What's up? You're reading Inclined AI. It’s muffin Monday. Okay, I made that up, but treat yourself anyways.

Here are the blueberries in our metaphorical muffin:

• Meta announces LLaMA, a collection of language models

• Schools are learning to adjust to AI in the classroom

• OpenAI describes their feelings about AGI

• Spotify’s DJ tool is not getting rave reviews

META’S LLaMA SHOWS ZUCKERBERG’S COMMITMENT TO STAYING OPEN

LLaMA stands for Large Language Model Meta AI, an entire family of models only accessible to researchers (at the moment). Meta released these models with one goal above all the others: to prove they’re more open than OpenAI.

Let’s pop open the hood and see what’s cooking:

1/ Their collection of models ranges from 7 billion-65 billion parameters. 

If you don’t know, parameters in AI are the variables that help determine the output (response) based on their measures of the input (prompt).

2/ LLaMA-13B, the 13 billion parameter version, can outperform GPT-3, according to Meta. 

This claim is big if true since OpenAI’s model is 165 billion parameters.

3/ The models run efficiently and perform better with fewer variables.

When you combine all that, you get a model that can potentially run on a single GPU.

But Wait, There’s More

The last point there is a biggy. Most models take several GPUs to run, making them inaccessible to the general public.

With this step forward, people can start to run their own LLMs on their computers and, in a short amount of time, even their smartphones.

I want to avoid digging into the weeds too far. So here’s the best way to grasp the unique part of the announcement:

OpenAI, Google, DeepMind, and others train their LLMs on private datasets. Meta created these models with data available to the public.

All that means it’s easier to replicate and more open than any of the big player’s systems that came before it.

Expect a Stability launch soon that gives you customizable chat AIs and much, much more.

AI BECOMES THE TEACHER’S PET

AI was off to a rocky start with teachers. Students everywhere walked into English classrooms with their heads held high and a solid B essay in their backpacks.

Of course, we know now OG chatGPT wrote many of those papers, and those students learned squat. It’s not a great look.

I’d argue it’s why Microsoft does not allow Sydney/Bing chat to write essays anymore.

This Semester Things Are Different

I heard a story on Reddit the other day about a professor who gave his class an essay written by ChatGPT. The student’s job was to proofread it and find all the errors.

In other words, college students start a second draft and learn from the AI’s mistakes. It’s brilliant

All over the country—excuse me—all over the world, education systems are learning to embrace AI and teach their students to use it with finesse.

That approach will pay off in a big way.

Every kid out there will tackle AI with undaunted confidence. It’s like when my English teacher in high school taught me how to properly use Wikipedia as a source aggregator instead of a one-stop shop. You learn to respect the tools and their features this way.

The grants and government funding are incoming. Watch this concept flourish in the next decade. I promise.

Quick Nuggets

⚙️ OpenAI plans for AGI and beyond in their latest blog

🔭 1 guy used AI to track down Chinese spy balloons

📌 The humans behind building AI aren’t treated well

⭐️ Hollywood stars and celebrities, do they support AI? Let’s find out.

🎥 Movies imagined AI over a number of films, so what did they get wrong?

✏️ People’s bios are filled with factual inaccuracies when AI makes them

💬 ChatGPT needs to cross the chasm into an investable product

🎤 1 voice created thousands of deepfakes—the latest AI reporting from The Information

🎧 Spotify’s DJ gets lambasted by WIRED

🛡️ Can AI protect itself from prompts that attempt to break it?

❗️ 1 year ago testers of Sydney raised concerns about inaccuracies and wild hallucinations in Sydney — BFD 🌟 

🏢 Some companies are trying to replace workers with ChatGPT prematurely

🏦 More banks are telling employees that ChatGPT is off limits

🔥 Tech’s hottest job is becoming a prompt engineer

😬 Scammers are using Chat AI to spread malware, be careful

💵 Voicemod raises $14.5m to catch up to the generative AI movement

🔥 Fresh Products

• autowrite - human-like SEO-generated articles (link)

• TLDR; - summarizes chats from your Discord (link)

• Change Style AI - switch up the style of your photos (link)

• AskSteve - talk with Steve Jobs bot (link)

• Synthesis - a tool for learners (link)

• AInventory - optimize your inventory (link)

• Berrycast - easily record and send messages (link)

• ProductListing - supercharge your Amazon product listing (link)

• AV Mapping - end tedious process of music license (link)

• AnyAPI - add AI power to your product (link)

• Accio - generate resume summaries (link)

• Addy AI - write emails 10x faster (link)

• Curipod - helps spark student creativity (link)

• Verb - a tool for authors (link)

• Test & Start - generate new businesses ideas (link)

• HireLakeAI - extract and match resume data (link)

• Latte - social-first clips from long-form videos (link)

• PoplarML - insane fast deploys (link)

• tl;dv - have fewer, more focused meetings (link)

Good Content, Wild Animation

Watch this video. This graphics team uses AI tooling to change the way we understand animation design.

_{That’s it for today. I hope you enjoyed the latest edition of inclined.ai - Davis.}

Want more?

Upgrade today for less than the price of a latte a month and get benefits like:

• A weekly Sunday deep dive into AI trends

• Access to Q&A privileges to get your biggest AI questions answered

• Exclusive discounts & invites to beta AI tools

Click down below to support our writing 🙂

Be a sponsor of our newsletter