Florian Camiade 🗝️

Perplexity Labs officially replaces (almost) all experts Mood board, McKinsey slide deck, landing page... One prompt every time 9 concrete AI use cases (+prompts):

@itsPaulAi It almost gives me too many ideas

Copilot Tasks is seriously good?! Even one of the best alternative to Claude Cowork Using a single prompt it was able to: → Use a cloud browser to find the right tool → Interact with the page to enter data → Interpret all the info given by the page → Generate a PowerPoint based on it → Draft an email containing the learnings Because it has access in real time to multiple Microsoft Office apps like Word, PowerPoint, Excel, etc. but also Outlook. Everything was seamless. First try. And you can even ask it to schedule this task (or another one) to run it every Friday, once a month, you name it. As with OpenClaw, the use cases are basically infinite.

@itsPaulAi Hope no one bought 12 Mac Mini

You can run local AI models up to 120B without a $10,000 Mac Studio This phone-sized device is your own server for open-source models. 100% local and private. And you can use it to: - Power an agent like OpenClaw 24/7 - Completely replace a chatbot - Literally anything that requires an API It’s called Tiiny, and, once again, you can run the latest open-source models on it. Link below

@itsPaulAi Make an article is a very good format to go into detail. Now just send someone back to it if they have any questions on this topic or if they don't agree.

@itsPaulAi It will help me for someone I work with!

@FCamiade Hahaha I swear. 🔗in is full of these ones What do you have in mind?

Google has just launched Nano Banana 2 🔥 You have basically the quality of the Pro version with: - Up to 14 objects/people combined - Accurate text generation (perfect for infographic) - Enhanced reasoning But way faster (and cheaper) than Nano Banana Pro! Already available in the Gemini App, AI Studio, etc. And many thanks to Google DeepMind for the early access!

@itsPaulAi And we continue to take steps Again, power is not the important factor

No. You don't need a Mac Mini for OpenClaw. You can actually host everything you need on an old Android phone. And you'll have a setup which is: - Much faster - Way cheaper - With the same features Even a $25 phone can do the job.

@mhdfaran @itsPaulAi I am a very reliable friend💕

@FCamiade @itsPaulAi you lowkey living under paul posts 👀

Gemini can now generate music 🔥 You can create tracks with (custom) lyrics from a text prompt, a photo or a video. → Works for all styles → Duration of 30 seconds → Available on the web (mobile soon) Just select "Music" in the tool section, see below And thanks to Google for the early access!

@itsPaulAi We had fun with Comet... Google is catching up on this ground.

You can now automate anything using Gemini in Chrome Just describe the task and that's it. - Nothing to configure - Works on tools without API - Fully integrated in Google's ecosystem And since 99% of tasks are now browser-based, there are no limits. Even if it's "slower than doing it manually" you can let it run in the background. Currently available for AI Pro and Ultra subs in the U.S.

@mhdfaran We've gone from Will Smith eating pasta in 2.5D to this. No need to go out anymore, the image quality is better than in real life.

Grok imagine is now available at higgsfield

@mhdfaran This is my post of the year, see you next year!

@FCamiade welcome back

@itsPaulAi Ok, that's impressive... I'm curious to see what people will do with it.

Here we are Google just made Project Genie available so you can create interactive world with AI 🔥 1. Describe the world & character 2. Nano Banana Pro makes an image (you can adjust anything) 3. Genie 3 generates everything in real time Incredible.

@itsPaulAi The official representative of open source has spoken🔥

That's just insane Kimi K2.5 (which is 100% open source) is as good as Claude Opus 4.5 and GPT-5.2... And even beats them in key benchmarks 🔥 - 8x cheaper than Opus 4.5 (!!) - Weights & code available on Hugging Face - Multimodal w/ image, video, etc. Closed source labs no longer have any advantages. Open source is winning.

@itsPaulAi Do you think it's ergonomic on the phone? Or that it is better to use it on a laptop?

You can setup Clawdbot on a server without typing a single command So you can even use it on your phone from anywhere. - Just duplicate a template - Choose your model - Pick a channel (Telegram, WA, etc.) And you're done. 5 minutes max. Quick steps below

@itsPaulAi I love it, it even picked up your accent haha

So you can clone any voice 100% locally using this new open source model?! Alibaba has released Qwen3-TTS on Hugging Face. You can easily: - Create custom voices - Clone any voice from a VERY short audio - Generate speech with style instructions Only 0.6B & 1.8B! Sound on🔊

@itsPaulAi Glad to see that all these open source alternatives are more and more effective!

This is so good 🔥 You can run this new model on a laptop which is: - 100% open source - Only 3B active parameters (!!) - Way better than GPT-OSS - Perfect for vibe coding (and more) And already available for free on Hugging Face or via API. Open source models keep winning!

@itsPaulAi Every time I speak about AI model, I know I can quote Open Source models that are really serious thanks to you !

Wow that's quite massive Zai has just released an open-source model which is: - Better than DeepSeek 3.2 (in most benchmarks) - Competitive with Sonnet 4.5 and GPT-5.1 High - 4x to 7x CHEAPER than them And it's insanely good at vibe coding with great UI quality (see below). Open source is still winning!