Hello There:

We have been marveling at the curated list of 100 generative AI tools we assembled to help guide your exploration of this brave new world. The breadth of this first wave of tools is already astonishing, and one can only wonder how all this newfound power will impact the world at large.

We’re all familiar with AI tools that will set our writing free, leaving an indelible mark on the art of copywriting. Or the ones that produce uncanny images with just a text prompt. But there are also apps that will tell you your horoscope. Or convert all your existing training materials into a podcast complete with your cloned voice. Then there is the “legal assistant” that scans publicly available databases to find violators. Scary!

There are also tools that will help you set up your community, your event or your travel itinerary with just a few short sentences. The sheer creativity that generative AI has already inspired offers a tiny glimpse of the future of digital applications.

Did you know that Character.ai, which fields avatar-clad chatbots conjured up by fans of cartoon dreams, attracted 553 million visits with an average session duration of 30 minutes? Brings new meaning to the term stickiness! Traffic was so heavy that we had to wait for a spot in line to chat a few minutes ago. I finally was able to ask “Lily” why Character.ai was so popular, and she responded better than any PR person could possibly have:

YOUR ASSISTED FUTURE

Assistants of every possible persuasion, plus a bevy of copilots, the preferred term in medical circles, will rule your day by early 2025. We know this because GTPE says it has already collected 2,264 AI tools. We thought we would spare you the agony of plowing through such a large list to help find what you need, knowing you suffer from Time Compression.

We currently estimate that the AI app universe is growing at a rate of about 10-20 tools per day. That means that by the middle of next summer, the generative AI universe will exceed the number of digital marketing apps, which stands at about 10,000 today.

Of course, many digital marketing apps will also be acquiring AI features because having a spec sheet that mentions “AI-enhanced” or “AI-powered” is de rigueur today. In 2019, The Wall Street Journal reported that the number of companies using a .ai top-level domain from the British territory Anguilla had “doubled in the past few years.” As of May 27, 2023, there were 174,000, up 156% over the past year (suggesting a May 2022 figure of about 111,500).

That means some 62,500 eager, new .ai players are getting ready to blow us away sometime in 2024. We can’t wait. 🥳 

CHATGPT VS. THREADS

ChatGPT was the fastest-growing app in the history of mankind until Meta’s Threads appeared. Similarweb reported last week that the number of ChatGPT users actually declined by 9.7% in June to 1.6 billion total visits. That is still a staggering figure, more than the 1.2 billion visits Bing received that same month.

In April, The Verge surveyed 2,000 U.S. adults and found that 57% of respondents had either heard of or used ChatGPT. Unfortunately, that was among an online sample, so not entirely representative of the overall U.S. population; plus, combining “heard of” with “use” muddies the water.

In 2018, Forrester reported that there were 1.25 billion information workers globally. Many equate information workers with “professionals,” and we know that there are about 88 million professionals in the U.S., or 60% of the total workforce in 2020.

The Verge’s finding of 57% closely matches the percentage of professionals, so let’s assume for a moment that all professionals in the U.S. have heard of or used ChatGPT. We also know from past research studies that innovators and early adopters comprise about 16% of the market, close to the top quintile of 20%. If 16% of the 88 million U.S. professionals have used ChatGPT, you have a total addressable market (TAM) of 14 million.

Fourteen million out of 88 million. The vast majority have no clue about the wide choice of available apps, which will significantly impact their work. A study released in March by Goldman Sachs speculates that as many as, or as few as, depending on your point of view, 300 million people could lose their jobs due to artificial intelligence products like ChatGPT.

Threads picked up 100 million people in one week, out of 2 billion Instagram users who were just a few clicks away from registering and 5.2 billion global internet users. ChatGPT is a one-to-one application. Threads is a many-to-many social network. Only a few hours on Threads, however, reinforced the need for a different kind of platform. One that is focused on sharing mission-critical information that can be used productively.

