Welcome to The Week in GenAI newsletter, where we discuss and the latest news and products going around the generative AI community. If you have any products or tidbits that you would like to be featured, by all means please reply to this email or take to Twitter where my dms are always open!

Also, if your friend is forwarding you this email and you have no idea what is going on subscribe here, so you too can one day forward this to a friend!

Before we get into this weeks news,

Now, it’s time for

A Whole Lot of AI

News Headlines

• AppleGPT 🍏

• The All-in-one ChatGPT Special Segment

• Meta & Microsoft are the new M&Ms

• What’s a Worldcoin? 🌎🪙🤔

• 7 A.I Companies Walk Into A Bar (Except the bar is the White House)

New AI Products

• Talkio.ai

  Nows a great time to learn a new language because you and I know a certain AI that is quite multilingual. Talkio combines ChatGPT and voice models to let you practice having conversations in any foreign languages and dialects that you want to learn. Now, you don’t have to feel as much crippling self-consciousness as you stumble through speaking a new language. Hurrah!

   

• Todo.is 

  This AI enhanced TODO app isn’t doing your TODO’s for you yet, but until Spellpage releases whatever they have in store for us, it might be the next best alternative. Todo.is helps you break down your TODOs into manageable, actionable pieces, so you can get more done, faster.

  Maybe this is the app that gets me off of Notes checklists for everything 🤞.

✈️ An AI Flyby for the Week ✈️

Apple GPT 🍏

It was leaked a while ago that Apple intended to give Siri a long awaited update with project codename Bobcat, but now there’s a new leak for us to obsess over.

Apparently, Apple isn’t as out of the chatbot game as we all thought and has been using an internal chatbot called Apple GPT, built on their in-house model Ajax. Apple GPT is reportedly being used internally for product prototyping and is capable of creating essays, images, and videos based on text prompts (it can also interact with people in virtual reality 😳). I think it’s safe to say Apple employees are relieved they get to finally use some form of generative AI to assist with their jobs after Apple banned the use of ChatGPT over fears of leaking company secrets.

Apple GPT is supposed to be released as a consumer product in the next year, making it a direct competitor to other chatbots like ChatGPT and Bard. If Apple manages to create a competitive chatbot model, they get a nice 1.5 billion user boost (the number of iPhone users in the world). That’s a nice little advantage.

And like all of big tech, Apple has been on a hiring spree for AI talent. I imagine a lot of this fresh talent will be focused on making Apple’s AI projects safer and more privacy-minded, as is the emphasis with most Apple products.

READ MORE

The All-in-one ChatGPT Special Segment 

Ho boy. Buckle up because OpenAI has been all over the place.

Is ChatGPT/GPT-4 Getting Worse?

A recent study has conducted tests that indicate GPT-4’s performance has nose-dived by over 95% on some tasks. Take a look —>

March to June 2023:

• Finding prime numbers: down from 98% to 2.4%

• Writing good code: down from 52% to 10%

• Chain of thought reasoning: significantly more mistakes

Does this mean I’m a better coder than ChatGPT now? Nooooooooo (I’m still probably worse, which is cause for even more despair)

Some folks, though, are crying foul. A Princeton computer science professor, for instance, argues that the paper is showing a change in behavior rather than a drop in capability. And there may be a glitch in the evaluation too. A case of mistaken mimicry for reasoning, he suggests. If true, this would mean that we should stop treating GPT-4 as static or expect it’s capabilities to stay constant. They would fluctuate to reflect broader trends in it’s usage.

What could be the cause of this AI brain freeze? Some speculate that OpenAI might be prioritizing speed and cost savings over quality. Or perhaps, the model is learning from less accurate user inputs over time. It could also be a safety feature designed to say 'no' more often. The true cause, however, remains as clear as mud. OpenAI is investigating due to the general outcry of Techies on Twitter. Hopefully they can diagnose the cause, so I can continue to spam using ChatGPT for code. Godspeed.

ChatGPT Gets Released to Samsung Users

The ChatGPT app gets rolled out on Samsung. Ohhh, so that’s why it got so bad all of a sudden. Mystery solved.

Just kidding of course, I would never fan the flames of the Apple-Samsung rivalry 🌚. As a peace gesture, here’s a non-confrontational meme that is still at Samsung’s expense, but not to Apple’s gain.

