Meta Introduces CM3Leon, the Revolutionary Image Generator Redefining Artistic Possibilities


Meta claims its new art-generating model is best-in-class

Meta’s tossed its hat in the ring with a new AI model, CM3Leon, that’s got a knack for generating pictures from words. Basically, you tell it what to create, and it whips up a digital image. This ain’t your average doodle bot though; CM3Leon can even write captions for the pictures it makes. That’s a step up from what’s out there and Meta’s calling it the cream of the crop.

CM3Leon pays attention to the details, taking into account the important parts of the prompt. This makes it faster to train and easier to scale up. Plus, Meta’s boasting that it’s more efficient, needing less computer power and a smaller learning playground than the competition.

To get CM3Leon up to snuff, Meta showed it millions of images from Shutterstock and taught it a trick called supervised fine-tuning (SFT). This helped CM3Leon not just in creating images, but also in caption writing, even following detailed instructions like “change the color of the sky to bright blue”.

But the tool isn’t just a one-trick pony. CM3Leon can tweak existing images, too. Say, adding a bottle to an image of a room with a sink and a mirror. Meta claims it’s got the upper hand here compared to others like DALL-E 2, which sometimes miss the mark. And let’s not forget, it can write captions, even with less text in its training data than other models.


Character.AI in Talks to Raise Funding as Meta Platforms Tests Rival

Character.AI, a startup that builds chatbots acting like famous folks such as TV’s Tony Soprano or Tesla big boss Elon Musk, is looking to raise more dough, as per someone in the know. Just a few months back, they bagged $150 million, boosting their worth to a cool billion.

This fresh cash injection could help the fairly new company, just a year and a half old, meet the growing hunger for their chatbots. These require a hefty amount of computing power to provide responses. However, they need to keep their millions of fans interested while also dealing with big tech companies, like Meta Platforms, hot on their trail. Meta is currently trying out a similar concept, letting folks chat with virtual characters like Abraham Lincoln.


Mastercard, eBay and Capital One talk equitable generative AI and innovation

Mastercard, eBay, and Capital One are mulling over the fairness and innovative potential of generative AI at the Women in AI Breakfast event. Emily Roberts from Capital One emphasized the need to build diverse and continuously learning organizations to tackle the challenges of AI. She expressed the importance of having a broad range of voices in the conversation.

eBay’s Xiaodi Zhang suggested building in checks and balances from the start to ensure fair results. She talked about how generative AI is a whole new ballgame, calling for constant learning, flexibility, and a willingness to experiment.

As generative AI finds its way into new fields, the companies are proceeding with caution. eBay had a hackathon solely about gen AI, while Mastercard encouraged internal innovation, but with guidelines to ensure things don’t go sideways. The ultimate goal is to produce unbiased AI applications before they hit the public.

Regulations around generative AI are still a gray area, with companies unsure what will be required for compliance. Zhang urged the need for flexibility to react to new regulations as they come. Roberts reiterated Capital One’s commitment to transparent and well-controlled experimentation and application development.


The Viral AI App That’s Triggering Baby Fever

The latest app craze, Remini, is giving folks a sneak peek into a future of family life they might not have considered. Remini outstripped Meta Platforms’ Threads to be Apple’s top free app, letting folks picture themselves in bridal gowns or maternity threads. Heck, it’ll even throw together a family snap with you and some AI-created kids.

The app’s got a similar vibe to Lensa. You upload a bunch of selfies, and the app uses some smart tech to create your images. It’s got a three-day free trial, but after that, it’s gonna cost you $10 a week for all the cool editing tools.

Remini started in 2019, focusing on sharpening blurry snaps. But last year, it added AI that lets you create some stylish avatars from your selfies. It’s also got a new feature that lets you check out what your kiddos might look like if you uploaded photos of you and your partner.

But don’t think it’s all roses. It takes a few minutes for the app to create images and some results can look a bit off, with weird hands and sometimes even mixed up races. But the company’s working hard on fixing these bugs.They also assure users they’re taking care of their snaps. Pictures, face data and AI images are deleted from their servers after 30 days. You can also ask the company to delete your data whenever you want.


Another player heats up generative AI race as China introduces interim laws, a big online shopping platform in China, has unveiled a new, bigger and smarter language model, called ChatRhino, upping the ante in the race of artificial intelligence (AI). The system can help in several industries, from healthcare to finance, making tasks easier and faster. ChatRhino outperforms its predecessor, Vega, and even some models from tech bigwigs like Microsoft and Facebook.

