OpenAI rolls out GPT-Image-1

Genspark just introduced a tool that helps with creating professional presentation that would usually takes hours of brainstorming, designing, formatting. With just a few clicks, AI Slides can turn your ideas, outlines, or even rough notes into a polished, ready-to-present slide deck. All you have to do is input your topic or upload a document, and the AI handles the heavy lifting. It researches, writes, organizes the flow, designs the slides, adds visuals, and even suggests speaker notes if you want them. Whether it’s for a business pitch, school project, marketing update, or a quick team meeting, AI Slides takes you from zero to professional faster than ever. It is also completely free right now. No expensive templates. As AI gets smarter about storytelling and visual communication, Genspark’s tool feels like the start of a trend: where presenting great ideas becomes more about what you say than how good you are at graphic design. If you are someone who’s tired of wasting time formatting presentations, Genspark AI Slides might just become your new favorite productivity hack.

OpenAI quietly rolled out GPT-Image-1 earlier this week, it is their new image generation model and it’s already making waves across creative tools like Figma and Adobe Firefly. Unlike older AI models that could generate stunning images but struggled with things like realistic text inside designs or maintaining consistent style, GPT-Image-1 has world knowledge baked into it. That means it understands what you’re asking for, why you might need it, and how to place things naturally. For example, you can prompt it to “create a poster for a summer music festival with funky handwritten text” and it will not only design a vibrant image but also generate actual realistic lettering, not messy gibberish. OpenAI has built GPT-Image-1 specifically to integrate inside real creative apps. You don’t have to jump between platforms anymore. Instead, you can summon AI art and designs right inside the software you’re already using — Figma, Adobe, and more coming soon.

If you have ever tried syncing someone’s voice to a video manually, you know it’s not as easy. Now, Tavus is changing that with their newly launched model: Hummingbird-0. Hummingbird-0 is not just another AI model, it is a zero-shot lip-sync system. What this mean is that you can take any short video of a person, feed it any voiceover or audio clip, and Hummingbird-0 automatically reanimates the lips to match the speech, all without needing to train on the speaker’s data beforehand. Before now, good lip-sync AI usually needed lots of customized training per person (think hours of video footage). Tavus’ breakthrough makes it instant and super realistic. The precision is impressive; facial details, expressions, and speech timing are almost perfectly preserved. For creators, brands, educators, and marketers, this unlocks massive opportunities. You can now instantly dub video content into different languages, creating personal messages at scale, or editing voiceovers without re-filming. The demo clips are jaw-dropping. It’s almost impossible to tell it’s AI doing the syncing. With Hummingbird-0, Tavus isn’t just improving video editing, they are completely reinventing how we think about content personalization. And it’s only the beginning.

Research just got a little easier and a lot faster with ChatGPT Deep Research Mini. Until now, OpenAI’s Deep Research tool was mainly a power feature for paying users or those needing heavy-duty reports. But now, they have launched a lightweight version for free users, with five free uses a month. Here’s how it works, instead of skimming through 20 tabs trying to find sources, you can just ask ChatGPT a deep question like “How has AI changed recruitment in the last 5 years?” The Deep Research Mini doesn’t just give you result promptly; it gives you a structured, detailed, cited report, almost like a mini analyst summary within a few minutes. It uses OpenAI’s new o4-mini model behind the scenes, built to balance speed and detail. Paid users (like ChatGPT Plus and Teams users) will get even more credits and deeper answers. But even free users now have access to a level of research quality that, honestly, beats a basic Google search.

Hummingbird-0 is Tavus’ new zero-shot lip sync model. No cloning. No training. Just upload an MP3 and MP4—and it syncs mouth movements better than any tool out there.