Today:
Create Custom GPTs | No code AI automation | OpenAI App Store | Complete Guide
GPT Builder, a cool tool for whipping up custom GPTs to make your life easier. First, you hop onto chat.open.com and use the GPT Builder to start shaping your own AI buddy. The process is pretty chill – you kick around ideas with the Builder, and it helps flesh out your vision.
Whether it's sorting your shopping spree mess or getting the scoop on Tesla's latest 10K, the GPT Builder is your go-to for creating personalized AIs. It's like having a bunch of mini-experts at your fingertips. Just remember, the more specific you are with your requests, the better your custom GPT will perform.
Biden, Xi set to pledge ban on AI in autonomous weapons like drones, nuclear warhead control
Biden and Xi, the big wigs from the US and China, are shaking hands on a deal to keep AI out of drones and nuclear weapons. They've got 36 countries on board, planning to hash out the details next year. The US already made a move in October, saying any fancy AI stuff needs a thumbs up from the government, especially to make sure it's not used for making bio or nuke weapons.
Despite their tiffs, the US and China are finding common ground on tech rules. Xi's been big on AI, aiming to beef up China's game in this area. Their army's been toying with AI in weapons too.
Experts are nodding their heads at this. They're saying it's key to keep humans calling the shots with nukes, given the hiccups AI's had. The Stanford brainiac Mastro says armies are getting all AI-ed up, so the first step is to agree not to let robots take over nuke decisions.
Mastro also chimes in that Biden and Xi will keep the nuke chat going, but don't expect them to dive deep into arms control this round. China's touchy about that, seeing as their arsenal's smaller than the US's.
Stability AI Gets Intel Backing in New Financing
Stability AI, the folks behind that cool Stable Diffusion image-making tech, just scored a big win. They've bagged nearly 50 million bucks in funding, and guess who's holding the purse strings? Intel Corp, the big-shot chipmaker.
This cash boost is a big deal for Stability AI, especially right now when they're at a critical point in their journey. Plus, Intel's not just writing checks; they're teaming up with Stability AI to build a monster AI supercomputer, with Stability AI as the star customer. But it's not all smooth sailing - the company's been dealing with some top-level staff heading out the door and some boardroom beefs
Investors use AI to glean the truth behind executives’ soothing words
Big-shot investors are now using AI to figure out what's really going on behind the smooth talk of company bigwigs. Take Francis deSouza from Illumina, for instance. He was trying to play it cool about some messy business with a $8 billion deal and all sorts of pushback, but this AI tech from Speech Craft Analytics caught him sounding nervous just by the way he talked about it. This tech listens for tiny changes in how folks speak, like stutters or even a gulp, to spot when they're tense.
This trick has got the attention of some major money managers. They're already using algorithms to sift through what these execs say in earnings calls and presentations, looking for clues in their choice of words. But now, they're also tuning in to how they say it. According to Mike Chen at Robeco, an investment firm, this audio analysis can pick up on stuff you won't catch just by reading a transcript, like tiny voice shakes we can't even hear.
Robeco's been mixing these audio clues into their investment strategies and they're seeing good results. They think more investors will jump on this bandwagon. It's like a high-tech game of cat and mouse, with fund managers trying to outsmart company heads who know they're being listened to. Turns out, since AI started analyzing their words, execs have been talking more positively to mess with the algorithms.
OpenAI's GPTs Go Live After Fending Off Hackers
Sam Altman, the big boss at OpenAI, was all set to roll out these new, fancy chatbots called GPTs for everyone last Thursday. But, hold your horses, because just as they were about to launch, the whole thing hit a snag. Turns out, there was a big-time cyber attack, a DDoS, that threw a wrench in the works. This attack wasn't just a little hiccup—it caused a massive blackout for ChatGPT and its API worldwide.
These hackers, going by "Anonymous Sudan," have a rep for causing trouble. They've been tagged as possibly being backed by Russia and have messed with Microsoft and even tried to mess with Sweden's NATO deal. They jumped on Telegram claiming they were the ones messing with OpenAI.
These Anonymous Sudan folks say they did it to show support for Palestine and because they're not cool with OpenAI's ties to Israel. Last June, Altman was talking up how Israel is gonna be a big player in the AI game, and that seems to have rubbed them the wrong way.
Bill Gates Just Described Exactly How A.I. Will Transform Computers
So, everyone's buzzing about A.I. these days. It's like the hottest topic out there, even beating out other tech crazes. But is it just hot air? Not really, according to some big brains in the field. A.I. is set to shake things up big time, kinda like how electricity did back in the day. Even President Obama's making comparisons to electricity when talking about A.I. That's a big deal, right?
Now, Bill Gates, you know, the Microsoft guy, he's saying that soon, using a computer will be a whole new ballgame. He's talking about everyone having their own A.I. buddy to help with everything – from writing emails to planning parties. No more juggling different apps for different stuff. Just tell your A.I. pal what you need, and bam, it's done. It's like having a super-smart friend who knows you inside out.
These A.I. helpers, or "agents", aren't a new idea. Remember Clippy, that annoying paperclip in Microsoft Word? Yeah, A.I. is way beyond that now. It's gonna be more like a personal sidekick that gets smarter over time.
Microsoft’s Chris Young on Bringing AI to Main Street
Chris Young, who used to run McAfee, hopped over to Microsoft in November 2020. His gig? He's all about business development, strategy, and ventures. Think of him as the guy trying to push Microsoft forward by making smart partnerships, setting up the game plan for the company, and finding cool new investments through Microsoft's venture arm.
