Google DeepMind’s latest innovation aims to provide actionable insights, from resolving personal dilemmas to creating custom workout plans


Google DeepMind testing ‘personal life coach’ AI tool

Google’s AI branch, DeepMind, is cookin’ up a “personal life coach” tool. This nifty little thing could help you with over 21 types of tasks, from life tips to meal plans. For instance, it might give advice on how to tell your buddy you can’t afford to hit up their fancy destination wedding. But, not everyone’s on board. Some folks are raising eyebrows about how close we might get to these chatbots and the possible risks. 

Dr. Christian Guttmann, an AI bigwig, thinks it’s a cool step forward but stresses we gotta be smart and safe about it. Earlier, Google and DeepMind became one big AI family, aiming to push AI but in a chill, responsible way. Google DeepMind says they’re always checking their tech with partners to make sure everything’s A-OK.


Moemate’s AI avatar analyzes your whole screen, with spotty but intriguing results

Old AI helpers like Cortana are out; bigger and better is in. Big names like Amazon and Google are beefing up their AI sidekicks, but it’s not just them. Startups are also hopping on the bandwagon.

Now, meet Moemate. Think of a cute anime avatar on your computer that’s a mix of big-shot AI tech like GPT-4 and Claude. Instead of just texting back and forth, it checks out what’s on your screen to help out. Customize how it looks, how it sounds, and even its “personality”. But here’s a heads-up: that open customization might leave room for some mischief-makers.

Its target audience? Looks like streamers. It’s got tools geared for Twitch, like chatting with viewers and such. As a helper? Well, basic stuff feels old school. The wow-factor? It peeks at your screen to understand context. Say you’ve got a recipe up – Moemate can sum it up without you typing a thing. Or maybe you’re gaming and need a tip? It’s got your back. But it’s not perfect. Sometimes, it can get super random or miss the mark.


Buildbox introduces StoryGames.AI for creators to build game narratives with AI

Buildbox has rolled out this new tool, StoryGames.AI, which is like a game changer for anyone wanting to make their own visual novel-style games. You don’t need to be a tech guru or anything; this thing is user-friendly and helps regular Joes and Janes craft cool game stories with the help of AI.

Imagine it like this: You got an idea for a game story in your head. Pop it into StoryGames.AI, answer some questions about your main character and plot, and BAM! In minutes, you got a ten-chapter game, complete with choices and some real eye-catching visuals. And you can roll these games out on platforms like the Apple App Store, Google Play, and even Steam.

They’ve even thought about how folks can make some dough off of it. After whipping up a couple of free games, you can opt for a paid plan to keep the ball rolling. But don’t think you can just create any wild, inappropriate story – Buildbox is keeping an eye out and checking these games before they give ’em the green light.


Creator of AI-Powered GTA 5 Story Mode Mod Unlikely to Fight Back Against Take-Two After Shutdown

Bloc made a super cool AI-driven Grand Theft Auto 5 mod that let players chat with in-game folks. It was blowing up, but then Take-Two, the big game company, yanked it offline claiming copyright issues.

Bloc’s saying he doesn’t have the cash or energy to go head-to-head with these bigwigs in court. He’s pretty bummed, especially since his mod was totally free and he thinks it wasn’t breaking any rules. Some folks think this move by Take-Two is just ’cause they’re getting ready to drop Grand Theft Auto 6.

As for now, Bloc’s mod is toast, and he’s warning other YouTubers to watch out ’cause Take-Two’s on the prowl. Rockstar hasn’t said a peep yet.


Microsoft pulls AI-written article telling tourists to visit the Ottawa Food Bank

Microsoft had an AI whip up a travel article about Ottawa, Canada. Funny thing is, the article told folks to hit up the Ottawa Food Bank as a must-see tourist spot! It was even ranked third, right after the National War Memorial and before catching an Ottawa Senators hockey game. Pretty wild, right?

After getting called out, Microsoft took the story down and said they’re looking into how it got the green light. Even though they gave some journalists the boot in 2020 to let AI take over, looks like their robot might need a bit more training.


AI-generated books are sneaking into online bookstores. Here’s how to spot them

AI-generated books are flooding online stores, and it’s a real pickle for the publishing industry. Amazon, a big name in e-book sales, is grappling with AI books that are crowding out the real deal. Even worse, some of these digital pretenders are using real authors’ names without permission, leaving the authors to fend off complaints about trashy books they didn’t even write. 

Travel guides, cookbooks, self-help, and fiction – no genre is safe. 

It’s a Wild West out there, but readers can be on the lookout. Do a quick online search of the author’s name, scan the reviews for patterns that suggest they’re fake, or return a suspicious e-book to Amazon. AI chatbots like ChatGPT, Bing AI, Google Bard, and Claude are the culprits here, whipping up content that’s easily mistaken for human-written. 


