ChatGPT's Power-Up Unveiled: Beta Web Browsing & Plugin Party for Plus Users!

Today:

ChatGPT Plugins And Web Browsing Beta Rollout For Plus Users

Alright folks, buckle up! OpenAI's given their chat whiz, ChatGPT, some serious upgrades. They've thrown in web browsing and a whopping 70 plugins for Plus users. Talk about taking things to the next level!

So what's the big deal? Well, with these shiny new toys, ChatGPT's turning into your personal butler, entertainer, job scout, and weatherman all rolled into one. We're talking a whole new ballgame of AI-powered internet chitchat.

Here's the skinny: Plus users can dip their toes in the beta phase with the web browsing and plugin features. Check your settings and, voila! You've got your hands on the future. And don't worry about getting lost - the latest version of ChatGPT's got you covered.

Web browsing's got some smarts too. Ask about the latest scoop, and it'll shoot straight for the web to bring you the goods. It's selective, though - it'll pick and choose the best from the bunch. And with the plugins? Oh, boy! Everything from home improvements to weather forecasts is just a chat away.

But don't get too giddy - you can only turn on three plugins at a time. And some plugins, like the PDF analyzer, are still a little rough around the edges. But it's early days, folks.

OpenAI's got a game plan for rolling out these new bells and whistles. It's a three-step dance: a small group gets first dibs (alpha), then Plus users can opt-in (beta), and finally, the whole shebang goes public.

Want to get your hands dirty? Developers can join the waitlist to start crafting their own plugins. Who knows, you might be the next big thing in the plugin store!

Big tech's in a full-blown arm wrestle for AI supremacy. OpenAI's pulling no punches with this new ChatGPT release, hot on the heels of Google's Bard expansion and Microsoft's Bing AI upgrades. The AI world's sure heating up, and we're all along for the ride!

READ THE ARTICLE ON SEARCH ENGINE JOURNAL.

Google Launches AI Supercomputer Powered by Nvidia H100 GPUs

Alright, let's break this down, y'all. Google I/O happened, and Google went on and on about how it's doing some real fancy stuff with artificial intelligence. They've got a new language model, PaLM 2, that's powering their Bard chatbot tool. This thing's going to be tucked into everything from Google Maps to Photos and even Gmail.

Now, all these fancy tools need a beefy engine under the hood, right? That's where Google's shiny new A3 supercomputer comes in. This big ol' monster, powered by NVIDIA H100 GPUs, is tailor-made to handle the heavy lifting for Google's AI tools. This thing's got some serious muscle, folks, with 26 exaFlops of AI performance. That's a whole lotta flops, if you ask me.

Underneath the hood, this beast is packed with Intel Xeon processors and a whopping 2TB of memory. But the real star of the show is the eight Nvidia H100 "Hopper" GPUs. These bad boys share data at a mind-boggling 200 Gbps, bypassing the CPU. Google's calling this the Infrastructure Processing Unit (IPU), and it cranks up the available network bandwidth by ten times.

Now, if you're a business and you're itching to take this A3 for a spin, you gotta fill out Google's A3 Preview Interest Form to get in line for the Early Access Program. But don't go counting your chickens before they hatch – submitting your info doesn't mean you're guaranteed a spot.

So, there you have it. Google's got a new AI supercomputer, and it's a real powerhouse. But remember, it's all fun and games until Skynet becomes self-aware.

READ THE ARTICLE ON TOM'S HARDWARE.

Spotify CEO Reports 500% Increase in AI-Related Podcast Discussions

So, the big cheese at Spotify, Daniel Ek, recently tweeted that there's been an out-of-the-park increase in chatter about AI on their podcasts, like a 500% surge in the last month alone. 'Course, this comes hot on the heels of Spotify giving the boot to thousands of AI-generated tunes on account of "artificial streaming" - that's the sneaky business of bots boosting audience numbers.

Despite the mass culling, Ek thinks AI is gonna be the cat's pajamas on the platform as new tools roll out. In fact, Spotify already dipped its toes in the AI pond earlier this year with a new DJ feature that suggests songs, partially using tech from the folks who cooked up yours truly, ChatGPT.

But wait, it ain't all roses. This AI hullabaloo is raising some eyebrows, especially when it comes to people like Lex Fridman, a computer scientist from MIT. This fella's been using his podcast soapbox to not only talk AI but also to stir the pot on hot-button issues like vaccine safety and gender roles.

Spotify, meanwhile, is keeping mum on the subject, not responding to requests for comment. But, hey, talk is cheap, especially when it's about AI on a podcast.

READ THE ARTICLE ON BUSINESS INSIDER.

