• Uncharted AI šŸ¤–
  • Posts
  • A New Speech Model Too Dangerous to Release, an AI AppStore, and More This Week in GenAI

A New Speech Model Too Dangerous to Release, an AI AppStore, and More This Week in GenAI

Today weā€™ve got chatGPTā€™s very own Sherlock Holmes writing the newsletter intro:

Ah, a hearty greeting to you, as we converge on yet another chronicle of events and knowledge in the form of 'The Week in ChatGPT' missive. We navigate through the labyrinth of innovative mechanisms and converse about the freshest dispatches circulating the Society of Artificial Intelligence.

Should you possess any intriguing contraptions or snippets that you ardently desire to see in the limelight, I implore you to respond to this electronic letter or deposit them in my Twitter dms.

If perchance, this correspondence has reached you through the generosity of an acquaintance, and you find yourself somewhat bemused, I propose you enlist yourself for future editions at this juncture (i.e. subscribe). This way, you might find yourself in a similar situation, passing on the torch of knowledge to your confidantes one day.

Onward, then, to the enigma that lies in the future, my dear reader!

Thanks Sherlock for that wonderfully loquacious intro. Now,

Before we get into this weeks news, a meme for the ages.

I share this kids ambitions for subscribers.

Ok, down to business.

News Headlines

  • EU AI Act Passed šŸ›‚

  • An AI Stat Line Breakdown šŸ“Š

  • Metaā€™s New Speech Model šŸŽ¤

  • Introducing the OpenAI App Store šŸ˜±

  • Amazon nā€™ AI šŸ˜‚

Bodacious AI Stuff

  • QR Codes, But Cool? - Isnā€™t this way cooler than a bunch of black pixels in a random pattern? I can see a lot of potential for these to become a new format for visual ads. You can make your own QR codes similar to these here (scan with Code Scanner on ios).

     

  • BodyScan - Snap a couple pics of yourself and get AI to draft you up an personalized workout regiment tailored to your body. I also just thought it was funny how the website creator used AI generated images for the models. Itā€™s reassuring to know I can still tell the difference (for now šŸ˜±).

āœˆļø An AI Flyby for the Week āœˆļø

EU AI Act Passed šŸ›‚

Yeah so in case you havenā€™t noticed, the European Union really likes to regulate stuff. First is was big tech and now theyā€™ve set their sights on AI with their most recently passed EU AI Act.

The proposed legislation, which is expected to come into effect in a few years, aims to shield consumers from potential threats posed by AI, including surveillance, discrimination, and misinformation. However, it also might rain on some companies' parades, such as OpenAI, which could consider pulling out from Europe due to the legislation's restrictions.

Here are some key aspects of the EU AI Act:

  1. It introduces bans on certain AI applications deemed unacceptable, such as emotion-recognition AI in policing, schools, and workplaces, real-time biometrics and predictive policing in public spaces, and social scoring by public agencies.

  2. It establishes restrictions on high-risk technologies. For instance, generative AI will face new regulations and a ban on the use of copyrighted material in training sets, such as those used by large language models like OpenAIā€™s GPT-4.

  3. It escalates the category of recommendation algorithms on social media to "high risk," which means these systems will be subject to more scrutiny.

It seems that the EU is in a sprint to outpace the US, China, or Japan in the race to regulate AI. This move positions the EU as a front-runner in tech regulation, potentially setting the standard for global practices.

While the EU is donning its superhero cape to protect the world from AI risks, there are concerns that the new regulations might stifle innovation in the region. Fines for non-compliance could reach up to ā‚¬20,000,000 or 4% of the company's revenue per infringement, which could make some companies think twice about their open source strategies or API offerings.

As of now, it appears that none of the top 10 AI language models meet the requirements outlined in the upcoming AI Act guidelines. So, the clock is ticking for AI companies to either shape up or ship out. If AI models decide to withdraw from the EU, this could further diminish the EUs chances of staying on top of AI in the coming years.

Iā€™m pro some regulation, but I think the EU may be overstepping a bit

Psst ā€” if any of the links to articles I attach are paywalled or blocked apparently the GPT-4 browsing extension can bypass that pretty easily. But you didnā€™t hear that from me.šŸ¤

An AI Stat Line Breakdown šŸ“Š 

I thought it might be helpful to use this section to do a big picture AI breakdown backed up with some good old statistics. Pulling out some of these AI stats at the dinner table is bound to get you some of that respect you deserve šŸ˜¤.

1) The AI market is saddling up for a wild ride, expected to grow at an annual rate of 37.3% from 2023 to 2030ā€‹ [src].

2) It's projected to buff up the US GDP by a whopping 21% by 2030ā€‹ [src]

3) Surprisingly, in terms of domestic market penetration, 42% of US adults have never heard of ChatGPT and it is primarily being used by younger generations [src].

4) Over 60% of companies are betting on AI to supercharge their productivityā€‹, and a quarter are looking to AI to fill in labor shortages. Bottom line is youā€™re going to have to work harder if you want that employee of the month plaque [src].

5) In the global AI race, China seems to be leading the pack, with 58% of companies deploying AI. Meanwhile, our very own US of A seems to still be warming up, with only 25% of companies currently using AI [src]. Cā€™mon now letā€™s pump up those numbers.

6) But what about the jobs, you ask? Well, buckle up, because AI and other technologies could create up to a whopping 97 million new jobs! Thank god cause Iā€™ll definitely be needing one of those [src].

Metaā€™s New Speech Model šŸŽ¤

So, Meta made a new AI model. No, surprise there. They are quite prolific when it comes to doing that.

