- Dotika
- Posts
- OpenAI acquires Mac AI interface Sky
OpenAI acquires Mac AI interface Sky
ALSO : Anthropic secures massive Google TPU deal

News of the day
1. OpenAI buys Sky, an AI interface for Mac, aiming to embed its tech into daily workflows for enhanced productivity in writing, planning, and coding. → Read more
2. Anthropic is set to secure up to one million Google TPUs by 2026, marking a significant expansion of its AI infrastructure to meet growing computational demands. → Read more
3. An AI app analyzes facial movements to measure pain, aiding non-verbal patients but raising questions about subjectivity and treatment options.→ Read more
4. Reddit sues Perplexity, an AI search engine, for illegally scraping its content from Google search results, alleging it bypasses anti-scraping measures and relies on others' investments. → Read more
Our take
Hi Dotikers!
On October 23, OpenAI announced the acquisition of Software Applications Incorporated, creator of Sky, a conversational interface for macOS capable of understanding what’s on your screen and acting within your apps. This isn’t just an acquisition ; it’s an offensive move to make ChatGPT a native desktop layer, not just a chatbot in a window.
Sky comes with a rare pedigree. Its founders previously sold Workflow to Apple, which became Shortcuts, and they promise a deep macOS integration that will be built into ChatGPT: the assistant hovers above your desktop, follows you through writing, planning, and coding, then takes action. Cupertino has Siri (which doesn’t work!!!!) ; OpenAI is targeting the entire Mac.
Apple has already opened the door to ChatGPT through Apple Intelligence, which can call on the model to answer via Siri or generate content. The Mac ecosystem is thus becoming the stage for an unprecedented coexistence between a tightly controlled OS and an increasingly autonomous third-party agent. But there’s a red line not to cross: these agents, capable of reading your screen and clicking on your behalf, represent a new attack surface, where prompt injections and context hijacking can expose data and actions. The recent incidents involving browsers and AI agents serve as reminders that the race toward agentic computing is as much about UX as it is about security.
An excellent product bet from OpenAI ; provided it delivers security guarantees worthy of a professional environment. OpenAI clarified that a fund linked to Sam Altman held a passive stake and that the deal was approved by its independent committees. Success will be measured in the coming months by OpenAI’s ability to integrate Sky into ChatGPT without sacrificing privacy, user control, or reliability of actions.
G.
Meme of the day

Reply