AMD bought... Who?

PLUS: Is ChatGPT slowing or growing?


Good morning, human brains. Welcome back to your daily munch of AI news.

Here’s what’s on the menu today:

  • Adobe’s spicy new AI image feature 🌶️ 🔥

    Generative Match lets you use references to make new images.

  • AMD’s newest feast? A heaping bowlful of 🦃 🥘

    It claims it bought the AI startup to provide turnkey AI solutions.

  • Is the ChatGPT app growing or slowing? 💸 🐮

    It beat a revenue record, but its growth slowed by over 10% from last month.


AMD gobbles up 🦃 🥘


In August, we covered AMD’s acquisition of Mipsology — the team behind the autonomous vehicle software, Zebra.

Yesterday, AMD slurped up — a company specializing in open-source AI software.

So what?

Well… Nod’s software reduces manual optimization through automation, which speeds up the deployment of high-performance AI implementation.

How many billions did it pay this time?

The financial details have not been disclosed. Sneaky, sneaky.

Ok, who is

It’s known for its complier-based methodologies, which presented a convenient alternative to traditional handwritten kernels.

It also developed the SHARK Machine Learning Distribution, which is recognized for its fast runtime capabilities for advanced BERT and Transformer-like models on GPUs.

I don’t trust people, what’s AMD's motive?

AMD claims it’s to provide turnkey AI solutions and enhance its commitment to open-source AI.

Our take: This is a direct countermove. NVIDIA partnered with Anyscale in September to accelerate LLM development (allegedly). AMD publically stated that by 2032, it wants to be a leader in the growing $594 billion AI market.


Horse - the internet browser where things get done.

😱 Are you stuck in tab hell?

🫠 Do you struggle to stay focused?

The internet is a wild, untamed beast - and tabs can’t help.

You deserve better than this mess.

Every task turns into a circus of unrelated articles, half-watched documentaries, and internet shopping.

Sure, you can dust your tabs away into a dark, forgotten folder and pretend you’ll visit soon. But you know the truth. Folders are a fancy tab graveyard.

Tabs are the problem. Trails are the solution.

Horse organizes your internet sessions with Trails - simple hierarchical trees that capture the natural flow of each journey.

You can fold, unfold, organize, and rename each Trail - and it couldn’t be easier to find that article about obscure ChatGPT prompts that you read three months ago.

Just ask our countless happy users:

  • “Horse is how a browser should have always worked.”

  • “Horse helps me to navigate the labyrinth that is the Apple developer documentation.”

  • “Horse has quickly become my go-to browser for focused work.”

 🐴 Rent your own Horse today - for just $99/year.

You get a 14-day money-back guarantee - so saddle up!



Mathematical functions that map input data into a higher-dimensional space. This makes it easier to identify and classify patterns.

They are also known as kernel techniques or kernel functions.

They were first used in the 1960s when the kernel perceptron (a variant of the perceptron learning algorithm) was invented.


How to get beta access to Adobe’s latest AI tool 🌶️ 🔥

Bot likes sexy and new things.

On Monday, we reported on Project Stardust — Adobe’s upcoming photo editing engine.

At its annual Max conference yesterday, Adobe introduced Generative Match — a new feature of Firefly, its AI image generation tool.

What does it do?

It allows you to generate new images by uploading a reference image and providing a text prompt.

What makes it sexy?

It lets you create a series of images that maintain a consistent style for your branding and marketing campaigns.

It’s been a long day, just give me an example.

Ok, so if you upload a picture of a cartoon cat and put “bear in the woods” as your text prompt, Firefly will produce an image of a bear in the cartoon style of the cat.

Sounds good, how do I use it?

It’s available to try today on Firefly beta and is “coming soon” to Creative Cloud apps.

We tried it here:


ChatGPT’s app is so “new money.” 💸 🐮

A cash cow or bottomless pit?

In September, the app experienced nearly $4.58 in revenue. It also saw 15.6 million downloads, which contributed to a lifetime total of 52.2 million installations.

So, it’s obviously the most popular AI app.

Actually, there’s a competitor called Ask AI that holds the number 1 spot. Its revenues peaked at $6.55 million in August.

Well, will it be the number 1 AI app soon?

Its growth rate is slowing. Its growth rate slowed to 20% in September from over 30% in the previous months.


AMD canceled production of the highly anticipated 8900 XTX graphics card and prioritized the upcoming M1300 AI chip. We covered that, here.


Think Pieces

Geoffrey Hinton on “60 Minutes.” He discusses the dangers and current state of AI.

Disney’s controversial AI-generated promotion. A Loki season 2 poster breaks the platform’s licensing rules on AI content.

The problems faced when creating commercially successful AI. Prioritize value at the beginning for your AI startup to stand a chance.

Startup News

Google unveiled a generative AI search for healthcare. It claims it will save healthcare workers time and energy.

IBM’s CEO claims he’s hiring more developers. Earlier this year, he claimed he was replacing 8,000 workers with AI.

AutoGPT’s hackathon results are in. AI engineers from around the world flew to San Francisco to compete.


Teaching LLMs to “forget” things — a paper on how Microsoft Researchers got AI to forget Harry Potter.

Visualization library — visualizations of basic concepts in statistics and signal processing.

MSA-Conv — a method that allows Vision Transformers to handle different sizes of images without the need to change designs or resize images.


Hotshot — a text-to-HQ GIF generation tool.

CreatorMind — An intuitive, versatile chatbot builder.

Code Companion — AI-powered tool that gives real-time coding assistance.

Airparser — automate data extraction from emails, PDFs, and more.


💡 Read Alts — Unique investment ideas for alpha-seeking investors, traders, and finance workers. Learn about exotic markets and alternative asset classes.


Renown Wharton University professor shows that GPT-4V is biased towards dogs. It absolutely loves them. It rates one dog’s cuteness a 15/10. Other dogs received 13/10 and 14/10.

Tag us on Twitter @BotEatBrain for a chance to be featured here tomorrow.


Until next time 🤖😋🧠

What'd you think of today's newsletter?

Login or Subscribe to participate in polls.