AI-Powered Summaries on Amazon Reviews

Today:

Amazon taps generative AI to enhance product reviews

Amazon's upping their review game with some fancy AI magic. They're rolling out this tech that's gonna boil down what folks say about products into a short and sweet paragraph. No need to scroll through endless comments; you'll get the main vibes right there on the product page. They're also adding clickable buttons for key features, so if you're curious about, say, "ease of use," you can get right to those reviews.

Now, this ain't totally new. Amazon had these clickable words before, but they're making it smarter with AI. Right now, this shiny new tool is for some US shoppers on mobile, but if it works out, they'll push it to more products.

With AI getting smarter, it might start spewing out fake reviews that sound totally real. That's bad news for Amazon's AI summaries if they don’t keep those fake, AI-written reviews out. But Amazon's on it. They promise they'll only use verified purchase reviews for their summaries, and they're throwing a bunch of tech and expert folks at the problem to keep fakes away.

READ THE ARTICLE ON TECHCRUNCH

Humane will share more about its mysterious ‘Ai Pin’ the same day as October’s eclipse

Mark your calendar for October 14th! Not only is there a solar eclipse, but Humane, started by former Apple folks, will spill the beans on their new AI gadget called the "Humane Ai Pin". Word is, this gadget, which you wear on your jacket, might replace your smartphone.

At a recent TED talk, the co-founder showed off answering calls without a button, speaking in AI-translated French, and checking emails without needing a smartphone nearby. Big claims!

The company seems to think the solar eclipse is the perfect time for a reveal, symbolizing fresh starts. Bethany Bongiorno, another big name at Humane, can't wait for people to use their creation.

READ THE ARTICLE ON THE VERGE

MetaGPT Lets You Create Your Own Virtual Software Company

MetaGPT on Github is the new big thing, getting close to a whopping 10,000 likes. Think of it like this: you give it a one-liner idea, and bam! It gives you everything you'd need to get going – from the big picture stuff to the nitty-gritty details.

The whole thing runs smoothly like a well-oiled machine. The idea? Make code as simple as getting your team together and following a game plan. Plus, it's super cheap – you can get a taste for just 20 cents, and bigger projects cost you about two bucks. So, with MetaGPT, we're looking at the next level of making software.

READ THE ARTICLE ON ANALYTICS INDIA MAG

Intel partners with Synopsys for designing chips for contract manufacturing

Intel, that big ol' chip company, is opening its factories to other chip designers. They're partnering with Synopsys, a heavy-hitter in chip design, to make this magic happen. Their game plan? Help chip designers make better use of Intel's top-notch factory tech.

Intel's working on new tech where the chip circuits are super tiny (about 18 angstroms apart). For context, that's ridiculously thin. And this partnership means designers can tap into this crazy tech for their chip designs.

The real goal? Get these new chips out by 2024 and 2025. To paint a picture, these new chips might have tens of billions of tiny switches, a wild jump from the 2,300 on a chip back in 1971. These chips will be like a Swiss Army knife, good for loads of stuff – networking, AI, phones, computers, you name it.

READ THE ARTICLE ON VENTUREBEAT

After Raising $54M, Futureverse Unveils ‘JEN 1’ Text-To-Music AI Generator

Futureverse, fresh off raising a cool $54M, has just dropped JEN 1, their hot new tool that turns text into tunes. They say this AI beats out big dogs like Google and Facebook in the music-making game. How did they get here? Well, last year, 11 companies joined forces to create Futureverse. They're also dropping big bucks on other tech projects, like their $50M "Born Ready" fund.

Let's break down the JEN 1 magic: A team led by some real brainiacs looked into how music has all these different sounds and vibes. The result? JEN 1 can whip up some quality tracks. In fact, some music pros gave it a thumbs up, scoring it higher than other similar tools. The best part? It doesn't just sound good; it sounds REALLY good. Even Google's got something kinda like it, where you type stuff like "chill jazz for chillin'" and out pops a song. And Meta? They've got their own gig going with AudioCraft.

