133 - Llama Unleashed and Midjourney Has Competition

This week's episode of TheChatGPTReport kicks off with a spotlight on Fcamiade, the creator of the week, who's sharing a clever method for crafting pro infographics using ChatGPT 4o and Canva, suggesting it's now easier than ever to whip up viral content. This naturally leads to a bigger question: is Midjourney, the go-to for many image creators, starting to feel a bit behind the curve? The seamless text integration in 4o and its convenient placement right within ChatGPT are definitely turning some heads, drawing comparisons to Grok's integrated experience. Even @DavidShapiro has voiced strong opinions on this, prompting a curiosity about what @DavidSHolz thinks.

Beyond the creative tools, there's a flurry of activity in the broader AI world. Meta has just dropped its new Llama 4 models, boasting improvements in handling different types of data and longer context. Google is also making waves with its experimental Gemini 2.5 Pro, claiming it's their smartest yet, and they're playing with more personalized AI experiences. Amazon's got a new speech model called Nova Sonic that's aiming to understand the nuances in our voices, and they're experimenting with AI in retail, like a feature that finds products based on your interests and even one that lets you buy from other sites through the Amazon app (though it sounds a bit clunky). On the hardware front, Broadcom is rolling out new chips designed for the heavy demands of AI, and in the investment world, Perplexity AI has launched a fund for AI startups, while OpenAI is reportedly swimming in new funding. And for a bit of fun, @iruletheworldmo shared a tweet joking about Microsoft's Co-pilot being like a time machine to the past.

@DavidSHolz@iruletheworldmo@DavidShapiro@fcamiade