Google just dropped Gemini 3.1 Flash Live, which is their newest and highest quality audio model designed to make real time conversations with AI feel more natural and reliable than ever. - Blazing fast response times and lower latency mean we can finally have fluid voice interactions without those awkward AI pauses. - A massive 90.8% score on complex task benchmarks shows it can handle multi step instructions way better for automation workflows. - Better tonal understanding lets the AI recognize pitch and pace so it can actually sense when you are frustrated or confused. - Conversation memory has doubled in Gemini Live which helps the AI keep track of your long brainstorm sessions without losing the thread. - Multilingual support is expanding to over 200 countries to make these voice tools accessible for global teams and automated customer service.

Are we about to see the end of the $20 monthly AI subscription? I just read a great article exploring why the future of AI might be moving away from giant data centers and right onto our local workstations. - Open source models are catching up to big names like OpenAI and Anthropic in as little as 4 months. - The gap in performance between paid frontier models and free open weight models shrank from over 8% to just 1.7% in a single year. - Running models locally is becoming a reality as the latest MacBook Pro M5 Max can handle models with up to 134.9 billion parameters. - Big providers are facing massive losses with OpenAI projecting a 14 billion dollar loss for 2026 which could lead to much higher subscription prices for us. - Small specialized models can achieve GPT 4o level performance at only 2% of the cost when fine tuned for specific tasks. - Local AI gives us total privacy and eliminates those annoying rate limits that slow down our automation workflows. Read more: https://tombedor.dev/open-source-models/

Stop pitching ideas and start showing working AI demos on day one. About 80% of prospects are impressed before even discussing pricing. If you want to stand out fast, this approach changes everything. 👉 btw you can create a working demo in 5 minutes on Trillet 😎

Hesitation isn’t random… it’s a signal. Ming Xu, CIO of Trillet, explains how he treats it as a cue to surface hidden objections, ask direct questions, and move the deal forward. That mix of speed and confidence is what turns conversations into cash.

In this video, I break down my proven 10-step onboarding process for building and selling voice AI receptionist services. ✅ Resource files in classroom: https://www.skool.com/automate-what-academy/classroom/dc7edadd?md=fc01fe4ae3ce4b94803486bd3da47fb5