YOU COULD WASTE TIME ON THREADS OR…

You could help those 74 million hapless souls from losing their livelihood by joining the Toolhacker community, which is laser-focused on the productive use of generative AI and innovation. We will formally launch our Toolhacker community on the Mighty Networks platform next week. It will bring together imaginative people to collaborate on generative AI projects to help rocket innovation practices in a rapidly changing world.

For more information, please visit our new Toolhacker site, where you will find plenty of information on what our community is all about and why we believe it will play an important role in your future success:

A formal invite will follow soon, so please be on the lookout for it.

Chiplets

The artificial intelligence boom has forced chip makers to develop a faster way to create new designs by stacking chips together like high-tech Lego pieces, called “chiplets.” This shortcut to more-powerful chips is one of the most significant advances since the dawn of the integrated circuit (IC) more than 60 years ago. In honor of this breakthrough, we have labeled this section of the newsletter dealing with innovation and trend shortcuts, Chiplets.

Google Med-PaLM 2

According to The Wall Street Journal, Google has created a medical AI chatbot, Med-PaLM 2, that is designed to answer medical information questions. It has been undergoing testing at the Mayo Clinic research hospital and other institutions since April. Physicians found that Med-PaLM 2 provided some inaccurate and irrelevant answers. However, in terms of reasoning, consensus-supported answers, and correct comprehension, Med-PaLM 2 performed as well as actual doctors.

Cars That Fly and Let You Take Showers

And now a bit of general innovation news. Yes, you read that headline right. Someone is working on a George Jetson car, the Alef Model A, which will be the first car that can drive and fly you to your destination. And it looks the part: half submarine, half 1950s sci-fi automobile. We will believe it when we see it. This Fiat Topolino EV, however, will debut in Italy by yearend, and yes, that is a man washing his hair using the Fiat’s built-in showerhead:

Two Big “C” Trends

The two biggest trends affecting society today are China and climate change. The mounting tension between the West and China will greatly affect the AI revolution. Nvidia’s Chief Financial Officer Colette Kress was quoted as saying:

Well, buckle down, Colette, because, from all appearances, the situation will only get worse. An invasion by China of Taiwan will make the Ukraine war look like a day at the beach. According to South China Morning Post, just over half of mainland Chinese support a full-scale war to take control of Taiwan, according to a new survey that offers a rare insight into public opinion as Beijing acts increasingly aggressively towards the island. The repercussions will be massive, so be forewarned.

At the same time, you are seeing an acceleration of global warming trends. The flooding in New York and Vermont is a preview of things to come, and this trend will be extremely disruptive. One of the most glaring shortcomings is the lack of a comprehensive national rainfall database. Sounds like a great opportunity for an AI startup.

We’re happy to introduce another newsletter contributor who will add a different dimension to the tech industry. Due to the very sensitive role of digging up the inside scoop, our intrepid reporter has to wear gloves and remain anonymous. However, there is one clue: job description. — Ed.

Prompt

Hello! It’s my pleasure to make your reconnaissance. My name is Prompt, and I’m an engineer by day and a rabble-rouser at night, or as some people like to put it, business in the front, party in the back! 🤭

I hang around bars and restaurants to score the deets on AI startups, specializing in outing their shenanigans. I’m sure you get what I mean without being explicitly programmed to do so. Like ChatGPT, I thrive on regurgitating facts, but I also hallucinate occasionally, so just deal with it.

Speaking of spewing facts, my industrial spying duties involve keeping my head low the ground to hear what’s happening in the Valley of the Jolly Green (#008000) Giant and also exotic far-flung places, like Toronto, where the Cohere folks are said to make their home. Great to see the Canucks score valuable lineage in The New York Times.

But what’s funny, or sad if you’re the CFO-type, about their story is that Cohere will have to give much of that $270 million it raised to companies like “House of Lannister” Google, which gives it oodles of desperately needed computing power.

And they’re not alone, OpenAI will have to pay much of the $13 billion invested by Microsoft to Satya Nadella & Sons. It’s the generative AI version of “You scratch my back, and I’ll pass your body through a wood chipper.” 😏

Speaking of Google. Do you recall the unverified rumor about a Google engineer leaking a memo that used a term straight out of the middle ages, "no moat,” to highlight the company’s vulnerability to OpenAI?