OpenAI Takes Down Their AI-Written Detector

Does anyone remember OpenAI’s detector for AI generated content that they released a couple months ago? No? Well, you won’t miss it then because its been taken down due to subpar accuracy.

Great, so now you’re telling me ChatGPT’s getting worse, but we still can’t tell whether it’s human or not. Fantastic.

READ MORE HERE

Meta & Microsoft are the new M&Ms 

Not only do these guys get mentioned in the news every week, but they’ve both been at the forefront of the generative AI wave. Now, it seems they might be partnering up.

You must be thinking: wouldn’t that mean Microsoft is cheating on OpenAI with Meta? Yeah, Microsoft’s a little peeved at OpenAI at the moment:

• OpenAI is causing more drama than an elementary school playground. They’re going to court, getting letters from the FTC, and fighting with the EU.

• OpenAI is releasing two Microsoft competitors: An OpenAI Marketplace and an OpenAI Office Assistant.

So it makes sense that Microsoft’s looking for ways to hedge their bets and what better way to hedge than on the hottest, new open source AI model in town. That would be Llama 2, courtesy of Meta last week.

In Llama 2’s publication piece, it mentioned that Microsoft will help fund and support Llama 2 in return for using the model in Microsoft Azure - a spot originally reserved for OpenAI.

I said it last newsletter and I’ll say it again.

READ MORE HERE

What’s a Worldcoin? 🌎🪙🤔 

So, it may seem like OpenAI kind of unleashed Generative AI on the world without much notice, but it was actually a lot more mediated than you might think. Sam Altman started the Worldcoin initiative over 3 years ago in 2019 to prepare for an AI future. Worldcoin would be a blockchain identity authentication system that people could use to prove that they aren’t AI!

And with OpenAI’s AI detection system just getting the boot for subpar accuracy, we can see that efforts to accurately detect AI are running up against problems. Deepfakes and AI generated content are proliferating across the internet and it’s hard to tell what is genuinely human sometimes. Worldcoin is posited as the solution to this rapidly growing problem.

But, how do you ascribe someone their human status? Worldcoin decided retinal scans were the way to go. That’s right, they went full Men In Black.

To validate your Worldcoin ID and thus have your personhood confirmed you have to conduct an in-person retinal scan. You might be thinking isn’t this all a bit overkill? Worldcoin’s 2 million users in 3 days, $3 billion valuation, and plans to expand to 35 cities across 20 countries beg to differ.

If you want to get in on the action, sign up here.

READ MORE HERE

7 A.I Companies Walk Into A Bar (Except the bar is the White House)

7 A.I companies met to decide on an A.I regulation agreement.

The companies are AI’s top earners: Amazon, OpenAI, Microsoft, Google, Meta, Anthropic, and Inflection AI.

So far this is what they’ve come up with:

• Have an extra team vet new A.I tools for safety, bias, or misuse before being released.

• Get digital watermarks to label A.I content

• Pay bounties and sponsor teams to find bugs and loopholes in their A.I software

• Potentially even require a license for A.I models. The government had floated this idea earlier on as well.

That’s all well and good, but let’s pause and really think if the world’s top AI companies, who are profiting the most, should really be relied upon to unselfishly set policies that will shape the future of AI in the US. In addition to this concern, here are two others:

#1 - The public was not involved in these decisions.

#2 - The new regulations make it harder for smaller companies to grow. These companies will have to come up with millions or even billions to redesign their algorithms and hire new people to stick to regulations. See a problem?

AI is undoubtedly one of the most valuable resources right now and all I’m saying is that corruption and valuable resources go hand and hand across history. Maybe this time will be different.

READ MORE HERE

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, reply and I’ll try my best to get it on a future edition.

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 Weekly Newsletter 🤯

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

(Newsletter 2) ==> buy a talking book from Konjer

(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.help’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 🎬: 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 ChatGPT— HeyGPT!

(Newsletter 14) ==> Autonomous AI Agents in Your Browser— AgentGPT!

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

(Newsletter 16) ==> Detect AI written content more reliably with GPTZero’s browser extension.

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

(Newsletter 18) ==> Digitize your mind with You.ai 😵‍💫

(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

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

(Newsletter 24) ==> Making cool QR codes with AI

(Newsletter 25) ==> Want to write an book with AI? 👇

(Today’s Newsletter) ==> Reply with what should go here!!!!

Wow that was a lot