Google AI and OpenAI Left Scrambling: Open Source Projects Outpacing Giants in the Race for AI Dominance, as Leaked Documents Reveal Vulnerabilities and Challenges Ahead
Google AI Documents Leak “We Have No Moat, And Neither Does OpenAI” – Natural 20
The text below is a very recent leaked document, which was shared by an anonymous individual. Source is a researcher within Google.
A Google researcher dropped a truth bomb in a leaked document, admitting that both Google AI and OpenAI are falling behind in the AI arms race. Turns out, open-source projects are kicking their butts and takin’ names. Major advancements thought to be “open problems”? Already in people’s hands, thanks to open source.
Imagine having a personal AI fine-tuned on your laptop in a single evening, or using your phone to run AI models faster than a rabbit on a hot tin roof. Open source made it happen, and they’re doin’ it faster, cheaper, and with more pizzazz than either Google AI or OpenAI.
The key takeaway? Google AI and OpenAI better start learnin’ from open source and consider where their value really lies. The best models come from rapid iteration, and the open-source community is proving that size doesn’t always matter when it comes to AI parameters.
How’d this happen? Meta’s LLaMA got leaked, and like ants to a picnic, the open-source community swarmed in. They innovated like there was no tomorrow, and now regular folks can dabble in AI with nothing more than a laptop and a free evening. Talk about a game-changer.
Well, folks, the open source LLM world’s blowing up, and we should’ve seen it coming. Just like image generation, cheap methods like low rank adaptation (LoRA) made it easy for the public to jump in and create some amazing stuff. It’s like a gold rush of ideas, and the big players are being left in the dust.
Turns out, using LoRA makes updating these models a breeze, and it’s way cheaper than starting from scratch. People are coming up with updates that cost only around a hundred bucks, and it’s making big models like ChatGPT look like they’re standing still.
These open source projects are all about quality, not quantity. They use smaller, carefully crafted datasets that actually work better than big ol’ piles of data. And here’s the kicker: trying to compete with open source is like trying to nail Jell-O to a tree.
The truth is, we need each other. Researchers are switching companies faster than a New York minute, so there’s no point in keeping secrets. And you know what? It doesn’t matter what OpenAI is doing. Open source is gonna take over unless they change their tune too.
So, let’s join the open source party and start learning from each other. We might have to let go of some control, but that’s the way the cookie crumbles. It’s about time we embrace the future and hop on the open source bandwagon.
Google AI Documents Leak about “Google and OpenAI”
See the full article:https://natural20.com/google-ai-documents-leak/
New LinkedIn AI Feature Might Actually Help Get You Hired
LinkedIn’s testing a new AI feature that might actually be helpful for job seekers. The platform’s offering paying subscribers an auto-generated message feature, which creates personalized notes to hiring managers based on the user’s profile. The messages are like short cover letters, taking info from the user’s LinkedIn bio and making a professional appeal.
The AI uses info from your profile, the hiring manager’s profile, the job description, and the company to create a tailored draft message. But, like all AI tools, you’ll need to double-check its work before sending it off.
There’s a catch, though. This feature is only being tested on LinkedIn’s Premium customers, who pay a pretty penny for their subscriptions. Plus, it’s not available to all Premium users just yet – it’s rolling out slowly.
AI isn’t perfect, and it’s definitely not the end of human writing, but for tedious tasks like cover letters, it might be a time-saver. If all goes well, this feature could help job hunters with just a couple of clicks. But there’s always the risk of flooding hiring managers’ inboxes with spam. LinkedIn, however, claims the tool will make messages more informed and contextual, which might just make them less spammy in the end.
WordPress Explores AI Integration
WordPress peeps are chit-chatting about AI joining the party, but don’t go thinking they’re marrying it to the core just yet. They’re just testing the waters, seeing what’s what.
Seems like they agree that AI would be better as a plugin sidekick rather than a core teammate. Matt Cromwell’s got a point – with a full plate of projects, adding AI to the mix could be too much too soon. But talking about it? Why not!
Ollie Jones pitched the idea of an AI co-pilot, something to diagnose problems like plugin conflicts and offer fixes. Imagine if AI could spot trouble and keep your site running smooth – pretty slick, right?
Robglidden threw in another two cents, suggesting an AI chatbot as a user in the upcoming WordPress phase focused on collaboration. Hey, more hands (or bots) on deck never hurt, right?
But surprise, surprise! AI’s already made its way into WordPress. There are plugins for SEO and content optimization that’ve brought AI into the mix. So, the future’s kinda already here. At least, in plugin form.
The WordPress gang’s talking about AI shows they’re keeping up with the times, so don’t count ’em out yet. Who knows where the convo will lead? If you’ve got ideas, join the official discussion and share your thoughts.
READ THE ARTICLE ON SEARCH ENGINE JOURNAL.
Aisera embraces Microsoft AI copilot for advanced enterprise service experience
Aisera, a company providing AI-based tools for customer and employee service, has joined forces with Microsoft to offer an AI copilot using Azure OpenAI service, including the fancy GPT-4. This high-tech helper aims to make service experiences more personalized and smooth for businesses, while also saving them some cash.
By hitching its AI service experience platform (AISX) to the AI copilot, Aisera plans to boost self-service, productivity, and user engagement while speeding up problem-solving. The result? Chit-chat that feels more human and makes sense for each company.
With the AI copilot, folks can ask questions using everyday lingo through a chatbot or voice, and get answers that are actually relevant to their workplace. The copilot is all set to handle a bunch of support tasks, making life easier for swamped support teams.
A few big names, like Chegg and Gap, have already taken the AI copilot for a spin. Chegg, for example, has seen impressive improvements like 75% of support tickets solved automatically, happier employees, and a more productive service desk.
Aisera is just one of many companies that are jumping on the OpenAI bandwagon to enhance their products, showing that AI is here to stay and make our lives a whole lot easier.
READ THE ARTICLE ON VENTUREBEAT.
Private ChatGPT? Microsoft to offer AI alternative for businesses with sensitive information
Hey, y’all! Microsoft’s cookin’ up a solution for businesses that are sweatin’ bullets over data leaks and playin’ by the rules. They’re workin’ on a ChatGPT just for them, hosted on their fancy-schmancy Azure cloud servers. Now, those folks in healthcare and finance can jump on the AI bandwagon without worryin’ about their secret sauce spillin’ all over the place. And hey, this’ll keep their info from helpin’ train ChatGPT too. So, it’s a win-win, y’know?
READ THE ARTICLE ON INTERESTING ENGINEERING.
Snapchat is bringing ads to Spotlight and ‘My AI’
Snap’s hoppin’ on the ad train, adding ’em to Spotlight (their TikTok lookalike) and “My AI” chatbot. These features have been hot, so Snap’s hoping to bring some green back to their ad business that’s been down in the dumps lately.
With over 350 million users watching snappy short vids on Spotlight, it’s a big deal. Creators might get a piece of the pie too, with possible revenue sharing. Snap’s been hush-hush about that, but it’s not too wild to think it might happen.
Ads are also sneaking into My AI chats. These sponsored links will try to fit right in, like recommending a local eatery or game store when you’re chatting about grub or gaming. It’s still an experiment, but Snap’s diving in headfirst, despite some folks not being too keen on their chatbot buddy.
Long story short, it’s gonna be interesting to see if Snap can make bank on these AI features without ticking off users.