Now, Young's got his eyes set on AI, specifically how it can be a game-changer for folks who aren't all that tech-savvy. He sees a huge opportunity here, like a gold mine waiting to be tapped. He's thinking about how AI can be more mainstream, helping out the average Joe and Jane on the street. We're talking about bringing AI to Main Street, making it more accessible and useful for everyday people and businesses.
And here's a kicker – the pandemic and all that remote work jazz gave him some unique insights. He's been thinking about how work is going to change in the future, especially with AI in the mix. It's not just about making things more high-tech; it's about making sure everyone can get on board, making it as inclusive as possible.
Google Cloud demonstrates the world’s largest distributed training job for large language models across 50000+ TPU v5e chips
Google Cloud just showed off a huge project where they trained a big language model using over 50,000 TPU chips. That's like having a small army of supercomputers working together!
They used something called Google Kubernetes Engine (GKE) to manage all these chips. They also used XPK, which is a tool that sets up and manages these computer clusters. It's like having an assistant for the conductor, making sure every part of the orchestra is in sync.
To speed things up, they trained the model using a method called Accurate Quantized Training (AQT). This is a fancy way of saying they made the training process faster without losing much quality.
In short, Google Cloud pulled off a massive, complex task, like directing a huge, high-tech symphony, and made it work smoothly!
NVIDIA Shows Intel Gaudi2 is 4x Better Performance Per Dollar than its H100
NVIDIA's flaunting their new supercomputer named Eos, packed with a whopping 10,752 of their H100 GPUs. If you were to buy this beast, it’d set you back over 400 million bucks. That's a lot of cheddar!
But here's the kicker: Intel's stepping up their game with this Gaudi2 thing. Turns out, it's giving you way more bang for your buck compared to NVIDIA’s gear. Like, we're talking about four times the performance per dollar. That's like getting four burgers for the price of one – sweet deal, right?
Now, in this tech race, NVIDIA’s still super fast, especially with their H100 GPUs. But when you compare the cost and all, Intel's Gaudi2 is a real eye-opener. It's cheaper and simpler, yet it still packs a punch.
NVIDIA's been busy tuning their H100, getting about twice as fast with their fancy FP8 support. But Intel? They're not too far behind, offering something that's sorta in between NVIDIA's older A100 and the new H100, but at less than half the price.
Getting emotional with ChatGPT could get you the best outputs
In a recent study, researchers found that if you talk to AI chatbots like ChatGPT with a bit of emotion, they give you better answers. Imagine adding lines like "This is super important for my job" to your usual questions. The study shows that this emotional touch makes the AI work harder, improving responses by about 8%. It's like the AI gets the feels and tries harder.
This is a big deal because it hints that AI might be getting closer to thinking and understanding like us humans, a concept known as artificial general intelligence (AGI). So, next time you chat with an AI, throwing in a little heart might just get you further!
AI is not good software. It is pretty good people.
This article argues that treating AI like traditional software is a mistake. AI, especially Large Language Models (LLMs) like ChatGPT, behaves more like humans than software. Unlike typical software, which is expected to deliver consistent outcomes, LLMs are unpredictable and can produce varying results for the same query. They lack the reliability and repeatability of traditional software.
It also suggests that thinking of AI as akin to people is more pragmatic. AI excels at tasks that are inherently human, like writing, analysis, and coding. They’re not so great at typical machine tasks, like performing consistent processes or doing math without external tools. AI's capabilities range widely, and understanding what they're good at requires working closely with them, much like you would with a human colleague.
However, this anthropomorphic view of AI raises uncomfortable questions. For example, if AI can replace certain human tasks, what does that mean for employment? And as AI becomes increasingly human-like, we must grapple with what tasks should be automated and which should remain exclusively human.
The Future of Music: How Generative AI Is Transforming the Music Industry
AI's been making some serious waves in music lately. First up, there was this AI song that went viral, "Heart on My Sleeve." It was so good it got everyone buzzing about AI's potential in music. Google's been in on the action too with MusicLM, a tool that cooks up tunes from text prompts. Then there's Paul McCartney, who used AI to bring John Lennon’s voice back for a new Beatles jam. And get this, Grimes is letting folks use her AI voice for their tracks and splitting the dough with them.
There are five big ways AI's changing the game:
Real-time Music Streaming: Apps like Endel and Brain.fm use AI to cook up endless playlists that change based on what you're doing. Spotify's got this AI DJ that curates music based on your taste. The future might even have AI making whole new songs.
AI Covers: After "Heart on My Sleeve" dropped, AI covers blew up. There are apps that can change your voice into someone else's, but there's some legal gray area here.
Royalty-Free Tracks: For folks making YouTube videos or podcasts, finding good, free music is tough. AI's stepping in to make unique, royalty-free tracks, making it easier for creators and small businesses.
Music Generation: This is huge for bedroom producers. AI can help flesh out a tune, convert audio to MIDI, and even mix tracks. It's a game-changer, especially for folks without formal training.
Professional Tools: There are browser-based tools, AI-powered plugins for music software, and even whole new digital audio workstations built around AI. This is big for pros and newbies alike.
What's the dream? A tool where you give some input - like text, audio, or even a video - and AI helps you make a song. The goal isn't to replace human music, but to help more people make music easily.