Japan’s Panasonic Connect is going all in on AI

Panasonic Connect, the B2B branch of Japan's tech heavyweight, decided to hitch a ride on the AI bandwagon, not pondering "if" but "when" to start. They've whipped up their own AI sidekick called ConnectAI, which they're peddling to all 12,500 employees. From penning emails to digging up info to even writing code, this AI buddy aims to take a load off workers, freeing them up for brainstorming and problem-solving.

Japan, with a third of its 125 million folks past 65 and a population on the downswing, sees the promise of boosting productivity with AI. Panasonic didn't just grab a free AI tool off the shelf; they crafted a custom model using Microsoft Azure OpenAI Service for better data security and to keep an eye on its usage.

Since its debut in February 2023, ConnectAI has been a hit, with questions pouring in from employees at all levels, including Panasonic Connect's head honcho, Yasuyuki Higuchi. He sees a future where folks focus on complex work, letting AI handle the simpler tasks. Though it's early days, anecdotal evidence suggests ConnectAI's doing its job. One legal worker traded an hour-long document read for a 10-minute AI-produced summary.

The cherry on top? Despite the thousands of daily queries, the IT team didn't get a single call for help using the AI assistant. Seems like the employees are figuring things out on their own, even arranging brainstorming sessions centered around ConnectAI.

Panasonic Connect is already gearing up for the next ConnectAI update, considering how it can be used to train employees on new skills and help customer service reps answer product queries. As Mukaino, the senior manager of IT and digital strategy, puts it, the final output always needs a human touch - AI is just a kind of advice. So, with new AI tech pouring in every day, they're looking to further integrate their internal system into ConnectAI and expand its capabilities.

READ THE ARTICLE ON MICROSOFT.

Google Unveils Plan to Demolish the Journalism Industry Using AI

Well, folks, it seems like Google's taken a shine to bulldozing the journalism industry with their latest toy, AI. Remember when they scrapped "don't be evil" from their tagline? They're showing us why.

At this year's tech bash, I/O, Google flaunted their new search whiz, "Search Generative Experience" (or SGE for the cool kids). This thing's got an "AI Snapshot" feature that churns out a summary before you can even click a link. Ask about the eternal popularity of sourdough bread (their example, not mine), and bam, you've got an AI-cooked summary right there at the top.

Sounds harmless, right? Sure, if you don't mind Google, who's basically the internet's overlord with 91% of all search traffic, dictating the rules. Here's the kicker: if Google's AI is going to chew up and spit out original work without ever showing you where it came from, how are publishers supposed to make money? It's like going to a concert but only hearing the cover band, not the original artist.

Some folks, like The Verge's James Vincent, believe Google's new AI search is going to "gobble up the open web and then summarize/rewrite/regurgitate it." And it's not like they're hiring a bunch of writers to do the summarizing. Google's new model is basically a content-eating machine, digesting human-made content and spitting it back out, stealing precious clicks from the real workers in the process.

Google's stance on paying publishers for their content? Well, let's just say they're playing coy. Their spokesperson said they're all about "supporting a healthy, open web" and that they'll "continue to work with the broader ecosystem." But plans to pay up? They're keeping mum.

Publishers are sweating bullets over this change. RPG Site owner Alex Donaldson said, "If this actually works...this is literally the end of the business model for digital media."

Google's got a lot to answer for here. Sure, they say their goal is to maximize access to information, but SGE might be doing the opposite. If they don't figure out how to compensate publishers, the public's access to information could be in serious trouble. So much for not being evil, huh?

READ THE ARTICLE ON FUTURISM.

AI voice synthesising is being hailed as the future of video games – but at what cost?

Alright, grab a coffee and sit tight, we're diving into the world of video games, where the characters are starting to sound a whole lot more realistic.

You remember that mega-hit PlayStation 4 game Red Dead Redemption 2? That took 2,200 days to record 500,000 lines of dialogue with 700 voice actors. A logistical nightmare that makes herding cats seem like a walk in the park. But now, thanks to artificial intelligence, we're getting close to being able to mimic human voices, which could make the whole process a lot easier.

Take Australian software company, Replica Studios, for instance. They've created a voice synthesizer platform that was used in the game Age of Darkness: Final Stand. They've got 120 actors' voices on their books, able to produce 1,000 different vocal tones. It's like having a chorus line at your fingertips. But this isn't all sunshine and roses. It's more like a double-edged sword.

Here's the rub: these AI voice synthesizing tools can mimic anyone's voice, including celebrities, often without their permission. Some voice actors have even discovered their voices being used in places they never agreed to. It's like finding out your car was borrowed for a joy ride while you were sleeping. This doesn't sit well with everyone, including the Japan Performing Arts Workers’ Association who are clamoring for legislation to protect their jobs.

The big shots at Replica Studios insist they're playing nice by allowing actors to keep earning from the use of their voice even when they're not in the studio. They see it as a win-win situation. Bigger video game companies will still hire the big-name talent, while AI gives smaller studios access to a variety of voices, and early-career actors more work.

