Tag

Ai Platforms And Assistants

All articles tagged with #ai platforms and assistants

ai-platforms-and-assistants10 days ago•145 min saved

Rambling Prompts, Real Results: How I Rewired My ChatGPT Use

The author experiments with ChatGPT’s voice mode by rambling for about 10 minutes on topics like holidays, gifts, and family plans, then lets the AI sort through the chaos. The result is surprisingly coherent insights that surface patterns and priorities polished prompts might miss. While careful prompts remain valuable for specific tasks, thinking out loud can reveal underlying themes and goals, showing a new way to interact with AI rather than a blanket replacement.

via TechRadar|

#ai #ai-platforms-and-assistants #chatgpt

ai-platforms-and-assistants4 months ago•74 min saved

Perplexity CEO says iPhone won’t be displaced by AI, thanks to on‑device processing and privacy

Perplexity CEO Aravind Srinivas argues the iPhone won’t be disrupted by AI, attributing this to Apple’s on‑device AI processing and strong privacy protections. He says the iPhone’s role as a personal hub for wallets, health data, photos, and communication makes it resistant to AI replacement, suggesting Apple could stay ahead in the AI race as AI improves.

via TechRadar|

#ai #ai-platforms-and-assistants #apple

ai-platforms-and-assistants5 months ago•62 min saved

Five prompts that push Gemini 3.1 Pro beyond basic

TechRadar showcases five prompts that demonstrate Gemini 3.1 Pro’s versatility: using adversarial reasoning to stress-test plans, mapping video moods to real-world locations (cinematic geography), applying spatial reasoning to fix a Lego model, generating and previewing interactive SVG animations, and leveraging Deep Research to produce a practical, high-end guide for niche projects like removable book bindings.

via TechRadar|

#adversarial-reasoning #ai-platforms-and-assistants #cinematic-geography

ai-platforms-and-assistants7 months ago•51 min saved

Signal founder unveils Confer: a privacy-first AI that encrypts chats end-to-end

Signal founder Moxie Marlinspike launches Confer, a privacy-first AI that encrypts chats by default and runs in a trusted execution environment, so conversations can’t be read by server admins and data isn’t stored or used for training. Messages are encrypted before they ever reach the server, keys stay on the user’s device, and remote attestation is used to verify code integrity. Confer aims to offer a private alternative to ChatGPT, with device syncing and a design that emphasizes confidentiality—even by the creators. It’s pitched as particularly appealing to privacy-conscious users and institutions.

via TechRadar|

#ai-platforms-and-assistants #artificial-intelligence #confer