Aside from language, ChatRhino can also whip up visuals like marketing posters in no time, saving costs and hours of work. In healthcare, it helps create AI models in minutes, a task that used to take a week and a bunch of scientists.


Actors say Hollywood studios want their AI replicas

Hollywood actors confirmed they’re striking in response to studios proposing to use and own their AI replicas indefinitely, for free. This revelation was made by Duncan Crabtree-Ireland, the top negotiator for the Screen Actors Guild‐American Federation of Television and Radio Artists (SAG-AFTRA).

Studios proposed that actors could be scanned, paid for one day, and then the studios could use that digital likeness forever, on any project, without further permission or payment. The studios’ group, the Alliance of Motion Picture and Television Producers (AMPTP), says that’s not accurate and claims digital replicas of actors could only be used in the film for which they were made, and any further use would need the actor’s agreement and extra payment.

The issue of using AI to generate replicas of actors has become a major bone of contention in the ongoing negotiations. SAG-AFTRA president, Fran Drescher, warned that actors risk being replaced by machines if they don’t push back now. The strike kicked off at midnight.


The EU Urges The US To Join The Fight To Regulate AI 

EU Justice Commissioner, Didier Reynders, is nudging the US to join the fight in regulating artificial intelligence (AI) and safeguarding user data. While the US has big tech, it’s pretty lax on legislation, leading to the EU becoming the watchdog. Reynders has had his fill of the US’s all-talk-no-action approach.

Reynders has been talking shop with US officials like Attorney General Merrick Garland. He’s pushing for stronger data privacy laws, similar to the EU’s General Data Protection Regulation (GDPR) that’s been running for five years. While tech bigwigs and Congress members have suggested similar measures, they haven’t walked the walk yet.

Despite the US Federal Trade Commission getting tech companies to play nice with user data or face fines, Reynders believes they’re not biting hard enough. He’s worried the same thing will happen with AI regulation, leaving this powerhouse tech unchecked.

Reynders wants the US to help establish international standards on AI, but if the US doesn’t match the EU’s upcoming AI Act, it’ll be tougher to get tech giants to toe the line.

OpenAI’s chatbot, ChatGPT, is under the microscope for both privacy and AI-specific regulations. An investigation into its GDPR compliance is due later in the year. While OpenAI’s CEO, Sam Altman, is keen on new AI rules, he’s wary of overregulation. Reynders believes Altman has business reasons to play nice with the EU, but he also wants more AI tech to be open-source, fostering an open market and encouraging new businesses.


U.S. companies are on a hiring spree for A.I. jobs—and they pay an average of $146,000

The US job market’s thirst for artificial intelligence (AI) skills is skyrocketing, and these roles pack a hefty paycheck, new findings from job search platform Adzuna show. In June, 169,045 job openings in the US cited AI requirements, with a substantial 3,575 specifically seeking generative AI expertise. Typical roles include software engineers, product designers, data scientists, and deep learning architects, but tax managers are also in high demand as firms seek to streamline operations using large language models. AI-oriented tax manager roles average an annual salary of $100,445, while overall, jobs requiring AI skills average $146,244.

According to a recent LinkedIn survey, nearly 70% of top companies already see speed and efficiency gains from AI use, with an additional 32% expecting more benefits in the future. Big players like EY, Wells Fargo, and Kaiser Permanente are highlighting AI as a hiring priority and embedding it into workflows. James Neave, Adzuna’s head of data science, encourages those looking for high-paying roles to focus on AI, as it’s an increasingly sought-after skill across industries.

While the US leads the way with approximately 173,000 AI job ads in June, India and the UK trail behind, posting 25,900 and 16,825 jobs, respectively. Those eager to join this AI wave can explore online certification and training courses from the likes of the University of Michigan, Coursera, and other e-learning platforms.


China takes major step in regulating generative AI services like ChatGPT

China’s internet watchdog has laid down new rules for AI tech like ChatGPT. They’re one of the first countries to put up guardrails around this exploding industry. These rules kick in on August 15.

The new rules are less tight than an earlier draft, and it looks like China’s seeing a chance to boost its economy and create jobs. Just last week, they hit Ant Group, a huge tech company, with a fine just shy of a billion dollars. Now, big names like Alibaba, Baidu, and are working on their own AI chatbots.

The rules only count for services available to the public in China. Stuff being cooked up in research labs or for folks outside China gets a pass. They’ve also scrapped some steep fines for breaking the rules.

China’s government is all for using AI in different industries and backs the development of reliable tech stuff, like chips and software. They’re also pushing for a say in setting international rules for AI.