OpenAI's Journey to a $80 Billion Valuation

Today:

OpenAI in Talks for Deal That Would Value Company at $80 Billion

OpenAI might soon be worth a whopping $80 billion! That's triple its worth from just half a year ago. They're chatting about a deal with Thrive Capital, a big investment group. If it goes through, OpenAI would be the top dog among San Francisco start-ups, trailing only global big shots like ByteDance and SpaceX.

Thrive is looking to buy more OpenAI shares, shooting the company's worth to that $80 billion mark. What's more, this won't be new stocks; it'll let folks at OpenAI sell their current ones. Why all the fuss over AI? Well, when ChatGPT came out, it blew folks away with its skills—from answering questions to writing poems and even computer code. This is part of a tech trend where machines can create content like text or pictures. While lots of companies are diving into this, only a few big ones, backed by a ton of cash and smart people, are making real moves.

READ THE ARTICLE ON THE NEW YORK TIMES

Amazon’s Andy Jassy Plans to Crash the AI Party

Andy Jassy, the big boss at Amazon, is gearing up to dive headfirst into the artificial intelligence (AI) frenzy. In a lengthy chat at Amazon's Seattle spot, Jassy made it crystal clear: he's itching to chat about AI.

With Microsoft teaming up with the AI wizards at OpenAI, and Nvidia raking in the big bucks with its must-have chips for AI training, the tech world is in a full-blown shake-up.

But Jassy, a two-year Amazon vet, isn't sweating it. He's playing the long game saying "We're just getting started," all while jamming to Foo Fighters and setting a chill vibe for his team.

READ THE ARTICLE ON THE INFORMATION

NVIDIA and AMD Deliver Powerful Workstations to Accelerate AI, Rendering and Simulation

NVIDIA and AMD are cookin' up some new desktops so folks can run AI stuff right from home. These ain't your grandma's computers; they're packed with NVIDIA's latest RTX graphics and AMD's beefy Ryzen processors.

Usually, big AI tasks need monster computers, like the ones in big data centers. Think of training a super smart language AI – it's like teaching an elephant to tap dance. But now, these new desktops can handle smaller versions of those tasks. You can even choose how many graphics cards you want.

These new machines are also great because they can take a load off those big data centers. And if you ever need to move your work from your desktop to a bigger machine or cloud, NVIDIA's got software to make that a breeze.

AMD's new processor has a whopping 96 cores – that's like having 96 brains in one chip. And NVIDIA's latest graphics cards are double the power of the old ones. From designing buildings to making movies to healthcare, professionals are gonna love these machines for all sorts of heavy-duty tasks.

READ THE ARTICLE ON NVIDIA

How AI Can Make Cancer Treatment More Equitable

The U.S. government has this big plan, the Cancer Moonshot, aiming to cut cancer deaths in half by 2047. It’s a team effort, involving everyone from the feds to healthcare folks, scientists, and regular folks dealing with cancer.

AI tests that can check out tumor samples and tell doctors not only how bad it is but also which treatments might work best. It's like a roadmap for each patient. And if AI's trained right, with data from all types of folks, it'll work wonders for everyone, even those who often get overlooked.

With AI, patients get more say in their treatment. They see the big picture, understand their choices, and feel good about the plan. That's super important because if folks don’t buy into their treatment, they might not stick to it, and that's bad news.

READ THE ARTICLE ON TIME

AI Can Flag Skin Cancer With Near-Perfect Accuracy

UK researchers have whipped up some fancy AI that can spot the meanest skin cancer with a perfect score. This thing caught over 99% of all skin cancers. Why does it matter? Well, skin cancer's a big deal in the US. 1 out of 5 folks here will deal with it by the time they hit 70. And if you catch it early, especially the big bad melanoma, you've got a 99% shot at kicking its butt. Problem is, we're only spotting about 75% of these early.

In this study, 6900 UK patients thought to have skin cancer got their pictures taken and those snaps were shown to the AI. The AI's guesses were then checked by a real-life skin doc. This clever AI nailed 59 out of 59 melanomas and caught almost all skin cancers. It also spotted over 90% of almost-cancers. That's a big step up from the older version from last year. And, bonus, it saved over 1000 in-person visits.

READ THE ARTICLE ON MEDSCAPE

How ChatGPT is transforming the postdoc experience

Rafael Bretas, a postdoc from Brazil based in Japan, had trouble with formal Japanese writing. But when OpenAI's ChatGPT came into the scene, it became a game changer for him, making his Japanese letters sound on point. Even though there's been a bunch of talk about ChatGPT shaking up jobs and industries, many researchers have jumped on the bandwagon to get a hand with tasks like drafting abstracts and working on code.

A survey by Nature showed that while a third of the postdocs used chatbots, the majority didn't feel any major shift in their daily grind. But, the times are a-changing. Postdoc Mushtaq Bilal believes we're just scratching the surface, and folks in academia usually take a hot minute to catch up with tech.

READ THE ARTICLE ON NATURE

The world’s biggest AI models aren’t very transparent, Stanford study says

Big AI models aren't being super clear about what they're doing, says a study from Stanford. They checked out the top 10 AI models from big names like OpenAI and Meta to see if they're telling folks about their work. Turns out, even the best ones, like Meta’s Llama 2 and OpenAI’s GPT-4, didn't score too hot on the openness scale.

The Stanford team looked at stuff like if these companies tell us where they get their data and if they’re using private info. Meta got a score of 54%, with OpenAI's GPT-4 at 48%. One big thing? None of these AI makers are talking about how their models might affect society. Stanford’s goal is to give governments and companies a way to measure transparency. There’s buzz about new laws making these AI companies share more.

Despite the name, OpenAI isn’t really sharing much anymore because they're worried about competition and safety. The Stanford team might look into more models later, but they're sticking with these 10 for now.

READ THE ARTICLE ON THE VERGE