Meta's AI Studio for Your Custom Chatbots

Today:

Meta debuts AI Studio to let developers build custom chatbots

Meta dropped a new tool called AI Studio at their big Connect developer shindig. It's for businesses to cook up their own chatbots for Facebook, Instagram, and Messenger, kicking off with Messenger. Mark Zuckerberg, the head honcho at Meta, says it's mostly for selling stuff and helping customers.

The tool’s just getting started, and Meta plans to ramp it up next year. They're also working on a sandbox tool so anyone can try making their own AI. They're planning to hook this up to their metaverse platforms, like Horizon Worlds, so you can have these chatbots in different online games and experiences.

READ THE ARTICLE ON TECHCRUNCH

Jony Ive and OpenAI’s Altman reportedly collaborating on mysterious AI device

Jony Ive, who used to design Apple’s biggest hits, and Sam Altman from OpenAI, are apparently brainstorming some new AI gadget. No one knows exactly what this gadget is yet, but people are guessing it might be a super innovative phone focused on using advanced AI. However, details are slim, and it's unclear if this device will actually be made or who would produce it, and the buzz has folks throwing around all sorts of guesses.

Some believe it’s likely to be a consumer device, considering Ive’s background, maybe something with an AI system that can create apps on the spot or pull up info from conversations around you. Yet, others doubt it’s a smartphone. Former Microsoft bigwig, Steven Sinofsky, said building a phone isn't groundbreaking and hinted it might not be the smart move.

READ THE ARTICLE ON ARS TECHNICA

OpenAI is reportedly raising funds at a valuation of $80 billion to $90 billion

OpenAI might be selling shares, bumping its value from $29 billion to around $80-$90 billion. Workers could sell their current shares instead of the company making new ones.

Earlier this year, OpenAI scored $300 million from big-name backers and got a hefty chunk from Microsoft too. Their AI chat tool, ChatGPT, has been a big hit since it came out about nine months back. Now, they're saying you might soon be able to chat with it using your voice.

Even though Microsoft owns almost half of OpenAI, the AI company is on track to make $1 billion in 2023. TechCrunch is waiting on a comment from OpenAI.

READ THE ARTICLE ON TECHCRUNCH

Major ChatGPT Update: AI Program No Longer Restricted To Sept. 2021 Knowledge Cutoff After Internet Browser Revamp

OpenAI announced that their AI program, ChatGPT, got a major upgrade. So, instead of being stuck with info only up to September 2021, it can now surf the web for the latest scoop! It’ll use Microsoft’s Bing to find info and answer questions.

This fresh feature is rolling out to the folks who are shelling out for premium plans first, but it’s gonna be available to everyone pretty soon. The date limit was a big thorn in the side for some users, so this should smooth things over with them.

READ THE ARTICLE ON FORBES

The CIA says it’s building a ChatGPT-style generative AI chatbot to aid data analysts

The CIA is whipping up a chatbot, kind of like ChatGPT, to help its analysts sort through tons of data for their investigations. This chatbot, made by the CIA’s Open Source Enterprise unit, uses fancy AI to help agents dig through open-source data. They ain’t keeping it just to themselves—it's part of a big push to step up AI-powered intelligence across multiple agencies to keep pace with China.

Randy Nixon, the big wig of the CIA’s AI unit, says there’s so much info out there, from newspapers to big data, and the agency needs AI to sift through it all. “Finding needles in a needle field,” he calls it. And the U.S. is feeling the heat to keep up with China's fast-growing AI capabilities. China's got big plans in AI—they even got a “prosecutor” bot nailing convictions at a 97% rate.

READ THE ARTICLE ON SILICON ANGLE

Mistral AI makes its first large language model free for everyone

French AI startup, Mistral AI rolled out their first big language model, and guess what? It's free for everyone! They say it's a big deal because it works better than other models its size.

So, anyone can download this Mistral 7B model, and there's no catch. It doesn’t matter if you’re just playing around with it, or if you’re a big-time company, or even the Pentagon—you can use it. You just need the tech to run it, or you gotta be cool with paying for cloud services to get it going.