READ THE ARTICLE ON MUSIC BUSINESS WORLDWIDE

You can build your own customer service AI chatbot with this drag-and-drop tool

Botpress is this rad tool where you can whip up your own chatbots. Think of it like digital LEGO – you drag, drop, and connect different bits to get your bot up and running. At first glance, it seems simple. You make these question cards, and based on user replies, the chat moves from one card to another. But the real juice? It's got AI smarts. You can feed it all sorts of info, from documents to web pages, and it uses the ChatGPT API to get all analytical.

While it's user-friendly in some ways, there are these weird codes and expressions that can trip you up. Oh, and while you can start using it for free, once your bot gets chatty past 1,000 interactions, you'll have to drop some cash.

READ THE ARTICLE ON ZDNET

World's first mass-produced humanoid robot? China start-up Fourier Intelligence eyes two-legged robots with AI brains

Fourier Intelligence, a Chinese tech start-up, recently showcased its cool humanoid robot, the GR-1, at a big AI event in Shanghai, turning heads and raising eyebrows. Everyone's been talking about AI chat stuff since OpenAI rolled out ChatGPT, but Fourier's two-legged robot—able to walk 5km per hour and carry a hefty 50kg—made folks sit up and remember that walking robots are the next big thing. Companies worldwide, from Tesla to Xiaomi, are chasing this dream.

The big question, though: How do you make money off these robots? They're cool to look at, but they're also super pricey. Yet, big shots like Elon Musk of Tesla and Lei Jun of Xiaomi are all in. Musk recently showed off Tesla's robot, Optimus, which can dance and do chores. He thinks robots are best for boring or risky jobs, but Gu believes they can also be our pals, offering some feels and company.

With advancements in AI, especially with chatbot tech like ChatGPT, robots can get pretty smart, making them more human-ish. For now, Fourier plans to stick to building the robot's body and let AI pros handle the brainy part. Some AI companies are already knocking on Fourier's door to collaborate.

READ THE ARTICLE ON YAHOO! FINANCE

IBM Plans To Replace Nearly 8,000 Jobs With AI — These Jobs Are First to Go

IBM's boss, Arvind Krishna, recently dropped a bombshell: the company's sidelining almost 8,000 jobs for AI tech. This comes after an earlier announcement of a hiring freeze. So, what's first on the chopping block? Back-end gigs in the HR world. IBM's job postings also show a growing appetite for AI-focused roles.

What’s the game plan? In the coming half-decade, IBM's gunning for machines to snap up about 30% of the gigs that don't need face-to-face human touch. That means if you're in finance, HR, accounting, and such, you might soon be squaring off against Mr. Robot for your paycheck.

READ THE ARTICLE ON YAHOO! FINANCE

AI Is Building Highly Effective Antibodies That Humans Can’t Even Imagine

In an old cookie factory in South London, James Field's company, LabGenius, is using AI to whip up cutting-edge antibodies, not desserts. Naturally, antibodies fight off diseases. Since the 1980s, we've made synthetic versions for things like cancer treatment. But crafting these antibodies takes humans ages.

Enter LabGenius: They're automating the process. Using machine learning, robots, and some human oversight, they're exploring vast possibilities of antibody designs. They test a bunch, see what works, feed the results back into their computer system, and then test some more, constantly refining.

This ain't your grandpa's protein engineering – traditional methods might miss out on the best options. But LabGenius' way finds solutions in about six weeks that might surprise even the sharpest scientists. They've banked $28 million in investment and are teaming up with big pharma. In the end, it means better drugs for patients – and maybe fewer side effects. The future of medicine? It's looking techy.

READ THE ARTICLE ON WIRED

AI Pinpoints Parkinson’s Subtypes Using Stem Cell Images

Scientists from the Francis Crick Institute and UCL teamed up with tech company Faculty AI. They found out that computers can spot different types of Parkinson’s disease using close-up pics of stem cells. The highest accuracy? A whopping 95%!

The main things helping to identify the right type were parts of the cell called mitochondria and lysosomes. Basically, these tiny bits play a big role in how Parkinson’s develops. Before this, doctors didn’t really have a way to tell the different types apart, which meant patients sometimes didn’t get the best treatment.

They did all this by taking stem cells from patients, making 'em into four Parkinson’s types in the lab, and then teaching a computer program to tell 'em apart. They're hoping this new method will make treatments more personalized and even help test out new drugs.

READ THE ARTICLE ON NEUROSCIENCE