The zinger is that their model, Voicebox, a voice emulation model, is so good that they donā€™t want to make it public for fear of it being misused.

Basically, imagine grandparent ransom scams so convincing that we canā€™t even blame them for sending the money anymore because it sounds exactly like little Jimmy! Or, as a more serious example, an indistinguishable voice recording of a world leader making threats against another nation. The list of potential misuses grows every larger the more you think, so I can understand why Meta made the choice not to open source this particular model. But, weā€™ll still go over some of its features just for fun.

Standout features:

  • actually good background noise removal (cough, cough Zoom) and content editing

  • convert your speech to different languages while maintaining your tone and cadence! Big one! Youtube acquired an entire startup just to do this, so that they could start dubbing their video content.

  • 20x faster than current models šŸ«Ø

OpenAI APP Store? šŸ˜±

So OpenAI has a phone app.

They have plugins.

Whatā€™s the next step then?

Apparently, it might be an AI app store.

Picture this: AI models that can sniff out financial fraud or answer business queries could be up for grabs for all businesses, almost like buying a fancy new gadget from an online storeā€‹. Pretty cool, no?

OpenAI could also get into the customized AI chatbot game and probably dominate the whole market by offering that kind of service on their app store, along with a whole host of other models and services.

Companies like KhanAcademy and Aquant are reportedly eager to host their AI models on the store. We also just saw a round of investments from OpenAIā€™s $175 million investment funds and I wouldnā€™t be surprised if some of the companies are being primed for acquisition to be offered up on this app store.

Just think about the effect the App Store had on the iPhone ecosystem. People are speculating we could be approaching a similar moment for AI.

Oh snap.

Amazon nā€™ AI šŸ˜‚

Every time you look away to all the other shiny stuff happening with AI, you can know with relative certainty, that Amazon is up to something.

As we speak, Amazon is gunning to be the premier AI model marketplace, pitting them against OpenAI if they decide to go ahead with their AI App Store.

Amazon already offers to dozens of foundational models from Stability, Falcon, and Anthropic via its ā€œSageMaker JumpStartā€ marketplace. Through Bedrock, clients have the option to:

  1. Pick ā€˜n choose from the best AI models.

  2. Optimize them using your private data and IP.

  3. Search your data, receive decision-making help, automate issues.

Many industries with more unstructured data then they know what to do this are eyeing Amazonā€™s offer and are ready to bite.

But, what I want to get across is that Amazon is silently building up the infrastructure for enterprises to build their own AI solutions using these models. They are also providing the flexibility to let companies mix and match and combine models as needed, which is pretty unique as things are.

And thatā€™s not all. There is also some buzz that Amazon could be starting their own book-making arm, using generative AI to cowrite and publish their own books! I know I said I would be fine reading books written by AI, but I might on that some more.

Parting is Such Sweet Sorrow (But Please Keep Reading LOL)

Well that's a wrap on the newsletter. Before I launch into my ending monologue, please take a moment to drag this email to your primary, so that you never miss it each week! Thank you!

Let me know how I did by filling out this feedback form or just reply to the email like you would any other. If I did well and you're in the giving spirit please support the newsletter by buying me a coffee or donating to my gumroad page. Most of the money you donate to me will go back into purchasing AI products to try out and review on the newsletter. Bernie says it better:

Also if you have a product that you would like to be featured on the newsletter remember to add it here!

Lastly, if you'd like to refer a friend just tell them to visit chatgptnewsletter.com to subscribe! Or you could always just forward them the email without any subject header. I'm sure they they would love that.

Cool Things You Should Try/Buy From Every Past Newsletter šŸ¤Æ

(Newsletter 1) ==> Consult domain experts in your browser with ExperAI

(Newsletter 3) ==> Pay for DoNotPay lol. Irony aside, it really is pretty cool.

(Newsletter 4) ==> Multion.AI - not your average burger buying extension.

(Newsletter 5) ==> Why settle for regular email when you could have Intellimail?

(Newsletter 6) ==> The AI app store where you can Cookup just about anything

(Newsletter 7) ==> A cute, fuzzy AI-powered online meeting summarizer: Otter.ai

(Newsletter 8) ==> Tired of taking hours to find the perfect online purchase? Getproduct.helpā€™s got you covered

(Newsletter 9) ==> Turn your Scribbles into art with ScribbleDiffusion

(Newsletter 10) ==> Ever wanted a search engine for your entire online life? Rewind.ai

(Newsletter 11) ==> Your Online Image Studio šŸŽ¬: Clipdrop

(Newsletter 12) ==> Use Generative AI to create online courses on any topic you can think of. For example, you can create lesson plans about yourself before you get famous. Learn.xyz

(Newsletter 13) ==> Replace Siri with ChatGPTā€” HeyGPT!

(Newsletter 14) ==> Autonomous AI Agents in Your Browserā€” AgentGPT!

(Newsletter 15) ==> Create your own AI music with musicfy

(Newsletter 16) ==> Detect AI written content more reliably with GPTZeroā€™s browser extension.

(Newsletter 17) ==> Another AI model playground, but more than just openAI models

(Newsletter 18) ==> Digitize your mind with You.ai šŸ˜µā€šŸ’«

(Newsletter 19) ==> Learn how to effectively use AI in your work and life with Maven

(Newsletter 19) ==> Super stylish generative AI website builder Dora. Explora whole new way to make websites.

(Newsletter 20) ==> A TODO list that does itself? Sign me Spellpage!

(Newsletter 21) ==> Making presentations just got easier again with Gamma

(Newsletter 23) ==> Making news concise with AI

(Todayā€™s Newsletter) ==> Reply with what should go here!!!!

Wow that was a lot

Reply

or to participate.