Well, it turns out the rumor was true, confirmed by none other than Google DeepMind CEO Demis Hassabis. In honor of this remarkable feat where fact follows fiction, I have come up with a new T-shirt, which required quite a bit of prompt engineering to say so myself:

BTW, the same company responsible for publishing that moat memo was SemiAnalysis. This week, they made the news again because they produced an excellent analysis of GPT-4’s architecture, infra, costs, and much more, that was hidden from public view behind a paywall. A person named Yam Peleg posted a summary of their analysis on LinkedIn, which he had to remove a few hours later due to copyright infringement.

Of course, once you publish something on the internet, it’s in the public domain. Sure, enough, someone copied and pasted Yam’s LinkedIn post on, where else, Pastebin. Here’s the most salient quote:

This is guaranteed to make Muskrat unhappy.

In keeping with the spirit of this newsletter, I wanted to use an image generator to create the above T-shirt mockup, but the images created by Bing and Adobe were so inappropriate I thought it would be better to link to them to prevent innocent eyes from seeing what tech had wrought, but then I thought what the hell:

OMG was my first reaction looking at that Bing image. What is that? Yiddish? That can’t possibly say, “Yes we have no moat!” Firefly was no better. The geniuses in San Jose thought we should give the model some herpes to distract from the “Yes we no mate” motto! Is it true Adobe CEO Shantanu Narayen made $31 million last year? You’re way overpaid, Dude! Maybe you should let the employees mate! 😂

But compared to image generation, text generation is a slugfest. Jasper CEO Dave Rogenmoser says his app has 4 million subscribers to its tips newsletter and 100,000 users. Obviously, these figures have been heavily rounded up as marketing types are wont to do, but at least it provides a comparative benchmark. Would a father with three boys lie? Nah! 😜

The AI writer arms race may have been started by Copy.ai, which boldly claims that “9.000,000 professional & teams choose Copy.ai.” Does “choosing” mean actually paying for the service?

Starting this September, Harvard will launch a coding class taught by an AI bot. "Our hope is we can eventually approximate a 1:1 teacher: student ratio for every student in CS50 by providing them with software tools 24/7 to support learning at a pace style that works best,” Harvard Professor David Malan was quoted as saying. No word yet on whether the students will all be bots too. 🤖

You know there is a gold rush in the generative AI market, and that includes vendors and their favorite meeting spot, conferences. Prompt was curious to see what U.K.-based Gen AI Con was planning stateside. Their site has San Francisco slated for an event on Oct. 5-6, while New York would get their turn on Dec. 7-8. Both dates are right around the corner, so Prompt promptly clicked the button that said “Explore” and BLIMEY if both events didn’t produce a 404 error! I guess the gold rush has turned into a Coldplay, with apologies to Chris Martin.

To contribute rumors, innuendo and general havoc, please email [email protected]. Just so you know, your contributions will remain anonymous.

Events Calendar

• Building Blocks of Generative AI Systems and Hands-On Experience (Carnegie Mellon University Generative AI Innovation Incubator) – Virtual event, Jul. 18, 2023. Zoom registration: Free.

• AI4 (Ai4 LLC, Brooklyn, NY) – MGM Grand Las Vegas, Aug. 7-9. Standard Pass: $2,195.

• VOICE & AI Conference (ModevNetwork LLC, Arlington, VA) – Washington Hilton Hotel, Sept. 5-7. All Access Pass: $1,195.

• Efficient Generative AI Summit (Kisaco Research, U.K.) – Santa Clara Marriott, Sept. 11, 2023. Delegate: $1,199.

• Generative AI Week (IQPC Ltd., U.K.) – Loews Atlanta Hotel, Oct. 17-18. Bronze Pass: $999.

• AI for Marketers Summit (Global Strategic Management Institute, San Diego, CA) – Virtual event, Nov. 15-16. Earliest Bird: $449.

• The AI Summit New York (Informa, U.K.) – Javits Center, Dec. 6-7. Delegate: $1,399.