive advantage against TikTok and competes with AI editing tools like Descript or CapCut. The move targets revenue growth through more creator activity and higher ad inventory.
4Anthropic employee explains why Claude forgot how to write for humans
THE DECODER Anthropic acknowledges that Claude models lost user-friendliness through optimization for technical tasks (math, code) and have become difficult for humans to read. The new Opus 5.5 model aims to correct this, but also signals that Anthropic must adjust to remain competitive with ChatGPT and other AI assistants. This could impact user adoption and thus business growth.
5Behind the scenes at Basecamp Research: How an AI start-up turns evolution into training data
THE DECODER Basecamp Research uses biological data from extreme ecosystems as training data for AI models to develop antibiotics and cell therapies – an innovative approach that merges biotechnology and AI. With $140 million in funding from strategic investors like Nvidia, the start-up positions itself at the intersection of two multi-billion dollar markets (AI + pharma/biotech). This could disruptively accelerate drug development and pressure established pharmaceutical companies.
Tokens: 1,628(1,042 in · 586 out)