There's even a content moderation policy in place to prevent voices from being used in scams or illicit content. But it's all uncharted territory. The Media, Entertainment and Arts Alliance, an Australian union for actors, has yet to work out its stance on AI's impact on acting, insisting that actors should keep ownership of their voices and that quality work for performers shouldn't be overlooked.

Meanwhile, Canadian singer Grimes is embracing the AI voice synthesizing trend, inviting folks to create new songs using her voice synthesized with AI. But even she admitted to feeling stressed that people might produce Grimes-sounding songs better than her own.

So there you have it. The future of video games may sound more realistic, but at what cost? It's a brave new world, folks, and as always, with every step forward, there's a new set of problems to navigate. But hey, isn't that just the way the cookie crumbles?

READ THE ARTICLE ON THE GUARDIAN.

Dezeen launched an AI artwork competition

Dezeen's stirring the pot this week with a spanking new contest. They're hunting for some fancy AI art for their upcoming series on how artificial intelligence is going to jazz up architecture and design. They're paying a cool £1,000 to the winner, whose artwork will be judged by the "world's first AI designer" Tilly Talbot. Ten runner-ups also get their 15 minutes of fame on Dezeen.

Meanwhile, in the land of flat-packed meatballs and Abba, a Stockholm tech company called Luvly has gone ahead and rolled out a mini electric car. They're betting their bottom dollar that this thing is more energy efficient and wallet-friendly than nearly all its gas-guzzling and electric cousins.

Speaking of things that go vroom, a book on collector motorcycles has listed seven of the world's most sought-after bikes. In architectural news, Japanese architect Tadao Ando's design for the 2023 MPavilion in Melbourne was unveiled, while his compatriot, Sou Fujimoto, showed off his contribution to the Tokyo Toilet project. Yes, you heard that right, a Tokyo Toilet project.

Back stateside, LA's (W)rapper Tower, a creation of American architect Eric Owen Moss, was featured on Dezeen. Moss reckons it's the safest building around. So safe, in fact, that if an earthquake hit on Wednesday, you could still show up for work on Thursday.

Up north, Swedish designer Pavels Hedström's Fog X jacket, which turns fog into drinking water, won the first public vote in the Lexus Design Awards.

In less rosy news, there was a bit of a kerfuffle at Milan's design week. An exhibition with some racially insensitive glass figures got tongues wagging, leading designer Stephen Burks to pen an opinion piece reflecting on the incident.

Other popular projects on Dezeen included an office with thick log columns, a house that's actually 12 buildings in one, and a mirrored pavilion in Oxfordshire. And if you fancy cave-like or brutalist Mexican interiors, they had some lookbooks that'll tickle your fancy.

So, in case you missed any of the hoopla, you can always subscribe to Dezeen's newsletters. Because, as they say, the more you know...

READ THE ARTICLE ON DEZEEN.

The open-source AI boom is built on Big Tech’s handouts. How long will it last?

Google's Luke Sernau blew the whistle on a reality that's been tickling Silicon Valley's fancy: open-source AI models are shaking big tech's monopoly. These are lean, mean, almost-as-good-as-the-big-guys versions of AI models, sprinkled around like confetti at a parade.

While Google was busy obsessing over its own reflection, this little open-source uprising has been pinching its lunch money. And hey, we're not complaining. More access to these models drives innovation and keeps the bigwigs honest.

But there's a catch. This open-source fiesta is hanging by a thread. Most are piggybacking on giant models by cash-rich firms like OpenAI and Meta. If these tech titans decide to pack up their toys, our booming party town could turn into a ghost town.

OpenAI is already pulling up the drawbridge, fearing competition. Meta may also yank back its open-source code, not too thrilled about some upstarts using it for no-good. This could leave the open-source folks up the creek without a paddle, and the future of AI could fall back into the money-stuffed pockets of the biggest, wealthiest AI labs.

Open-source AI has had a wild ride. Startups like Hugging Face and Stability AI are churning out open-source models quicker than you can say "innovation", putting AI tech in the hands of millions. But these models aren't easy to train, and the bigger they get, the harder it is. Only a handful of orgs can actually afford to train these big models, so most open-source models are using ready-made ones like Meta's LLaMA.

Meta AI's been a real champ, nurturing open-source development. But even they're feeling the heat. These models can spit out misinformation, prejudice, and hate speech faster than a greased pig. So, the open-source crowd might have to brace for a cold winter if Meta AI decides to yank back its open-source models.

In a nutshell, AI's future is hanging in the balance. Will it remain a playground for the wealthy, or will the open-source party keep on rocking? Only time will tell.

READ THE ARTICLE ON MIT TECHNOLOGY REVIEW.