Introducing the GenAI models you haven’t heard of yet

S&P Global is testing out new artificial intelligence models, including Llama 2 and other open-source models from Hugging Face. Many businesses start with OpenAI but eventually explore other options, such as Google’s Bard and Databricks, a data pipeline platform that also offers open-source language models like Dolly and Llama 2. These models will soon be available on Microsoft as well, as the field evolves rapidly.

KPMG is focusing on building flexible architectures for AI models, as the field changes quickly. The firm is experimenting with systems that can use models from OpenAI, Dolly, Claude, and Bard. Meanwhile, global consulting firm AArete is using Document AI from Snowflake, which allows for secure access to unstructured documents without risking private data.

Large foundation models are popular in the generative AI space, but some companies are looking for smaller, specialized models. Databricks, IBM, and AWS offer such models for various tasks, including code generation, image description, and specialized scientific tasks. Gartner analyst Arun Chandrasekaran says that these models are less complex and cheaper. Companies can use public versions of generative AI models or run them in private clouds, access them via APIs, augment them with embeddings, or develop new models by fine-tuning existing ones.


Consulting giant McKinsey unveils its own generative AI tool for employees: Lilli

McKinsey, a big-shot consulting company that’s been around for nearly 100 years, is making waves with its new AI chat tool called Lilli. This isn’t just any chatbot – it’s like the sum of all McKinsey’s knowledge in a chatbox. Imagine being able to ask everything McKinsey knows and getting an answer. Cool, right?

This new tool can dig into over 100,000 documents and interviews to give insights, data, and even suggest the best McKinsey expert for a job. They’ve been testing it since June, and so far, it’s been a hit – cutting down research time from weeks to hours or even minutes. And in just two weeks, Lilli answered a whopping 50,000 questions!

Now, what can Lilli do? It’s like a Swiss Army knife for McKinsey consultants. Need to know about a big online retailer? Lilli’s got you. Need the lowdown on the future of clean energy in the US? Ask Lilli. And even though it might take a second or two longer than other AI tools to answer, McKinsey says it’s all about quality over speed.


Meet Marqo, an open source vector search engine for AI applications

Vector databases are like the behind-the-scenes MVPs for AI, holding onto stuff like photos, videos, and words. This makes it easier for folks and computers to search for random content. Think of them as the digital storage heroes helping big AI brainiacs like GPT-4 (hey, that’s me!) do their thing in real-time.

Marqo’s founders, Jesse Clark and Tom Hamer, used to work at Amazon. Their big goal with Marqo? Making it easier to deal with messy data. They claim that Marqo’s got everything you need – from creating to storing and finding vector data. And, they’ve put it all under one roof, so you don’t have to juggle between different tools.


The Associated Press sets AI guidelines for journalists

The Associated Press (AP), a big shot in the news world, told its writers to tread carefully with using ChatGPT, an AI tool. They said, “Hey, don’t swap out real reporters for robots!” but gave some rules if journalists want to play around with it. 

Basically, if you get something from an AI, think of it like a tip you haven’t checked out yet. And don’t trust AI to mess with photos or videos. If they show an AI-made image, they’ll slap a label on it. AP’s also reminding writers to be sharp-eyed and not get fooled by fake AI content. 


The Desperate Hunt for the A.I. Boom’s Most Indispensable Prize

In the tech world, there’s a mad dash for a specific type of computer chip known as the graphics processing unit (GPU), and it’s getting wild out there. These chips are like gold for companies that deal with artificial intelligence because they’re great for running the calculations needed to sift through tons of data. Jean Paoli, CEO of an AI start-up called Docugami, has been on the hunt for GPUs, calling up anyone who might help, applying for government grants, and even using old gaming chips.

The frenzy began last year when chatbots got everyone excited about AI, causing a shortage of these chips. The situation is worse because Nvidia, a chip provider, dominates the market and can’t meet the skyrocketing demand. Usually, tech companies get their chips through cloud computing services like Google, Microsoft, and Amazon. But now, there’s a long waitlist, nearly a year in some cases, making it hard for businesses wanting to use generative AI (which can create its own images, text, and video).


Bill Gates says A.I. could transform education: It ‘will be like a great high school teacher’ who always gives useful feedback

Bill Gates chatted on his “Unconfuse Me” podcast about how AI could revamp education, comparing it to an ace high school teacher. Imagine a chatbot that offers solid advice on essay writing. Current tech? Not really cutting it, he says. But don’t think Gates wants bots replacing our hardworking teachers. No, he sees AI helping them out, especially in under-resourced spots. Teachers’ input is crucial, though. 

Without their feedback, bringing tech to the classroom might backfire, making educators feel sidelined. The Khan Academy’s got this AI tutor, Khanmigo. It’s kinda like a helpful classmate, but it’s got kinks to work out. Some teachers worry it hands out answers too fast. Khan’s dream? Using chatbots for group discussions, like mini teaching assistants. But remember, the face-to-face classroom chit-chat? Super vital for kids’ growth, especially after all the Covid remote learning. Gates gets it: school’s as much about hanging with pals as hitting the books.