Mistral’s team said that they want to lead the way in supporting open AI and take open models to the next level. This model, Mistral 7B, is like proof of what smaller models can do. They said they built it from the ground up in three months. They knew their stuff because they had worked on similar things at Meta and Google DeepMind before.

READ THE ARTICLE ON TECHCRUNCH

Cloudflare ignites AI platform efforts with serverless inference, vectorize database and AI gateway

Cloudflare is revamping its game, focusing on AI with a bunch of new updates. They've rolled out a new service called Workers AI, which lets AI models work all over the world without needing a constant server – it’s a pay-as-you-go deal. The company is big on managing and keeping data safe, and now it’s diving deep into AI, teaming up with Hugging Face and Microsoft. The partnership means that users can easily deploy models on the Workers AI platform, powered by Microsoft's ONNX runtime model.

On top of this, there’s the new AI Gateway by Cloudflare. This isn’t just a regular gateway; it lets organizations keep an eye on their AI apps, controlling how they are used, and gives insights into usage patterns. It’s like a watchdog for AI applications, managing, monitoring, and making sure everything’s under control. It even helps with caching and rate limiting to ensure everything runs smoothly.

Cloudflare is also introducing Vectorize, a new vector database, to store embeddings and other vectorized data, making data accessible where it’s needed. This helps in scaling AI applications more efficiently by having the data close to where the inference happens with Workers AI.

READ THE ARTICLE ON VENTUREBEAT

AI, implants form ‘digital bridge’ to help paralyzed man move arms, hands

A Swiss man, paralyzed after a fall, is regaining movement thanks to a groundbreaking surgery involving AI implants. This dude had an implant put in his brain that reads his thoughts about moving and sends them to another implant in his belly. This tech is the brainchild of a Dutch company, Onward. It’s all about turning thoughts into movements, allowing the guy to move his arms, hands, and fingers again. It's still early days, but things are looking up.

Dr. Gregoire Courtine, a big brain from France, and Dr. Jocelyne Bloch from Switzerland teamed up for this. They had this wild idea more than ten years ago but had to wait for tech to catch up. Bloch is the surgeon who puts in the electrodes. The whole thing works wirelessly, activating spinal cord stimulation.

They pulled off a similar stunt earlier with a Dutch guy who lost the use of his legs. It’s not perfect, but the guy is walking. The more he walks, the more his body is repairing itself, it seems like it's doing some healing mojo on the nervous system. But making hands and arms move is a tougher nut to crack than getting legs moving, it's trickier and needs more refinement.

READ THE ARTICLE ON CNN

AI Hurricane Predictions Are Storming The World Of Weather Forecasting

So, Hurricane Lee was out in the sea and it wasn’t clear where it was gonna hit, but meteorologists were keeping a close eye, right? Usually, they’d use models based on the atmosphere to make a call on where it would land, but this time they had something new - AI models from Nvidia, Huawei, and Google’s DeepMind. These AI models were on the money, predicting a path between Rhode Island and Nova Scotia which was in line with the traditional models.

Mark DeMaria, an atmospheric scientist, was initially doubtful about AI models but changed his mind when he saw them working well. Hurricane Lee did land within the range of the AI predictions, hitting Nova Scotia. These AI models, they’re good - sometimes even better - at tracking tropical storms, and they do it super fast compared to the old school methods.

READ THE ARTICLE ON WIRED

Microsoft and Mercy collaborate to empower clinicians to transform patient care with generative AI

Microsoft and health system Mercy are teaming up for a long-term project, leveraging advanced AI to give medical staff more time with patients, enhancing patient experience. The focus is on generative AI, which is like a smart tool that can help doctors and nurses do their jobs better and make health care organizations more efficient.

In practical terms, patients at Mercy will get clearer insight into their lab results and health status, making discussions with providers more informed. The AI will also help with administrative tasks like scheduling appointments and answering staff queries about company policies and HR matters, saving time and reducing the need for follow-up calls.

The collaboration is utilizing Microsoft's Azure OpenAI Service, aiming to launch multiple AI use cases by mid-next year to personalize and improve care and the working experiences of healthcare providers. These innovations can help in real-time decision-making and are a part of ongoing efforts by both organizations to better healthcare through technology.

READ THE ARTICLE ON MICROSOFT